threads
listlengths
1
2.99k
[ { "msg_contents": "\nJust heard on tv, Japan is either considering or planning on trying\nDaylight Savings Time this year. Didn't one of the Tom's just finish\nthe timezone stuff?\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Wed, 24 Feb 1999 18:36:21 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": true, "msg_subject": "timezones revisited" }, { "msg_contents": "> Just heard on tv, Japan is either considering or planning on trying\n> Daylight Savings Time this year. Didn't one of the Tom's just finish\n> the timezone stuff?\n\nI just finished documenting the existing time zone info (and fixing a\nfew bad values). To accept a new Japanese TZ abbreviation, we would just\nneed to add it to the relevant table in dt.c. For output, we rely on the\nunderlying OS to provide the info.\n\n - Tom\n", "msg_date": "Tue, 02 Mar 1999 03:17:01 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] timezones revisited" } ]
[ { "msg_contents": "Thomas,\n\nIt looks like you committed some fixes in datetime.c and dt.c to the\nREL6_4 branch without committing changes in dt.h that they depend on...\nthey do not compile right now :-(.\n\nI got my copy to build by un-ifdefing a couple of things, but I assume\nthat's not the intended answer so I won't commit it.\n\n\t\t\tregards, tom lane\n", "msg_date": "Wed, 24 Feb 1999 20:50:34 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": true, "msg_subject": "BTW, datetime is busted in REL6_4 branch" }, { "msg_contents": "> It looks like you committed some fixes in datetime.c and dt.c to the\n> REL6_4 branch without committing changes in dt.h that they depend \n> on... they do not compile right now :-(.\n\nOK, thanks for the tip. I'll look at it...\n\n - Tom (just back from a short vacation)\n", "msg_date": "Tue, 02 Mar 1999 03:35:44 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: BTW, datetime is busted in REL6_4 branch" } ]
[ { "msg_contents": "Hi,\n\n I have seen what I concider to be a bug in either the libpq library or\nin the backend. To see the effects, first, install the execq() function\nfrom the SPI section of the Programmers Guide. Then do this in psql:\n\ntemplate1=> select execq('create user fred', 1);\nBackend sent D message without prior T\n\n At this point psql will hang. I have a patch for libpq that fixes this\nbut I am not sure if this is the right place for it. My questions are: Is\nit not reasonable to run \"utility\" queries from inside SPI? Is the problem\nin the client-side libpq interface or should something be done at the server\nend? If anyone is interested, I can explain why I want to do this.\n\n Any advice would be greatly appreciated.\n\nJerry\n\n\n", "msg_date": "Thu, 25 Feb 1999 14:02:58 +0900", "msg_from": "\"Gerald L. Gay\" <[email protected]>", "msg_from_op": true, "msg_subject": "libpq and SPI" }, { "msg_contents": "\"Gerald L. Gay\" <[email protected]> writes:\n> I have seen what I concider to be a bug in either the libpq library or\n> in the backend. To see the effects, first, install the execq() function\n> from the SPI section of the Programmers Guide. Then do this in psql:\n\n> template1=> select execq('create user fred', 1);\n> Backend sent D message without prior T\n\nThat would be a backend bug, for sure. It's a violation of the FE/BE\nprotocol to send data row(s) without sending a row description first.\n\n> At this point psql will hang. I have a patch for libpq that fixes this\n> but I am not sure if this is the right place for it.\n\nI do not believe it is really possible to \"ignore\" this error inside\nlibpq. Without the initial T message you have no idea how many fields\nare in a row, and thus you cannot even parse a D message to skip over\nit --- there's no way to know the length of the null-fields bitmap.\n\n> Is it not reasonable to run \"utility\" queries from inside SPI?\n\nSeems reasonable offhand, but I have no idea whether it really is or\nnot. If the context that the SPI procedure is executing from is a\nSELECT, as you illustrate above, then I could see where it would be\na bad idea to allow utility statements to execute before the SELECT\nfinishes. (Examples of no-nos: altering or dropping tables that the\nSELECT has already started using; VACUUM; perhaps other stuff.)\n\nBut either way it's definitely a backend bug: the SPI interface\nshould either handle utility statements or reject them cleanly.\n\n\t\t\tregards, tom lane\n", "msg_date": "Thu, 25 Feb 1999 09:54:18 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI " } ]
[ { "msg_contents": "OK, who deleted recipe.h and recipe.c without getting rid of\nthe references to them? Current backend sources do not build.\n\n\t\t\tregards, tom lane\n", "msg_date": "Thu, 25 Feb 1999 00:15:00 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": true, "msg_subject": "Current tree is busted" }, { "msg_contents": "> OK, who deleted recipe.h and recipe.c without getting rid of\n> the references to them? Current backend sources do not build.\n\nIt was me. FIx applied. Sorry.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Thu, 25 Feb 1999 12:26:52 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Current tree is busted" } ]
[ { "msg_contents": "What is the possibility of recording foreign keys (a partial foreign key\nimplementation) in the database such that the ODBC interface can correctly\nreport them to client applications? This would require an enhancement to\nPostgres and to the ODBC driver for Postgres. I have a problem with Access\n97 not working properly with forms that contain sub forms representing\nparent/child or master/detail tables. For example: orders and orderlines.\nI think this is caused by the lack of foreign key support. I am basing this\nconclusion on the SQL statement that Access sent through ODBC when looking\nup records for the sub form. The where clause of the SQL statement was\nbased on the primary key of the child or detail table. The where clause\nneeds to be based on the column that is linked to the parent or master\ntable. I already have indexes installed on both the primary key and the\nfield that links to the master/parent table.\n\nI am willing to explore a work around and/or help with these enhancements.\nThis issue, at least for the moment, appears to be a show stopper for me.\nAny Suggestions?\n\nThanks, Michael\n", "msg_date": "Thu, 25 Feb 1999 09:32:20 -0600", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "Foreign Keys" }, { "msg_contents": "\n\nMichael Davis wrote:\n\n> What is the possibility of recording foreign keys (a partial foreign key\n> implementation) in the database such that the ODBC interface can correctly\n> report them to client applications? This would require an enhancement to\n> Postgres and to the ODBC driver for Postgres. I have a problem with Access\n> 97 not working properly with forms that contain sub forms representing\n> parent/child or master/detail tables. For example: orders and orderlines.\n> I think this is caused by the lack of foreign key support. I am basing this\n>\n\nThe driver does support SQLForeignKeys. It does say it supports this function\nin SQLFunctions.\n\nHowever, the current implementation is based on the SPI/trigger stuff from the\npg_trigger table. If you don't have any trigger info, then SQLForeignKeys will\nnot report anything. This is most likely the problem.\n\nByron\n\n", "msg_date": "Thu, 25 Feb 1999 10:37:57 -0500", "msg_from": "Byron Nikolaidis <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [INTERFACES] Foreign Keys" } ]
[ { "msg_contents": "Excellent! I will start working on it. \n\n\t-----Original Message-----\n\tFrom:\tByron Nikolaidis [SMTP:[email protected]]\n\tSent:\tThursday, February 25, 1999 8:38 AM\n\tTo:\tMichael Davis\n\tCc:\t'[email protected]'; pgsql postgres;\[email protected]\n\tSubject:\tRe: [INTERFACES] Foreign Keys\n\n\n\n\tMichael Davis wrote:\n\n\t> What is the possibility of recording foreign keys (a partial\nforeign key\n\t> implementation) in the database such that the ODBC interface can\ncorrectly\n\t> report them to client applications? This would require an\nenhancement to\n\t> Postgres and to the ODBC driver for Postgres. I have a problem\nwith Access\n\t> 97 not working properly with forms that contain sub forms\nrepresenting\n\t> parent/child or master/detail tables. For example: orders and\norderlines.\n\t> I think this is caused by the lack of foreign key support. I am\nbasing this\n\t>\n\n\tThe driver does support SQLForeignKeys. It does say it supports\nthis function\n\tin SQLFunctions.\n\n\tHowever, the current implementation is based on the SPI/trigger\nstuff from the\n\tpg_trigger table. If you don't have any trigger info, then\nSQLForeignKeys will\n\tnot report anything. This is most likely the problem.\n\n\tByron\n", "msg_date": "Thu, 25 Feb 1999 10:45:41 -0600", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "RE: [INTERFACES] Foreign Keys" } ]
[ { "msg_contents": "Hi all\n\nI'm now building up a new server, Red Hat 5.2 with updates and the new\n2.2.2 kernel.\n\nI may have missed it, but do not recall seeing any thing about semaphores\nin the \"make menuconfig\", is it call some thing else? or is it N/A for\nLinux? or did I just not see it?\n\nI would like to build a kernel that can handle the future as I want to go\nfor a year of uptime (had 6months+ twice now.)\n\n\nThanks and have a great day\nTerry Mackintosh <[email protected]> http://www.terrym.com\nsysadmin/owner I'm excited about life! How about YOU!?\nProfessional Web Hosting and site design to include programming\nProudly powered by R H Linux 4.2, Apache 1.3.x, PHP 3.x, PostgreSQL 6.x\n-----------------------------------------------------------------------\nOnly if you know where you're going can you get there.\n\n", "msg_date": "Thu, 25 Feb 1999 12:39:10 -0500 (EST)", "msg_from": "Terry Mackintosh <[email protected]>", "msg_from_op": true, "msg_subject": "Linux: semaphores: How?" }, { "msg_contents": "Hi!\n\nOn Thu, 25 Feb 1999, Terry Mackintosh wrote:\n> I'm now building up a new server, Red Hat 5.2 with updates and the new\n> 2.2.2 kernel.\n> \n> I may have missed it, but do not recall seeing any thing about semaphores\n> in the \"make menuconfig\", is it call some thing else? or is it N/A for\n> Linux? or did I just not see it?\n\n It is \"System V IPC\".\n\nOleg.\n---- \n Oleg Broytmann National Research Surgery Centre http://sun.med.ru/~phd/\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Thu, 25 Feb 1999 20:42:28 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Linux: semaphores: How?" }, { "msg_contents": "The defaults for RH5.2 (from /usr/src/linux/include/linux/sem.h:\n\n#define SEMMNI 128 /* ? max # of semaphore identifiers\n*/\n#define SEMMSL 32 /* <= 512 max num of semaphores per\nid */\n#define SEMMNS (SEMMNI*SEMMSL) /* ? max # of semaphores in system\n*/\n#define SEMOPM 32 /* ~ 100 max num of ops per semop\ncall */\n#define SEMVMX 32767 /* semaphore maximum value */\n\nWhich means the default max semaphores is 128*32 = 4096, I don't\nthink you'll have a problem. :)\n\nDwD\n\n> -----Original Message-----\n> From: [email protected]\n> [mailto:[email protected]]On Behalf Of\n> Oleg Broytmann\n> Sent: Thursday, February 25, 1999 12:42 PM\n> To: Terry Mackintosh\n> Cc: PostgreSQL-development\n> Subject: Re: [HACKERS] Linux: semaphores: How?\n>\n>\n> Hi!\n>\n> On Thu, 25 Feb 1999, Terry Mackintosh wrote:\n> > I'm now building up a new server, Red Hat 5.2 with\n> updates and the new\n> > 2.2.2 kernel.\n> >\n> > I may have missed it, but do not recall seeing any\n> thing about semaphores\n> > in the \"make menuconfig\", is it call some thing else?\n> or is it N/A for\n> > Linux? or did I just not see it?\n>\n> It is \"System V IPC\".\n>\n> Oleg.\n> ----\n> Oleg Broytmann National Research Surgery Centre\n> http://sun.med.ru/~phd/\n> Programmers don't die,\n> they just GOSUB without RETURN.\n>\n>\n\n", "msg_date": "Thu, 25 Feb 1999 18:16:45 -0500", "msg_from": "\"Daryl W. Dunbar\" <[email protected]>", "msg_from_op": false, "msg_subject": "RE: [HACKERS] Linux: semaphores: How?" }, { "msg_contents": "I was trying to 'rename' a column\nduring an inherit. Since I could not\nfind a way to do this, I renamed\nthe column in a view, and then\nI inherited from the view.\n\nAnyway, I think being able\nto inherit a view is a bug,\nbesides confusing the \nbackend ... *evil grin*\n\nClark\n", "msg_date": "Sat, 27 Feb 1999 00:08:06 +0000", "msg_from": "Clark Evans <[email protected]>", "msg_from_op": false, "msg_subject": "Inherit View" } ]
[ { "msg_contents": "subscribe [email protected]\n\n", "msg_date": "Thu, 25 Feb 1999 23:27:14 +0400", "msg_from": "budin <[email protected]>", "msg_from_op": true, "msg_subject": "None" } ]
[ { "msg_contents": "\nShouldn't pg_dumpall also dump user privileges? Every time I've updated\n(just did it a little while ago) I have to go back into the databases and\ngrant all the users their privileges. Or am I doing something wrong?\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Thu, 25 Feb 1999 17:50:14 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": true, "msg_subject": "shouldn't pg_dumpall..." }, { "msg_contents": "Vince Vielhaber <[email protected]> writes:\n> Shouldn't pg_dumpall also dump user privileges?\n\nI believe pg_dumpall -z does it right. That's what I usually use.\n\nI suggested making -z the default a while ago, but no one has gotten\naround to doing something about it. (It'd be easy enough, but tedious,\nsince you'd need to invent a new switch for turning off privilege-\ndumping and then update all the documentation. Sigh.)\n\n\t\t\tregards, tom lane\n", "msg_date": "Thu, 25 Feb 1999 19:14:51 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] shouldn't pg_dumpall... " } ]
[ { "msg_contents": "\n\n>\"Gerald L. Gay\" <[email protected]> writes:\n>> I have seen what I concider to be a bug in either the libpq library\nor\n>> in the backend. To see the effects, first, install the execq() function\n>> from the SPI section of the Programmers Guide. Then do this in psql:\n>\n>> template1=> select execq('create user fred', 1);\n>> Backend sent D message without prior T\n>\n>That would be a backend bug, for sure. It's a violation of the FE/BE\n>protocol to send data row(s) without sending a row description first.\n>\n>> At this point psql will hang. I have a patch for libpq that fixes\nthis\n>> but I am not sure if this is the right place for it.\n>\n>I do not believe it is really possible to \"ignore\" this error inside\n>libpq. Without the initial T message you have no idea how many fields\n>are in a row, and thus you cannot even parse a D message to skip over\n>it --- there's no way to know the length of the null-fields bitmap.\n\nWhat I did in libpq was not to ignore the T message. Instead, if I get a T\nmessage, I remember it and ignore any intervening C messages until I either\nget a D to satisfy the T, or I get a C message of type \"SELECT.\" This\nhappens if the select returns no rows.\n\n>\n>> Is it not reasonable to run \"utility\" queries from inside SPI?\n>\n>Seems reasonable offhand, but I have no idea whether it really is or\n>not. If the context that the SPI procedure is executing from is a\n>SELECT, as you illustrate above, then I could see where it would be\n>a bad idea to allow utility statements to execute before the SELECT\n>finishes. (Examples of no-nos: altering or dropping tables that the\n>SELECT has already started using; VACUUM; perhaps other stuff.)\n\nWhat I have seen is: Any internal select/insert/update/delete calls that\nare performed inside the SPI function get suppressed but any utility\nfunctions get their status sent to the front end. So the T message is\ngenerated for the return type of the SPI function, then the C messages for\nany utility functions called, and then the D message for the actual return\nvalues.\n\nThe reason I discovered this in the first place was because the create user\n.... in group thing doesn't work yet. I am porting an application from\nSybase to Postgres and I need to ensure that the group stuff gets created.\nSo I wrote a create_user SPI function that creates the user and then updates\npg_group appropriately. This is when I saw this bug. Off-hand, I can't\nthink of anything else you might need this for. But I can envision\nsite-specific triggers on things like create user/drop user that may be tied\nto groups. Or maybe something like this: Deleting a group causes all the\nusers in that group to also get deleted. In that case you would have quite\na few \"drop user\" calls.\n\nAnother alternative might be something like Sybase. In Sybase, when you\ncreate a server-side procedure, you don't call it via SELECT. You just type\nit's name (they normally start with sp_, i.e., sp_spaceused - shows how much\nspace is available in a database). So the current paradigm in Postgres\nrequires the SELECT protocol. I personally like this because it works\nnicely for getting the return status. But it doesn't necessarily have to be\nSELECT. It could be something like EXEC or CALL or something.\n\n>\n>But either way it's definitely a backend bug: the SPI interface\n>should either handle utility statements or reject them cleanly.\n>\n> regards, tom lane\n\nI don't think it would be good to reject utility functions. This seems to\nme to fall into the category of \"what if I need to .....\"\n\nJerry\n\n\n", "msg_date": "Fri, 26 Feb 1999 09:59:49 +0900", "msg_from": "\"Gerald L. Gay\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] libpq and SPI " }, { "msg_contents": "Hi All,\n\tThis question of an XML based frontend/backend protocol\nhas come up once before in the last few months on this list (or is this\nthe same thread even?) I am guessing that the underlying motivation is\nthat many, if not most, users of Postgres want to connect the database\nto web-page user interfaces, and they would like the connection to be as\nseamless as possible. From that point of view the proposal seems\nreasonable, however I think that that point of view is limited, and that\ntying the frontend/backend protocol to a specific frontend technology\nwould be a design mistake. Here are\ntwo reasons:\n\n1) Frontend technology is notoriously short lived. Postgres -- or at\nleast Ingres -- predates the internet, and since the beginning of\nPostgres there have been at least three protocols for transmitting\nformatted data over the internet (gopher, html, and now XML). I would\nexpect that the basic design of Postgres is good for at least another 10\nyears, could the same be said about the design of XML?\n\n2) Although the majority of applications for Postgres are likely to use\nweb-based interfaces (or their successors), there are a significant\nnumber of applications that do not. My use for Postgres is as an indexed\ndata store for large quantities of signal data, a typical front end for\nme is a scripting language embedded in a numerical application. Fast and\nsimple are my primary requirements for a frontend/backend protocol. \n\nMore generally, I think that the strength of Postgres' design is that it\ncaters to a broad range of applications, and encourages experimentation\nwith the internals of the DBMS at a fundamental level. GIST, RTREEs, the\ngenetic optimizer, the myriad locking schemes, MVCC are all evidence of\nthis. If you need special support for XML, include it as a configurable\nmodule, don't replace an existing generic solution with one that tailors\nthe system to a specific application.\n\nBernie Frankpitt\n", "msg_date": "Sat, 27 Feb 1999 06:47:30 -0500", "msg_from": "[email protected]", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI" } ]
[ { "msg_contents": "Just an FYI:\n\nThe web page mail archive has not been archiving new messages since Sun Feb 21. \n\n-Ryan\n\nP.S. I like the new organization of the web pages.\n\n", "msg_date": "Fri, 26 Feb 1999 16:06:19 -0700 (MST)", "msg_from": "Ryan Bradetich <[email protected]>", "msg_from_op": true, "msg_subject": "Web page mail archive" } ]
[ { "msg_contents": "\nIgnore...\n-- \nMarc G. Fournier [email protected]\nSystems Administrator @ hub.org \nscrappy@{postgresql|isc}.org ICQ#7615664\n", "msg_date": "28 Feb 1999 07:49:59 GMT", "msg_from": "[email protected] (Marc G. Fournier)", "msg_from_op": true, "msg_subject": "Test 2...news2mail ..." } ]
[ { "msg_contents": "hello everybody,\ni want to know where i can find softwares that reveal passwords,\nespecially tricks that can help me to enter the web based mail like the\nhotbot, hotmail, etc.\nthank you\[email protected]\n\n\n", "msg_date": "Sun, 28 Feb 1999 18:23:32 +0200", "msg_from": "dana lapidot <[email protected]>", "msg_from_op": true, "msg_subject": "hello" } ]
[ { "msg_contents": "\n============================================================================\n POSTGRESQL BUG REPORT TEMPLATE\n============================================================================\n\n\nYour name\t\t: Monica Kroger\nYour email address\t: [email protected]\n\nCategory\t\t: runtime: back-end: SQL\nSeverity\t\t: non-critical\n\nSummary: Error with date_part when checking month\n\nSystem Configuration\n--------------------\n Operating System : RedHat 5.2\n\n PostgreSQL version : 6.3\n\n Compiler used : it came precompiled with RH5.2\n\nHardware:\n---------\nLinux localhost.localdomain 2.0.36 #7 Fri Feb 26 21:42:39 MST 1999 i586 unknown\nP133, 64 Megs of RAM\n\n\nVersions of other tools:\n------------------------\n\n\n--------------------------------------------------------------------------\n\nProblem Description:\n--------------------\nWhen using the date_part function it seems to be off by a day.\nIt excludes the first of this month and includes the first of next month in the query.\n\n\n--------------------------------------------------------------------------\n\nTest Case:\n----------\nGiven a table(ice) with a date(eventstart) in it when I use a select as follows:\nselect * from ice where date_part('month', \"eventstart\")='6';\nit will return all of the dates in June except if they're on 06-01-1999\nIt will also return all of the dates on July 1.\n\n--------------------------------------------------------------------------\n\nSolution:\n---------\n\n\n--------------------------------------------------------------------------\n\n", "msg_date": "Sun, 28 Feb 1999 11:38:41 -0500 (EST)", "msg_from": "Unprivileged user <nobody>", "msg_from_op": true, "msg_subject": "Port Bug Report: Error with date_part when checking month" }, { "msg_contents": "> PostgreSQL version : 6.3\n> Problem Description:\n> --------------------\n> When using the date_part function it seems to be off by a day.\n> It excludes the first of this month and includes the first of next \n> month in the query.\n\npostgres=> select date_part('month', '6-1-1999'::datetime);\ndate_part\n---------\n 6\n(1 row)\n\nIt appears to be fixed in more recent versions (though I haven't\nconfirmed that I can reproduce your symptom on v6.3).\n\nv6.4.2 is the current version, and has significant improvements in all\nareas.\n\n - Tom\n", "msg_date": "Wed, 03 Mar 1999 03:26:38 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [PORTS] Port Bug Report: Error with date_part when checking month" }, { "msg_contents": "> eventstart|date_part\n> ----------+---------\n> 06-15-1999| 6\n> The data was entered in this same time zone. It was entered as\n> 'mm-dd-yyyy' from a PHP3 form using a standard SQL insert.\n\nAh, it is the date data type. I now recall that I fixed some time zone\nconversion problems in this data type. Postgres v6.4.2, plus some\npatches on the web site, will fix this problem, which was first\nidentified at the end of December. The upcoming v6.4.3 will have the\nfixes already applied.\n\nIf you don't want to install and patch a new version of Postgres, try\nusing the datetime data type instead.\n\nSorry for the problems...\n\n - Tom\n", "msg_date": "Wed, 03 Mar 1999 15:28:10 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [PORTS] Port Bug Report: Error with date_part when checking month" } ]
[ { "msg_contents": "ignore...\n-- \nMarc G. Fournier [email protected]\nSystems Administrator, Acadia University\n\n \"These are my opinions, which are not necessarily shared by my employer\"\n", "msg_date": "28 Feb 1999 21:34:13 GMT", "msg_from": "[email protected] (Marc G. Fournier)", "msg_from_op": true, "msg_subject": "Testing news2mail gateway ..." } ]
[ { "msg_contents": "Hi,\n\n How did XML get into this discussion? What I was talking about was a\nbug where, utility queries in a SPI procedure send command complete messages\nbetween data type messages (T) and data value messages (D) which confuses\nlibpq.\n\nJerry\n\n-----Original Message-----\nFrom: [email protected] <[email protected]>\nTo: Gerald L. Gay <[email protected]>; [email protected]\n<[email protected]>\nDate: Sunday, February 28, 1999 1:54 AM\nSubject: Re: [HACKERS] libpq and SPI\n\n\n>Hi All,\n> This question of an XML based frontend/backend protocol\n>has come up once before in the last few months on this list (or is this\n>the same thread even?) I am guessing that the underlying motivation is\n>that many, if not most, users of Postgres want to connect the database\n>to web-page user interfaces, and they would like the connection to be as\n>seamless as possible. From that point of view the proposal seems\n>reasonable, however I think that that point of view is limited, and that\n>tying the frontend/backend protocol to a specific frontend technology\n>would be a design mistake. Here are\n>two reasons:\n>\n>1) Frontend technology is notoriously short lived. Postgres -- or at\n>least Ingres -- predates the internet, and since the beginning of\n>Postgres there have been at least three protocols for transmitting\n>formatted data over the internet (gopher, html, and now XML). I would\n>expect that the basic design of Postgres is good for at least another 10\n>years, could the same be said about the design of XML?\n>\n>2) Although the majority of applications for Postgres are likely to use\n>web-based interfaces (or their successors), there are a significant\n>number of applications that do not. My use for Postgres is as an indexed\n>data store for large quantities of signal data, a typical front end for\n>me is a scripting language embedded in a numerical application. Fast and\n>simple are my primary requirements for a frontend/backend protocol.\n>\n>More generally, I think that the strength of Postgres' design is that it\n>caters to a broad range of applications, and encourages experimentation\n>with the internals of the DBMS at a fundamental level. GIST, RTREEs, the\n>genetic optimizer, the myriad locking schemes, MVCC are all evidence of\n>this. If you need special support for XML, include it as a configurable\n>module, don't replace an existing generic solution with one that tailors\n>the system to a specific application.\n>\n>Bernie Frankpitt\n\n", "msg_date": "Mon, 1 Mar 1999 08:56:29 +0900", "msg_from": "\"Gerald L. Gay\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] libpq and SPI" } ]
[ { "msg_contents": "\nMorning all...\n\n\tI've had a couple of queries over the past few days about\n'cross-list' posting, where you answer email that is CC'd to a list you\naren't subscribed to.\n\n\tIt is recommended that everyone take a minute and subscribe to\[email protected] to avoid this happening. It is a\nnon-functional list, no mail will ever go out on it...its purpose is\npurely to ensure that you can post to a list without being subscribed to\nit.\n\n\tFor anyone wishing to make use of the newsgroups, for a posting to\nbe able to go from the newsgroup to the list, you *have* to be subscribed\nto pgsql-loopback *at the very least*, or the posting will be rejected at\nthe gateway.\n\n\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Sun, 28 Feb 1999 20:46:43 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "[email protected]" } ]
[ { "msg_contents": "\nMorning all...\n\n\tI've had a couple of queries over the past few days about\n'cross-list' posting, where you answer email that is CC'd to a list you\naren't subscribed to.\n\n\tIt is recommended that everyone take a minute and subscribe to\[email protected] to avoid this happening. It is a\nnon-functional list, no mail will ever go out on it...its purpose is\npurely to ensure that you can post to a list without being subscribed to\nit.\n\n\tFor anyone wishing to make use of the newsgroups, for a posting to\nbe able to go from the newsgroup to the list, you *have* to be subscribed\nto pgsql-loopback *at the very least*, or the posting will be rejected at\nthe gateway.\n\n\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n\n", "msg_date": "Sun, 28 Feb 1999 20:50:52 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "IMPORTANT NOTICE: [email protected]" } ]
[ { "msg_contents": "\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n---------- Forwarded message ----------\nDate: Sun, 28 Feb 1999 20:50:52 -0400 (AST)\nFrom: The Hermit Hacker <[email protected]>\nTo: [email protected]\nCc: [email protected]\nSubject: IMPORTANT NOTICE: [email protected]\n\n\nMorning all...\n\n\tI've had a couple of queries over the past few days about\n'cross-list' posting, where you answer email that is CC'd to a list you\naren't subscribed to.\n\n\tIt is recommended that everyone take a minute and subscribe to\[email protected] to avoid this happening. It is a\nnon-functional list, no mail will ever go out on it...its purpose is\npurely to ensure that you can post to a list without being subscribed to\nit.\n\n\tFor anyone wishing to make use of the newsgroups, for a posting to\nbe able to go from the newsgroup to the list, you *have* to be subscribed\nto pgsql-loophole *at the very least*, or the posting will be rejected at\nthe gateway.\n\n\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n\n\n", "msg_date": "Sun, 28 Feb 1999 21:20:07 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "IMPORTANT NOTICE II: [email protected] " } ]
[ { "msg_contents": "\nhad a bug in configuration,causing duplicates to be sent out...think its\nfixed, but if you see two fo these, let me know (I have anti-duplicate\nfilters in place)...\n\n\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 1 Mar 1999 03:11:06 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "news2mail gateway..." } ]
[ { "msg_contents": "Hi all,\n\nI just finished with a little password manager workaround for Postgres\nand I'd like to tell you to use it. Comments are welcome! I hope I could\ndo something with this user authentication and password altering misery.\n\nHere is the URL where you can download it from:\nhttp://westerlike.bay.u-szeged.hu/members/waiter/pwman/index.html\n\nThanks in advance and regards,\nPeter Blazso\n\n\n\n", "msg_date": "Mon, 01 Mar 1999 15:34:22 +0100", "msg_from": "Peter Blazso <[email protected]>", "msg_from_op": true, "msg_subject": "A small Password Manager server for PGSQL" }, { "msg_contents": "How will the new numeric type appear in Perl? Will it be a string or\nwhat?\n\n", "msg_date": "Mon, 1 Mar 1999 20:14:45 +0100 (CET)", "msg_from": "Kaare Rasmussen <[email protected]>", "msg_from_op": false, "msg_subject": "NUMERIC and Perl" }, { "msg_contents": ">\n> How will the new numeric type appear in Perl? Will it be a string or\n> what?\n\n Should be a string - or is Perl capable of handling numbers\n with hundreds of digits before and after the decimal point\n (ALL SIGNIFICANT)?\n\n AFAIK, only tools like bc(1) or when using some arbitrary\n precision packages from inside an application (where numbers\n are usually handled as strings for I/O) can operate correctly\n on such precise things.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Tue, 2 Mar 1999 21:47:19 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] NUMERIC and Perl" }, { "msg_contents": "> Should be a string - or is Perl capable of handling numbers\n> with hundreds of digits before and after the decimal point\n> (ALL SIGNIFICANT)?\n\ngolem> perl -e '{printf(\"%30.18f\\n\", 1234567890.2345678901234567890)}'\n 1234567890.234567880630493164\n\nLooks like not, unless I'm forgetting a feature...\n\n - Tom\n", "msg_date": "Wed, 03 Mar 1999 04:14:25 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] NUMERIC and Perl" }, { "msg_contents": "> Looks like not, unless I'm forgetting a feature...\n\nThat's why I asked. So I hope it will be compatible with Perl's\nMath-BigInteger module.\n\n", "msg_date": "Wed, 3 Mar 1999 19:33:55 +0100 (CET)", "msg_from": "Kaare Rasmussen <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] NUMERIC and Perl" }, { "msg_contents": ">\n> > Looks like not, unless I'm forgetting a feature...\n>\n> That's why I asked. So I hope it will be compatible with Perl's\n> Math-BigInteger module.\n\n NUMERIC isn't INTEGER data!\n\n Well - you can define an attribute as NUMERIC(33,0) and all\n values assigned to that column will be rounded to the closest\n integer (an integer that can hold values up to 999\n quintillions exactly). But as soon as you do calculations\n with other values that have fractions, there will a decimal\n point and digits after that appear in the result. Don't know\n how Perl might respond to it. Maybe you need to round the\n result explicitly on calculations.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 3 Mar 1999 20:17:08 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] NUMERIC and Perl" }, { "msg_contents": "> > > Looks like not, unless I'm forgetting a feature...\n> > That's why I asked. So I hope it will be compatible with Perl's\n> > Math-BigInteger module.\n> NUMERIC isn't INTEGER data!\n\nOh, btw numeric() seems to work great on my Linux/libc5 box. The last\ntime I reported that it had troubles...\n\n - Tom\n", "msg_date": "Thu, 04 Mar 1999 02:33:01 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] NUMERIC and Perl" } ]
[ { "msg_contents": "\n============================================================================\n POSTGRESQL BUG REPORT TEMPLATE\n============================================================================\n\n\nYour name\t\t:\tGiuseppe Tanzilli\nYour email address\t: [email protected]\n\n\nSystem Configuration\n---------------------\n Architecture (example: Intel Pentium) \t: Intel Pentium II\n\n Operating System (example: Linux 2.0.26 ELF) \t: Linux 2.2.1 ELF libc6\n\n PostgreSQL version (example: PostgreSQL-6.4.2) : PostgreSQL-6.4.2\n\n Compiler used (example: gcc 2.8.0)\t\t: gcc 2.7.2.3\n\n\nPlease enter a FULL description of your problem:\n------------------------------------------------\nINSERT is allowed over VIEWs, the row are not visible in the tables,\nbut in the tables file the row is present. \n\n\n\n\nPlease describe a way to repeat the problem. Please try to provide a\nconcise reproducible example, if at all possible: \n----------------------------------------------------------------------\nCREATE TABLE t1 (i1 int4, i2 int4);\nCREATE TABLE t2 (i3 int4, i4 int4);\nINSERT INTO t1 (i1,i2) values(1, 1);\nINSERT INTO t1 (i1,i2) values(2, 1);\nINSERT INTO t1 (i1,i2) values(3, 1);\nINSERT INTO t1 (i1,i2) values(4, 1);\nINSERT INTO t1 (i1,i2) values(5, 1);\nINSERT INTO t2 (i3,i4) values(1, 100);\nINSERT INTO t2 (i3,i4) values(2, 100);\nINSERT INTO t2 (i3,i4) values(3, 100);\nINSERT INTO t2 (i3,i4) values(4, 100);\nCREATE VIEW v1 as SELECT i1,i2,i4 from t1,t2 where i1=i3;\nSELECT * from v1;\nINSERT INTO v1 (i1,i2,i4) values (10,10,200);\nSELECT * from v1;\nSELECT * from t1;\nSELECT * from t2;\n\nDROP VIEW v1;\nDROP TABLE t1;\nDROP TABLE t2;\n\n\n\n\n\nIf you know how this problem might be fixed, list the solution below:\n---------------------------------------------------------------------\n\n\n", "msg_date": "Mon, 1 Mar 1999 16:23:42 +0100", "msg_from": "Giuseppe Tanzilli <[email protected]>", "msg_from_op": true, "msg_subject": "Bug report" } ]
[ { "msg_contents": "Hi all,\n\nSorry for setting up a bad link and thank you for informing me! (I\nwanted\nto be very quick but I messed up...) Ross has finally found the right\nlink to PwMan. Thanks Ross!\n\nThe URL is the same for the corrected page:\nhttp://westerlike.bay.u-szeged.hu/members/waiter/pwman/index.html\n\nRegards,\nPerter B.\n", "msg_date": "Mon, 01 Mar 1999 22:55:00 +0100", "msg_from": "Peter Blazso <[email protected]>", "msg_from_op": true, "msg_subject": "Re: pwman is now more available ... :-)" } ]
[ { "msg_contents": "\nI just built a v6.4.3beta2 tar file, hoping to try and get it released\nRSN, and am doing a test built on my machine...it fails at datetime.c,\nline 109, with a reference to day_tab[] ... \n\nday_tab is 'declared' at line 31 of that same file, but that section of\ncode is #ifdef 0'd out ...\n\nLooks like something that Thomas added:\n\nhub# cvs diff -r1.25.2.1 -r1.25.2.2 datetime.c | less\nIndex: datetime.c\n===================================================================\nRCS file: /usr/local/cvsroot/pgsql/src/backend/utils/adt/datetime.c,v\nretrieving revision 1.25.2.1\nretrieving revision 1.25.2.2\ndiff -r1.25.2.1 -r1.25.2.2\n10c10\n< * $Header:\n/usr/local/cvsroot/pgsql/src/backend/utils/adt/datetime.c,v \n1.25.2.1 1998/12/31 16:34:47 thomas Exp $\n---\n> * $Header:\n/usr/local/cvsroot/pgsql/src/backend/utils/adt/datetime.c,v \n1.25.2.2 1999/02/13 05:59:34 thomas Exp $\n30c30\n< \n---\n> #if 0\n35a36\n> #endif\n101a103\n> #if 0\n105a108\n> #endif\n\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Tue, 2 Mar 1999 00:47:17 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "datetime.c in v6.4.3beta2 ..." }, { "msg_contents": "> I just built a v6.4.3beta2 tar file, hoping to try and get it released\n> RSN, and am doing a test built on my machine...it fails at datetime.c,\n> line 109, with a reference to day_tab[] ...\n> Looks like something that Thomas added:\n\nYeah, Tom Lane already mentioned it. My 'net access was down this\nmorning; will look at it this evening. Sorry...\n\n - Tom\n", "msg_date": "Wed, 03 Mar 1999 04:36:22 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] datetime.c in v6.4.3beta2 ..." }, { "msg_contents": "On Wed, 3 Mar 1999, Thomas G. Lockhart wrote:\n\n> > I just built a v6.4.3beta2 tar file, hoping to try and get it released\n> > RSN, and am doing a test built on my machine...it fails at datetime.c,\n> > line 109, with a reference to day_tab[] ...\n> > Looks like something that Thomas added:\n> \n> Yeah, Tom Lane already mentioned it. My 'net access was down this\n> morning; will look at it this evening. Sorry...\n\nS'alright, must have missed Tom's :(\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 3 Mar 1999 00:50:17 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] datetime.c in v6.4.3beta2 ..." }, { "msg_contents": "> > > I just built a v6.4.3beta2 tar file, hoping to try and get it \n> > > released RSN, and am doing a test built on my machine...it fails \n> > > at datetime.c, line 109, with a reference to day_tab[] ...\n> > > Looks like something that Thomas added:\n\nI had patched my tree on postgresql.org, but had neglected to commit the\npatch for dt.h. afaik it is now OK, but I was not able to get a\nregression test since REL6_4 hangs while building on my system due to\ntcl scripting problems.\n\nI can work around that if necessary, but we *really* should have a\nsystem that manages to build cleanly...\n\nSorry for the dt.h omission.\n\n - Tom\n", "msg_date": "Wed, 03 Mar 1999 06:38:57 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] datetime.c in v6.4.3beta2 ..." }, { "msg_contents": "On Wed, 3 Mar 1999, Thomas G. Lockhart wrote:\n\n> > > > I just built a v6.4.3beta2 tar file, hoping to try and get it \n> > > > released RSN, and am doing a test built on my machine...it fails \n> > > > at datetime.c, line 109, with a reference to day_tab[] ...\n> > > > Looks like something that Thomas added:\n> \n> I had patched my tree on postgresql.org, but had neglected to commit the\n> patch for dt.h. afaik it is now OK, but I was not able to get a\n> regression test since REL6_4 hangs while building on my system due to\n> tcl scripting problems.\n> \n> I can work around that if necessary, but we *really* should have a\n> system that manages to build cleanly...\n\nWill put some time into this later this evening...thanks for fixign the\ndatetime.c stuff :)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 3 Mar 1999 03:09:36 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] datetime.c in v6.4.3beta2 ..." } ]
[ { "msg_contents": "Hello all,\n\n\"Hiroshi Inoue\" <[email protected]> writes:\n> \"Tom Lane\" <[email protected]> writes\n> > \"Hiroshi Inoue\" <[email protected]> writes:\n> > > AFAIC the relation between objects is not copied correctly\n> > > by copyObject() (i.e the same pointers to an object are copied\n> > > to different pointers by copyObject()).\n> >\n> > True, but it seems irrelevant to me --- as Jan Wieck was just pointing\n> > out, no code should ever depend on pointer-equality in parse trees or\n> > plan trees anyway.\n> >\n> > > There is a way to maintain the list of (old,new) pairs during\n> > > copyObject() operations.\n> >\n> > I think we'd be better off fixing any places that mistakenly assume\n> > pointer compare is sufficient. You didn't say which version you were\n> > testing,\n>\n> My environment is v6.4.2.\n> OK,I would test my cases again after the release of 6.5-BETA(v6.4.3?).\n>\n\nThis time,I tested the following 2 cases under CURRENT(99/02/23)\nenvironment and the result was same.\n\nThose bugs are caused by copyObject().\nThe removal of pointer comparison solves those bugs ?\nI don't think so.\nI think copyObject() should be more reliable.\n\nI made a patch on trial (see attached file) .\nAfter applying it,both return proper results.\nThis patch changes the implementaion of copyObject() entirely.\n\nComments ?\n\n\n[Case -1]\ncreate function subsel() returns int4\nas\n'\ndeclare\nk int4;\nbegin\nSELECT 1 AS one into k WHERE 1 IN (SELECT 1);\nreturn k;\nend;\n'\nlanguage 'plpgsql';\nselect subsel();\n\npqReadData() -- backend closed the channel unexpectedly.\n This probably means the backend terminated abnormally before or\nwhile pr\nocessing the request.\nWe have lost the connection to the backend, so further processing is\nimpossible.\n Terminating.\n\n[Case-2]\n\ncreate table grp (key int4);\ninsert into grp values(2);\ninsert into grp values(2);\ncreate function grup() returns int4\nas\n'\ndeclare\nk int4;\nbegin\nselect key into k from grp\ngroup by key\nhaving count(*) > 1;\nreturn k;\nend;\n'\nlanguage 'plpgsql';\nselect grup();\n\nThe result is\n\ngrup\n----\n\n(1 row)\n\n\nThanks in advance.\n\nHiroshi Inoue\[email protected]", "msg_date": "Tue, 2 Mar 1999 18:15:40 +0900", "msg_from": "\"Hiroshi Inoue\" <[email protected]>", "msg_from_op": true, "msg_subject": "copyObject() ? again " }, { "msg_contents": "I'm still not real happy about Hiroshi's proposed reimplementation of\ncopyObject. It would make copyObject much slower and more memory-\nhungry, which I think is a bad thing -- especially since we may use\ncopyObject more than now if we redesign memory allocation as was\ndiscussed a few weeks ago.\n\nI looked at the specific problem he's reporting (coredump involving\na SELECT inside a plpgsql function), and I believe I understand it;\nit's a pretty straightforward order-of-operations bug in copyfuncs.c.\nSpecifically, CopyPlanFields() tries to construct the new plan node's\nlist of subplans, but you can't do that until the node-type-specific\nfields have all been set, since they may contain references to subplans.\nAs it stands, a copied plan will have a subPlan list that is missing\nsome of its subplans.\n\nA simple localized fix would be to rearrange operations inside\ncopyfuncs.c so that the SS_pull_subplan operation is not done until\nthe end of the node-type-specific routines for copying plan nodes.\n\nMore generally, however, I think that the very existence of the subPlan\nlist is a bug. It violates the notion that \"pointer equality shouldn't\nmatter\" because the subPlan list *must* contain pointers to the actual\nmember objects of a plan node's other substructure.\n\nMoreover the subPlan list isn't even buying us much --- it wouldn't be\nthat expensive to scan the substructure directly for plan-type nodes\nwhen necessary, exactly as SS_pull_subplan does. Arguably this would\ntake less time than is expended on memory allocation to build the\nexplicit subPlan list. So I suggest getting rid of the subPlan list\nentirely, rather than hacking on copyfuncs.\n\nPlans also have an \"initPlan\" list that seems to have the same kind of\nproblem. I don't really understand the difference between initPlans\nand subPlans --- plannodes.h says\n List *initPlan; /* Init Plan nodes (un-correlated expr\n * subselects) */\n List *subPlan; /* Other SubPlan nodes */\nwhich doesn't convey a whole lot to my mind. Does anyone understand\nwhy this distinction is made? Is it really necessary?\n\n\t\t\tregards, tom lane\n", "msg_date": "Tue, 02 Mar 1999 11:52:18 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] copyObject() ? again " }, { "msg_contents": "Tom Lane wrote:\n> \n> More generally, however, I think that the very existence of the subPlan\n> list is a bug. It violates the notion that \"pointer equality shouldn't\n> matter\" because the subPlan list *must* contain pointers to the actual\n> member objects of a plan node's other substructure.\n> \n> Moreover the subPlan list isn't even buying us much --- it wouldn't be\n> that expensive to scan the substructure directly for plan-type nodes\n> when necessary, exactly as SS_pull_subplan does. Arguably this would\n> take less time than is expended on memory allocation to build the\n> explicit subPlan list. So I suggest getting rid of the subPlan list\n> entirely, rather than hacking on copyfuncs.\n\nFirst, I apologize for any inconveniences in my implementation\nof subselect. I agreed that current handling of subPlan lists\nis bad, but parent plan need in this list for some initializations\nand SUBPLAN_EXPR expressions inside parent plan also need in \ncorresponding subPlan nodes (to execute them).\n\nWell, this could be good solution:\n\n1. Implement TopPlan node (upmost plan, we told about this and\n seems that Jan wanted implement it).\n2. Construct list of _all_ subPlan inside planner and\n put this list into TopPlan node.\n3. Put into Plan->subPlan list of _indices_ of subPlans used\n in this Plan (indices in TopPlan->subPlan list).\n4. Use indices in SUBPLAN_EXPR, istead of subPlan pointers.\n5. Add to EState and ExprContext pointer to TopPlan->subPlan\n list (to enable index --> node conversion in\n ExecSubPlan() & ExecInitSubPlan()).\n \nUnfortunately, I missed this solution year ago -:(\nAnd more of that, I have no time to implement this now, sorry.\n\n> Plans also have an \"initPlan\" list that seems to have the same kind of\n> problem. I don't really understand the difference between initPlans\n\nNo problems here. As you see this list is just copied in copyfuncs\nand this is ok.\n\n> and subPlans --- plannodes.h says\n> List *initPlan; /* Init Plan nodes (un-correlated expr\n> * subselects) */\n> List *subPlan; /* Other SubPlan nodes */\n> which doesn't convey a whole lot to my mind. Does anyone understand\n> why this distinction is made? Is it really necessary?\n\nUncorrelated expression subqueries are processed differently\nfrom all other subqueries.\n\nVadim\n", "msg_date": "Wed, 03 Mar 1999 01:16:23 +0700", "msg_from": "Vadim Mikheev <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] copyObject() ? again" }, { "msg_contents": "Hello all,\n\n> -----Original Message-----\n> From: Tom Lane [mailto:[email protected]]\n> Sent: Wednesday, March 03, 1999 1:52 AM\n> To: Hiroshi Inoue\n> Cc: pgsql-hackers\n> Subject: Re: [HACKERS] copyObject() ? again \n> \n> \n> I'm still not real happy about Hiroshi's proposed reimplementation of\n> copyObject. It would make copyObject much slower and more memory-\n> hungry, which I think is a bad thing -- \n\nYes,it's a big defect of my implementation.\nBut isn't it simple and reliable ?\n\nCurrently,limited persons could implement copyObject() or \ncould rearrange structures which cause copyObject() bugs.\nAnd it seems that we should use copyObject() carefully.\nI don't know whether we are allowed to use copyObject() in \nvarious phase (parser,optimizer etc) without limitation.\n\nOK,there's a possibility to reduce the number of (old,new) pairs \nmaintained during copyObject() operations. \nIn fact,if only SubPlan and Aggref type nodes are maintained,\nthe cases I reported return proper results.\nIt may be risky but would improve the performance of my \nimplementation.\n\nComments ? \n\nAs to the 2nd case I reported,it was probably introduced by my \npatch. The patch was made to solve other Aggregate bugs.\nget_agg_tlist_references() is used to reconstruct aggs member \nnode of Agg type nodes as SS_pull_subPlan() does in CopyPlan\nFields(). (The name of function was set_agg_tlist_references() \noriginally. Probably Bruce changed it).\nWe may be able to patch by following related code closely.\nBut it seems strange that copying objects requires such \na complicated procedure. \n\nThanks in advance.\n\nHiroshi Inoue\[email protected]\n", "msg_date": "Wed, 3 Mar 1999 11:53:46 +0900", "msg_from": "\"Hiroshi Inoue\" <[email protected]>", "msg_from_op": true, "msg_subject": "RE: [HACKERS] copyObject() ? again " }, { "msg_contents": "\nJust sending this over to you in case it is helpful in the future.\n\n> Tom Lane wrote:\n> > \n> > More generally, however, I think that the very existence of the subPlan\n> > list is a bug. It violates the notion that \"pointer equality shouldn't\n> > matter\" because the subPlan list *must* contain pointers to the actual\n> > member objects of a plan node's other substructure.\n> > \n> > Moreover the subPlan list isn't even buying us much --- it wouldn't be\n> > that expensive to scan the substructure directly for plan-type nodes\n> > when necessary, exactly as SS_pull_subplan does. Arguably this would\n> > take less time than is expended on memory allocation to build the\n> > explicit subPlan list. So I suggest getting rid of the subPlan list\n> > entirely, rather than hacking on copyfuncs.\n> \n> First, I apologize for any inconveniences in my implementation\n> of subselect. I agreed that current handling of subPlan lists\n> is bad, but parent plan need in this list for some initializations\n> and SUBPLAN_EXPR expressions inside parent plan also need in \n> corresponding subPlan nodes (to execute them).\n> \n> Well, this could be good solution:\n> \n> 1. Implement TopPlan node (upmost plan, we told about this and\n> seems that Jan wanted implement it).\n> 2. Construct list of _all_ subPlan inside planner and\n> put this list into TopPlan node.\n> 3. Put into Plan->subPlan list of _indices_ of subPlans used\n> in this Plan (indices in TopPlan->subPlan list).\n> 4. Use indices in SUBPLAN_EXPR, istead of subPlan pointers.\n> 5. Add to EState and ExprContext pointer to TopPlan->subPlan\n> list (to enable index --> node conversion in\n> ExecSubPlan() & ExecInitSubPlan()).\n> \n> Unfortunately, I missed this solution year ago -:(\n> And more of that, I have no time to implement this now, sorry.\n> \n> > Plans also have an \"initPlan\" list that seems to have the same kind of\n> > problem. I don't really understand the difference between initPlans\n> \n> No problems here. As you see this list is just copied in copyfuncs\n> and this is ok.\n> \n> > and subPlans --- plannodes.h says\n> > List *initPlan; /* Init Plan nodes (un-correlated expr\n> > * subselects) */\n> > List *subPlan; /* Other SubPlan nodes */\n> > which doesn't convey a whole lot to my mind. Does anyone understand\n> > why this distinction is made? Is it really necessary?\n> \n> Uncorrelated expression subqueries are processed differently\n> from all other subqueries.\n> \n> Vadim\n> \n> \n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 6 Jul 1999 22:03:19 -0400 (EDT)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] copyObject() ? again" } ]
[ { "msg_contents": "I'm interested in Porting Postgres to Windows NT.\n\nIs there anyone, who has done this before. \n\nI read about CygWin, does anyone have the right configuration files ?!\n\nOr is there anyone who ported it with MS-Visual C++", "msg_date": "Tue, 2 Mar 1999 14:56:42 +0100", "msg_from": "\"Stefan Diestelmann\" <[email protected]>", "msg_from_op": true, "msg_subject": "Postgres under Windows NT" } ]
[ { "msg_contents": "I just noticed that the palloc.h problem still exists. That is we do not\ninstall utils/mcxt.h but is included by palloc.h.\n\nDid we find a solution for this yet?\n\nMichael\n-- \nMichael Meskes | Go SF 49ers!\nTh.-Heuss-Str. 61, D-41812 Erkelenz | Go Rhein Fire!\nTel.: (+49) 2431/72651 | Use Debian GNU/Linux!\nEmail: [email protected] | Use PostgreSQL!\n", "msg_date": "Tue, 2 Mar 1999 20:11:06 +0100", "msg_from": "Michael Meskes <[email protected]>", "msg_from_op": true, "msg_subject": "palloc.h again" }, { "msg_contents": "Michael Meskes <[email protected]> writes:\n> I just noticed that the palloc.h problem still exists. That is we do not\n> install utils/mcxt.h but is included by palloc.h.\n\nI was hoping that Jan would try to reduce the number of include files\nthat are now needed by palloc(). But definitely something will have\nto be done about this before we release.\n\n\t\t\tregards, tom lane\n", "msg_date": "Wed, 03 Mar 1999 10:14:48 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] palloc.h again " }, { "msg_contents": ">\n> Michael Meskes <[email protected]> writes:\n> > I just noticed that the palloc.h problem still exists. That is we do not\n> > install utils/mcxt.h but is included by palloc.h.\n>\n> I was hoping that Jan would try to reduce the number of include files\n> that are now needed by palloc(). But definitely something will have\n> to be done about this before we release.\n\n Hmmm - how should I reduce the number of include files? All\n the palloc stuff is macroized now, so using palloc() results\n in MemoryContextAlloc() calls. Reduces one brain dead\n function call (the old versions of palloc() etc. simply did\n that - wasting one stack frame).\n\n Thus, we definitely need mcxt.h where we need palloc() and\n that's why I've put the include of mcxt.h into palloc.h.\n\n What I'm in doubt about is, why (if) we don't want to install\n mcxt.h? It is required (since my changes) by every user\n built backend-loadable module. So it should be installed.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 3 Mar 1999 16:44:18 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] palloc.h again" }, { "msg_contents": "[email protected] (Jan Wieck) writes:\n> What I'm in doubt about is, why (if) we don't want to install\n> mcxt.h? It is required (since my changes) by every user\n> built backend-loadable module. So it should be installed.\n\nIf it's only mcxt.h that is newly required, we may as well just export\nit and be done with it. My recollection of the original report was\nthat *several* formerly backend-private headers needed to be added to\nthe export set, and I thought that wasn't a very good idea...\n\n\t\t\tregards, tom lane\n", "msg_date": "Wed, 03 Mar 1999 18:31:52 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] palloc.h again " }, { "msg_contents": "On Wed, 3 Mar 1999, Tom Lane wrote:\n\n> [email protected] (Jan Wieck) writes:\n> > What I'm in doubt about is, why (if) we don't want to install\n> > mcxt.h? It is required (since my changes) by every user\n> > built backend-loadable module. So it should be installed.\n> \n> If it's only mcxt.h that is newly required, we may as well just export\n> it and be done with it. My recollection of the original report was\n> that *several* formerly backend-private headers needed to be added to\n> the export set, and I thought that wasn't a very good idea...\n> \n\nIt's been a while since I switched to CVS but I remember other .h\ndependencies with each .h file I added starting with palloc.h. I ended up\ngetting tired of it and just copied the contents of the include dir over\nrather than fight with it. So unless somethings changed there are more\nheaders needed that palloc.h and mcxt.h Sorry I don't have better\ndetails as to which ones it needed. I'll be doing a fresh install this\nweekend, if it'd help I'll make a list. \n\n->->->->->->->->->->->->->->->->->->---<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<\nJames Thompson 138 Cardwell Hall Manhattan, Ks 66506 785-532-0561 \nKansas State University Department of Mathematics\n->->->->->->->->->->->->->->->->->->---<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<\n\n\n", "msg_date": "Wed, 3 Mar 1999 18:24:59 -0600 (CST)", "msg_from": "James Thompson <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] palloc.h again " }, { "msg_contents": "On Tue, 2 Mar 1999, Michael Meskes wrote:\n\n> I just noticed that the palloc.h problem still exists. That is we do not\n> install utils/mcxt.h but is included by palloc.h.\n> \n\nAfter doing a clean install of todays CVS, I tried to compile one of the\necpg based programs. These are the dependencies that failed as a copied\nthe needed files into place.\n\n[jamest@calvin obelib]$ make\nmake: *** No rule to make target `/usr/local/pgsql/include/utils/mcxt.h',\nneeded by `packet.lo'. Stop.\n[jamest@calvin obelib]$ make\nmake: *** No rule to make target\n`/usr/local/pgsql/include/nodes/memnodes.h', needed by `packet.lo'. Stop.\n[jamest@calvin obelib]$ make\nmake: *** No rule to make target `/usr/local/pgsql/include/lib/fstack.h',\nneeded by `packet.lo'. Stop.\n[jamest@calvin obelib]$ make\nmake: *** No rule to make target\n`/usr/local/pgsql/include/utils/memutils.h', needed by `packet.lo'. Stop.\n[jamest@calvin obelib]$ make\nmake: *** No rule to make target `/usr/local/pgsql/include/nodes/nodes.h',\nneeded by `packet.lo'. Stop.\n\n->->->->->->->->->->->->->->->->->->---<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<\nJames Thompson 138 Cardwell Hall Manhattan, Ks 66506 785-532-0561 \nKansas State University Department of Mathematics\n->->->->->->->->->->->->->->->->->->---<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<\n\n\n", "msg_date": "Sun, 7 Mar 1999 10:34:12 -0600 (CST)", "msg_from": "James Thompson <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] palloc.h again" }, { "msg_contents": "James Thompson <[email protected]> writes:\n> After doing a clean install of todays CVS, I tried to compile one of the\n> ecpg based programs.\n\nWait a sec ... ecpg? ecpg shouldn't be depending on backend internals\nat all, should it? I thought we were talking about support for\ndynamically loaded backend extensions.\n\n> These are the dependencies that failed as a copied\n> the needed files into place.\n>\tinclude/utils/mcxt.h\n>\tinclude/nodes/memnodes.h\n>\tinclude/lib/fstack.h\n>\tinclude/utils/memutils.h\n>\tinclude/nodes/nodes.h\n\nOK, thanks for the list. Now the question is do we want to expose that\nmuch of the backend internals to dynamically loaded extensions in order\nto allow palloc() to be a macro.\n\nIt looks to me like there's no real need for mcxt.h to import\nmemnodes.h; as far as most users of palloc are concerned, the type\nMemoryContext is just an abstract token. I propose putting\n\ttypedef struct MemoryContextData *MemoryContext;\n\ttypedef struct GlobalMemoryData *GlobalMemory;\nin mcxt.h instead of including memnodes.h. Then we avoid needing to\nexport more than mcxt.h... should speed up backend rebuilds a little bit\ntoo, as palloc.h is included in a lot of files that don't need the rest\nof that stuff...\n\nI'll try this and see if it compiles that way.\n\n\t\t\tregards, tom lane\n", "msg_date": "Sun, 07 Mar 1999 14:22:28 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] palloc.h again " }, { "msg_contents": "On Sun, 7 Mar 1999, Tom Lane wrote:\n\n> James Thompson <[email protected]> writes:\n> > After doing a clean install of todays CVS, I tried to compile one of the\n> > ecpg based programs.\n> \n> Wait a sec ... ecpg? ecpg shouldn't be depending on backend internals\n\nI should have left the ecpg part out as I don't think it was important.\nAll my code requires ecpg and it was easy to just build what I'm familure\nwith and submit the results. I didn't try a libpg based program.\n\n->->->->->->->->->->->->->->->->->->---<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<\nJames Thompson 138 Cardwell Hall Manhattan, Ks 66506 785-532-0561 \nKansas State University Department of Mathematics\n->->->->->->->->->->->->->->->->->->---<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<\n\n\n\n", "msg_date": "Sun, 7 Mar 1999 13:47:43 -0600 (CST)", "msg_from": "James Thompson <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] palloc.h again " }, { "msg_contents": "On Sun, Mar 07, 1999 at 02:22:28PM -0500, Tom Lane wrote:\n> Wait a sec ... ecpg? ecpg shouldn't be depending on backend internals\n> at all, should it? I thought we were talking about support for\n> dynamically loaded backend extensions.\n\nHmm, historically ecpg's ecpglib.h includes postgres.h which include\npalloc.h. But scanning through postgres.h I wonder why. \n\nI will check that as soon as I find time.\n\nMichael\n-- \nMichael Meskes | Go SF 49ers!\nTh.-Heuss-Str. 61, D-41812 Erkelenz | Go Rhein Fire!\nTel.: (+49) 2431/72651 | Use Debian GNU/Linux!\nEmail: [email protected] | Use PostgreSQL!\n", "msg_date": "Mon, 8 Mar 1999 08:42:56 +0100", "msg_from": "Michael Meskes <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] palloc.h again" }, { "msg_contents": "Michael Meskes <[email protected]> writes:\n> On Sun, Mar 07, 1999 at 02:22:28PM -0500, Tom Lane wrote:\n>> Wait a sec ... ecpg? ecpg shouldn't be depending on backend internals\n>> at all, should it? I thought we were talking about support for\n>> dynamically loaded backend extensions.\n\n> Hmm, historically ecpg's ecpglib.h includes postgres.h which include\n> palloc.h. But scanning through postgres.h I wonder why. \n\nIt'd probably be better if you could make ecpg code depend only on\npostgres_ext.h. However, I did commit fixes yesterday to solve the\nimmediate problem with mcxt.h.\n\n\t\t\tregards, tom lane\n", "msg_date": "Mon, 08 Mar 1999 09:39:05 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] palloc.h again " }, { "msg_contents": "On Mon, Mar 08, 1999 at 09:39:05AM -0500, Tom Lane wrote:\n> It'd probably be better if you could make ecpg code depend only on\n> postgres_ext.h. However, I did commit fixes yesterday to solve the\n> immediate problem with mcxt.h.\n\nI found I essantially need only c.h (and config.h of course). I will change\nmy include file to just include these.\n\nMichael\n-- \nMichael Meskes | Go SF 49ers!\nTh.-Heuss-Str. 61, D-41812 Erkelenz | Go Rhein Fire!\nTel.: (+49) 2431/72651 | Use Debian GNU/Linux!\nEmail: [email protected] | Use PostgreSQL!\n", "msg_date": "Mon, 8 Mar 1999 20:30:32 +0100", "msg_from": "Michael Meskes <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] palloc.h again" } ]
[ { "msg_contents": "Thank you to each of you who helped me with my PostgreSQL and\nSolaris7 troubles. A special thank you goes out to Marc Fournier,\nTatsuo Ishii and especially Tom Lane, without whom I would have lost\nwhat is left of my mind.\n\nAfter much research, it was determined that the primary cause of my\ndistress was incomplete transactions. Coding in perl and using\nDBI/DBD seems to encapsulate every command (including SELECTs) in a\ntransaction. Apparently we were abandoning the transactions,\nleaving the database to rollback, but not until after holding the\nlock for a very long time. Given the lock-happy behavior of the\ncurrent release, this made the database basically unusable under any\nload. As a matter of fact, a single insert amid dozens of queries\ncould throw the backend so far out to lunch that it never recovers\nand had to be restarted. We closed those holes and even tried every\nother access method available in DBI/DBD. Unfortunately, in a\nbenchmark, MySQL still outperformed PostgreSQL by such:\n\nOne Minute trial, 60 readers, one writer:\nEach reader attempts to retrieve ten rows:\n\t\t\t\tmysql\t\tpgsql\n# of inserts performed 19,000\t26\n# of random seek reads 4,000\t\t7\n\nNot to say this was the only error. We still saw lots and lots of\nthis behavior:\n\nFeb 28 13:17:19 constantinople POSTMASTER:\nFeb 28 13:17:19 constantinople POSTMASTER: FATAL: s_lock(ef68004d)\nat\nspin.c:125, stuck spinlock. Aborting.\n\nWhich leads me to think there are still some aberrant IPC issues\nlurking in Solaris7. I'm not sure if any of you are aware, but Sun\nwent back to sockets and away from streams in the 2.6-2.7 upgrade.\nThis fact alone could count for some of the behavior we have seen.\nWhen we reproduced our problem on Linux, we did not see any of the\nstuck spinlock messages.\n\nI know the new locking methods will improve this greatly, but for\nwhat we hope will be one of the busiest portals on the Internet,\nspeed is very important. We have taken our database portion of the\nsite down in order to port to MySQL. <Sigh> A sad move on my part,\nnow that I know PostgreSQL so well.\n\nThank you all again, be well. See you all on the flip side...\n\nDwD\n--\nDaryl W. Dunbar\nVP of Engineering/Chief Technology Officer\nhttp://www.com, Where the Web Begins!\nmailto:[email protected]\n\n", "msg_date": "Tue, 2 Mar 1999 15:22:58 -0500", "msg_from": "\"Daryl W. Dunbar\" <[email protected]>", "msg_from_op": true, "msg_subject": "So long and thanks for all the fish" } ]
[ { "msg_contents": "> Modified Files:\n> parse.h gram.c\n> Log Message:\n> Someone forgot to commit gram.c and parse.h after his latest\n> set of updates to gram.y.\n\nI didn't forget! istm that we risk cvs bloat by checking in derived\nfiles like gram.c, when there are probably minor differences in how each\nsystem generates them. So I don't usually check in those files when we\nare deep in the middle of hacker development. Of course, that could be\nthe wrong thing to (not) do...\n\n - Tom\n", "msg_date": "Wed, 03 Mar 1999 02:45:11 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [COMMITTERS] 'pgsql/src/backend/parser parse.h gram.c'" }, { "msg_contents": "On Wed, 3 Mar 1999, Thomas G. Lockhart wrote:\n\n> > Modified Files:\n> > parse.h gram.c\n> > Log Message:\n> > Someone forgot to commit gram.c and parse.h after his latest\n> > set of updates to gram.y.\n> \n> I didn't forget! istm that we risk cvs bloat by checking in derived\n> files like gram.c, when there are probably minor differences in how each\n> system generates them. So I don't usually check in those files when we\n> are deep in the middle of hacker development. Of course, that could be\n> the wrong thing to (not) do...\n\nWould have to agree here...developers *should* have the tools on their\nsystems required to generate this...these should only be\ngenerated/committed as part of our pre-release checklist...\n\nI think there are slightly more important things to worry about during\ndevelopment...no? :)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Tue, 2 Mar 1999 23:15:50 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [COMMITTERS] 'pgsql/src/backend/parser parse.h gram.c'" }, { "msg_contents": "The Hermit Hacker <[email protected]> writes:\n> On Wed, 3 Mar 1999, Thomas G. Lockhart wrote:\n>>>> Someone forgot to commit gram.c and parse.h after his latest\n>>>> set of updates to gram.y.\n>> \n>> I didn't forget! istm that we risk cvs bloat by checking in derived\n>> files like gram.c,\n\n> Would have to agree here...developers *should* have the tools on their\n> systems required to generate this...these should only be\n> generated/committed as part of our pre-release checklist...\n> I think there are slightly more important things to worry about during\n> development...no? :)\n\nWell, if you want to know why I was annoyed, I'll explain.\n\nOn my machine, gram.c/parse.h appeared to be newer than gram.y.\nJust a chance artifact of when I happened to do cvs updates,\nno doubt. (It doesn't help that cvs nearly always updates gram.y\nfirst when it has to update them all --- it probably gave me\nversion N of gram.y and version N-1 of the derived files in the\nsame update run, while managing to make the derived files look\nnewer.)\n\nSince gram.c/parse.h were in fact *not* up-to-date with respect\nto header-file changes elsewhere, they didn't compile.\n\nI thought this was breakage due to Bruce's removal of recipe.h/.c,\nand griped accordingly last Tuesday or so. It wasn't till Saturday\nthat I realized the guys whom I was expecting to fix it were on\nvacation and I had better do my own digging. Now it didn't take me\na lot of hours to realize what was wrong, but it very easily could've\n--- and in any case I'd already spent a couple of evenings doing other\nstuff when I had wanted to work on Postgres.\n\nMy feeling is this: if you want to remove gram.c/parse.h from the\nCVS file set entirely, expecting developer types to have the\nnecessary tools, that's a defensible approach. (Tarball drops\ncould and should include freshly-generated copies, of course.)\nOr, if you want to keep them in CVS and *keep them up to date*,\nthat works for me too. But don't be wishy-washy. You're just\nwasting download time and developer time if you don't treat\nthese files consistently.\n\n\t\t\tstill fairly annoyed, tom lane\n", "msg_date": "Wed, 03 Mar 1999 01:53:02 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [COMMITTERS] 'pgsql/src/backend/parser parse.h gram.c' " }, { "msg_contents": "Then <[email protected]> spoke up and said:\n> > Modified Files:\n> > parse.h gram.c\n> > Log Message:\n> > Someone forgot to commit gram.c and parse.h after his latest\n> > set of updates to gram.y.\n> \n> I didn't forget! istm that we risk cvs bloat by checking in derived\n> files like gram.c, when there are probably minor differences in how each\n> system generates them. So I don't usually check in those files when we\n> are deep in the middle of hacker development. Of course, that could be\n> the wrong thing to (not) do...\n\nYou *never* want to check in derived files when using CVS. Instead,\ncreate checkout rules (see the cvs docs) to automatically create those\nfiles upon checkin/checkout, whichever is more appropriate.\n\n-- \n=====================================================================\n| JAVA must have been developed in the wilds of West Virginia. |\n| After all, why else would it support only single inheritance?? |\n=====================================================================\n| Finger [email protected] for my public key. |\n=====================================================================", "msg_date": "3 Mar 1999 08:43:03 -0500", "msg_from": "[email protected]", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: [COMMITTERS] 'pgsql/src/backend/parser parse.h\n\tgram.c'" }, { "msg_contents": "> >> I didn't forget! istm that we risk cvs bloat by checking in derived\n> >> files like gram.c,\n> Well, if you want to know why I was annoyed, I'll explain.\n> On my machine, gram.c/parse.h appeared to be newer than gram.y.\n> Since gram.c/parse.h were in fact *not* up-to-date with respect\n> to header-file changes elsewhere, they didn't compile.\n\nOh! I haven't seen that behavior before, and am not sure why you did :(\nIf I updated gram.y, but did not update gram.c, then cvs and CVSup\nshould never bollix up the times on the files afaik.\n\nSorry for the diversion, but I'm confused as to how you got this\nsymptom. Been doing this for a couple of years now, and doing a *lot*\nwith gram.y so if it was a checkin or sync problem I would have expected\nto see it before this.\n\nbtw, one of the changes I made was to move the backend/parser/ build to\nearlier in the build list, since when debugging is enabled the node\nprinting functions (now) need to see the definitions of some of the\nparser nodes. I wonder if somehow that was involved??\n\n - Tom\n", "msg_date": "Wed, 03 Mar 1999 14:41:36 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [COMMITTERS] 'pgsql/src/backend/parser parse.h gram.c'" }, { "msg_contents": "\"Thomas G. Lockhart\" <[email protected]> writes:\n>> Well, if you want to know why I was annoyed, I'll explain.\n>> On my machine, gram.c/parse.h appeared to be newer than gram.y.\n>> Since gram.c/parse.h were in fact *not* up-to-date with respect\n>> to header-file changes elsewhere, they didn't compile.\n\n> Oh! I haven't seen that behavior before, and am not sure why you did :(\n> If I updated gram.y, but did not update gram.c, then cvs and CVSup\n> should never bollix up the times on the files afaik.\n\nI don't think it's cvs' fault. The recent checkin times for gram.y are:\n2.57 1999/02/23 07:42:41; author: thomas\n2.56 1999/02/21 03:49:00; author: scrappy\n2.55 1999/02/13 23:17:03; author: momjian\n2.54 1999/02/08 14:14:12; author: wieck\n2.53 1999/02/07 19:02:19; author: wieck\n\nand gram.c:\n2.75 1999/02/27 21:33:53; author: tgl\n2.74 1999/02/22 05:26:33; author: momjian\n2.73 1999/02/14 05:14:09; author: momjian\n2.72 1999/02/13 23:16:54; author: momjian\n2.71 1999/02/09 06:30:40; author: momjian\n2.70 1999/02/09 03:51:30; author: momjian\n2.69 1999/02/07 19:04:59; author: wieck\n\nUnfortunately I do not have a log of when I recently did \"cvs update\"s\n(unless someone knows a way to extract that info from cvs itself)?\nBut I will bet that when I updated on 2/24, it had been more than two\ndays since my previous update, and so cvs updated both gram.y (to 2.57)\nand gram.c (to 2.74, that being Bruce's latest commit) in the same\nupdate cycle. Since cvs timestamps updated files with the time of\n*retrieval*, gram.c ended up having a slightly newer timestamp locally\nthan gram.y, and so make didn't think it needed to be rebuilt.\n\nI'm using cvs 1.10 here, but I don't think its behavior has changed\nin this respect in any recent version. If it did not timestamp updates\nwith time of retrieval, it would create synchronization bugs with\nrespect to locally created files, which'd be no fun either.\n\nI still stand by my observation: either gram.c is a derived file and\nshouldn't be in the CVS repository at all, or else it is a master file\nand must be maintained just as faithfully as any other source file.\nThere is no middle ground that will work reliably with CVS. Especially\nnot in the face of multiple authors some of whom commit the derived file\nand some of whom don't.\n\n\t\t\tregards, tom lane\n\nPS: treating gram.c as a derived file that is made while building\na tarball would also eliminate our recurring problem with gram.c\nappearing to be out-of-date in releases, when it really isn't.\n", "msg_date": "Wed, 03 Mar 1999 10:08:13 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [COMMITTERS] 'pgsql/src/backend/parser parse.h gram.c' " }, { "msg_contents": "> I don't think it's cvs' fault.\n> I'm using cvs 1.10 here, but I don't think its behavior has changed\n> in this respect in any recent version. If it did not timestamp \n> updates with time of retrieval, it would create synchronization bugs \n> with respect to locally created files, which'd be no fun either.\n\nI use CVSup, and it seems to keep the creation date of files consistant\nwith the source cvs tree. Does anonymous cvs not do that? Again, I\n*don't* see the behavior that you do, at least given my CVSup/cvs-1.9\nsystem.\n\n> PS: treating gram.c as a derived file that is made while building\n> a tarball would also eliminate our recurring problem with gram.c\n> appearing to be out-of-date in releases, when it really isn't.\n\nSure. Who wants to work that out?\n\nbtw, could all of this be traced to bad dependencies on parse.h? Our\ncurrent Makefile system does not do this correctly. I applied some small\npatches recently to help, but it did not fix the fundamental problem\nthat all the backend/* directories refer to backend/parse.h, but they do\nnot know that backend/parse.h is a copy of backend/parser/parse.h.\n\n - Tom\n", "msg_date": "Wed, 03 Mar 1999 16:52:34 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [COMMITTERS] 'pgsql/src/backend/parser parse.h gram.c'" }, { "msg_contents": "\"Thomas G. Lockhart\" <[email protected]> writes:\n>> I'm using cvs 1.10 here, but I don't think its behavior has changed\n>> in this respect in any recent version. If it did not timestamp \n>> updates with time of retrieval, it would create synchronization bugs \n>> with respect to locally created files, which'd be no fun either.\n\n> I use CVSup, and it seems to keep the creation date of files consistant\n> with the source cvs tree. Does anonymous cvs not do that?\n\nI don't know anything about CVSup, but plain cvs works the way I\ndescribed, and *should* work the way I described. Otherwise, when\nyou update a \".c\" file from the repository, it might not look like\nit's newer than the corresponding \".o\" file that you generated locally\n(since that could have been made later than someone else committed a\nchange to the .c file).\n\nIf CVSup timestamps files with their repository timestamps, then the\nonly safe way to use it is to \"make distclean\" and rebuild from scratch\nafter every update. (Of course that's probably a good idea anyway since\nwe aren't maintaining reliable dependency info in the makefiles, but we\nshouldn't get forced into it because of a misdesigned tool.)\n\n> btw, could all of this be traced to bad dependencies on parse.h?\n\nNo. The problem was that parse.h itself was not up-to-date.\n\n\t\t\tregards, tom lane\n", "msg_date": "Wed, 03 Mar 1999 18:44:56 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: [COMMITTERS] 'pgsql/src/backend/parser parse.h\n\tgram.c'" }, { "msg_contents": "> > btw, could all of this be traced to bad dependencies on parse.h?\n> No. The problem was that parse.h itself was not up-to-date.\n\nYeah, I guess that was what I was asking. There is a problem in that\nmost of the backend/* subsystems have a dependency on ../parse.h, rather\nthan ../parser/parse.h. So they never know that the parse.h is out of\ndate :(\n\nIf you can fix up this dependency so it does the right thing, that would\nbe A Good Thing...\n\n - Tom\n", "msg_date": "Thu, 04 Mar 1999 02:26:21 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: [COMMITTERS] 'pgsql/src/backend/parser parse.h\n\tgram.c'" } ]
[ { "msg_contents": "============================================================================\n POSTGRESQL BUG REPORT TEMPLATE\n============================================================================\n\n\nYour name\t\t:\tBilly G. Allie\nYour email address\t:\[email protected]\n\n\nSystem Configuration\n---------------------\n Architecture (example: Intel Pentium) \t: Pentium\n\n Operating System (example: Linux 2.0.26 ELF) \t: UnixWare 7.0.1\n\n PostgreSQL version (example: PostgreSQL-6.4) : Current CVS version\n\n Compiler used (example: gcc 2.8.0)\t\t: Optimizing C Compilation\n\t\t\t\t\t\t System (CCS) 3.2 08/18/98\n\t\t\t\t\t\t (u701)\n\nPlease enter a FULL description of your problem:\n------------------------------------------------\nsrc/backend/utils/adt/numeric.c\tfails to compile due to a string having an\nembedded new-line character.\n\n\nPlease describe a way to repeat the problem. Please try to provide a\nconcise reproducible example, if at all possible: \n----------------------------------------------------------------------\nCompile the program. :-)\n\n\nIf you know how this problem might be fixed, list the solution below:\n---------------------------------------------------------------------\nApply the following patch:\n\n------------------------------ 8< CUT HERE >8 ------------------------------\n*** src/backend/utils/adt/numeric.c.orig\tMon Mar 1 10:13:27 1999\n--- src/backend/utils/adt/numeric.c\tMon Mar 1 23:09:00 1999\n***************\n*** 2370,2377 ****\n \tif (var->weight >= maxweight)\n \t{\n \t\tfree_allvars();\n! \t\telog(ERROR, \"overflow on numeric\n! ABS(value) >= 10^%d for field with precision %d scale %d\",\n \t\t\t\t\t\t\tvar->weight, precision, scale);\n \t}\n \n--- 2370,2377 ----\n \tif (var->weight >= maxweight)\n \t{\n \t\tfree_allvars();\n! \t\telog(ERROR, \"overflow on numeric \"\n! \"ABS(value) >= 10^%d for field with precision %d scale %d\",\n \t\t\t\t\t\t\tvar->weight, precision, scale);\n \t}\n \n------------------------------ 8< CUT HERE >8 ------------------------------\n\n\n\n____ | Billy G. Allie | Domain....: [email protected]\n| /| | 7436 Hartwell | Compuserve: 76337,2061\n|-/-|----- | Dearborn, MI 48126| MSN.......: [email protected]\n|/ |LLIE | (313) 582-1540 |", "msg_date": "Tue, 02 Mar 1999 22:51:06 -0500", "msg_from": "\"Billy G. Allie\" <[email protected]>", "msg_from_op": true, "msg_subject": "None" }, { "msg_contents": "Applied.\n\n\n> \n> ============================================================================\n> POSTGRESQL BUG REPORT TEMPLATE\n> ============================================================================\n> \n> \n> Your name\t\t:\tBilly G. Allie\n> Your email address\t:\[email protected]\n> \n> \n> System Configuration\n> ---------------------\n> Architecture (example: Intel Pentium) \t: Pentium\n> \n> Operating System (example: Linux 2.0.26 ELF) \t: UnixWare 7.0.1\n> \n> PostgreSQL version (example: PostgreSQL-6.4) : Current CVS version\n> \n> Compiler used (example: gcc 2.8.0)\t\t: Optimizing C Compilation\n> \t\t\t\t\t\t System (CCS) 3.2 08/18/98\n> \t\t\t\t\t\t (u701)\n> \n> Please enter a FULL description of your problem:\n> ------------------------------------------------\n> src/backend/utils/adt/numeric.c\tfails to compile due to a string having an\n> embedded new-line character.\n> \n> \n> Please describe a way to repeat the problem. Please try to provide a\n> concise reproducible example, if at all possible: \n> ----------------------------------------------------------------------\n> Compile the program. :-)\n> \n> \n> If you know how this problem might be fixed, list the solution below:\n> ---------------------------------------------------------------------\n> Apply the following patch:\n> \n> ------------------------------ 8< CUT HERE >8 ------------------------------\n> *** src/backend/utils/adt/numeric.c.orig\tMon Mar 1 10:13:27 1999\n> --- src/backend/utils/adt/numeric.c\tMon Mar 1 23:09:00 1999\n> ***************\n> *** 2370,2377 ****\n> \tif (var->weight >= maxweight)\n> \t{\n> \t\tfree_allvars();\n> ! \t\telog(ERROR, \"overflow on numeric\n> ! ABS(value) >= 10^%d for field with precision %d scale %d\",\n> \t\t\t\t\t\t\tvar->weight, precision, scale);\n> \t}\n> \n> --- 2370,2377 ----\n> \tif (var->weight >= maxweight)\n> \t{\n> \t\tfree_allvars();\n> ! \t\telog(ERROR, \"overflow on numeric \"\n> ! \"ABS(value) >= 10^%d for field with precision %d scale %d\",\n> \t\t\t\t\t\t\tvar->weight, precision, scale);\n> \t}\n> \n> ------------------------------ 8< CUT HERE >8 ------------------------------\n> \nContent-Description: uw7.patch\n\n[Attachment, skipping...]\n\n> ____ | Billy G. Allie | Domain....: [email protected]\n> | /| | 7436 Hartwell | Compuserve: 76337,2061\n> |-/-|----- | Dearborn, MI 48126| MSN.......: [email protected]\n> |/ |LLIE | (313) 582-1540 | \n\n[application/pgp-signature is not supported, skipping...]\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 14 Mar 1999 11:49:19 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: your mail" } ]
[ { "msg_contents": "\n> The second problem is this:\n> \n> conn=153237224, query='SELECT \"RentalOrders\".\"rentalorderlinesid\" FROM \"rentalorderlines\"\n> \"RentalOrders\" WHERE (\"rentalorderid\" = NULL ) '\n> ERROR from backend during send_query: 'ERROR: parser: parse error at or near \"null\"'\n> STATEMENT ERROR: func=SC_execute, desc='', errnum=1, errmsg='Error while executing the query'\n> \n> \n> Since postgres will not recognize the syntax (where 'col' = null)... it only recognizes\n> \"isnull\". I was hoping someone would have added the ability for the parser to handle this at\n> some point (Hey Dave, maybe you could contribute something here man :-).\n> \n> \n> Byron\n> \n\nI do not poke my nose into odbc as I have nothing to do with it, but\nthis parsing problem caught my attention. To me, 'isnull' and '= null'\nare the same token. So I fixed the aforementioned problem like this:\n\n1. In backend/parser.scan.l, find the line that reads:\n\nidentifier {letter}{letter_or_digit}*\n\nand put this following macro after it:\n\nisnull ={space}*(null|NULL)\n\nit does not matter where before the beginning of the rules section you\nwill put it, but it is better to keep related things close to each other.\n\n\n2. In the same file, find the line that reads:\n\n{identifier} {\n\nand insert the following rule before it\n\n{isnull} {\n int i;\n ScanKeyword *keyword;\n\n for(i = 0; yytext[i]; i++)\n if (isascii((unsigned char)yytext[i]) &&\n isupper(yytext[i]))\n yytext[i] = tolower(yytext[i]);\n if (i >= NAMEDATALEN)\n yytext[NAMEDATALEN-1] = '\\0';\n\n keyword = ScanKeywordLookup((char*)\"isnull\");\n return keyword->value;\n }\n\n3. run make && make install in the src directory, then stop and restart postmaster\n\n\nI understand it is an ugly hack but if you are desperate to get things running ...\nIf you don't try to use it as NULL = 'col', you should be OK.\n\n--Gene\n", "msg_date": "Wed, 03 Mar 1999 02:04:39 -0500", "msg_from": "\"Gene Selkov Jr.\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [INTERFACES] Foreign Keys " } ]
[ { "msg_contents": "Hi all,\n\nWhile compiling of 6.4.2 I've realized that int8's are not supported\non FreeBSD by default. The Configure script looks if %ld or %lld\nwill do the job but FreeBSD is using %qd as the long long int format.\nAfter a little fiddeling around with the configure script and \ninclude/utils/int8.h, I've got the int8 type working.\nSince I've read somtimes in this Mailinglists, that FreeBSD is one\nof the development platforms for PostgreSQL, I wonder why PGSQL\ndoesn't support FreeBSD's own format ...\n\nBTW: why the configure script think's that tcl and tk includes\n(tclConfig.sh,tkConfig.sh) must reside in the same directory ?\nThat's really an odd assumtion.\n\n\nHolm\n\nPS:\nsorry for my poor english.\n-- \nFreibergNet Systemhaus GbR Holm Tiffe * Administration, Development\nSystemhaus f�r Daten- und Netzwerktechnik phone +49 3731 781279\nUnternehmensgruppe Liebscher & Partner fax +49 3731 781377\nD-09599 Freiberg * Am St. Niclas Schacht 13 http://www.freibergnet.de/\n\n", "msg_date": "Wed, 3 Mar 1999 08:28:16 +0100", "msg_from": "Holm Tiffe <[email protected]>", "msg_from_op": true, "msg_subject": "int 8 on FreeBSD" }, { "msg_contents": "\nNetBSD is the same way. Any chance of someone coming up with a patch\nif int8 isn't supported?\n\n--Michael\n\nHolm Tiffe <[email protected]> writes:\n\n> Hi all,\n> \n> While compiling of 6.4.2 I've realized that int8's are not supported\n> on FreeBSD by default. The Configure script looks if %ld or %lld\n> will do the job but FreeBSD is using %qd as the long long int format.\n> After a little fiddeling around with the configure script and \n> include/utils/int8.h, I've got the int8 type working.\n> Since I've read somtimes in this Mailinglists, that FreeBSD is one\n> of the development platforms for PostgreSQL, I wonder why PGSQL\n> doesn't support FreeBSD's own format ...\n> \n> BTW: why the configure script think's that tcl and tk includes\n> (tclConfig.sh,tkConfig.sh) must reside in the same directory ?\n> That's really an odd assumtion.\n> \n> \n> Holm\n> \n> PS:\n> sorry for my poor english.\n> -- \n> FreibergNet Systemhaus GbR Holm Tiffe * Administration, Development\n> Systemhaus f�r Daten- und Netzwerktechnik phone +49 3731 781279\n> Unternehmensgruppe Liebscher & Partner fax +49 3731 781377\n> D-09599 Freiberg * Am St. Niclas Schacht 13 http://www.freibergnet.de/\n", "msg_date": "03 Mar 1999 00:32:46 -0800", "msg_from": "Michael Graff <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "On Wed, 3 Mar 1999, Holm Tiffe wrote:\n\n> Hi all,\n> \n> While compiling of 6.4.2 I've realized that int8's are not supported\n> on FreeBSD by default. The Configure script looks if %ld or %lld\n> will do the job but FreeBSD is using %qd as the long long int format.\n> After a little fiddeling around with the configure script and \n> include/utils/int8.h, I've got the int8 type working.\n> Since I've read somtimes in this Mailinglists, that FreeBSD is one\n> of the development platforms for PostgreSQL, I wonder why PGSQL\n> doesn't support FreeBSD's own format ...\n> \n> BTW: why the configure script think's that tcl and tk includes\n> (tclConfig.sh,tkConfig.sh) must reside in the same directory ?\n> That's really an odd assumtion.\n\nIts only an odd assumption to those on FreeBSD...we tend to be the only\nones that pervert most of the \"standards\" as far as installation\ndirectories are concerned, and require specific handholding to get things\nto install :(\n\nSee /usr/ports/databases/postgresql/Makefile, where it provides the\nvarious --with-tcl= directives required...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 3 Mar 1999 08:11:57 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "The Hermit Hacker wrote:\n\n> On Wed, 3 Mar 1999, Holm Tiffe wrote:\n> \n> > Hi all,\n> > \n[..]\n> > \n> > BTW: why the configure script think's that tcl and tk includes\n> > (tclConfig.sh,tkConfig.sh) must reside in the same directory ?\n> > That's really an odd assumtion.\n> \n> Its only an odd assumption to those on FreeBSD...we tend to be the only\n> ones that pervert most of the \"standards\" as far as installation\n> directories are concerned, and require specific handholding to get things\n> to install :(\n\n..but it is possible on FreeBSD to run several different tcl/tk pairs\nin parallel, so it turns out that it is the better way and I can't\nremeber that I have had such problems with an other package.\n\nPersonally I HATE GNU's autoconf, it is really a pain to get things\nworking on \"non standard environments\". I prefer a Makefile in which\nI can set options directly. But this is my own opinion and I can \nunderstand why others doesn't think so.\n\n> \n> See /usr/ports/databases/postgresql/Makefile, where it provides the\n> various --with-tcl= directives required...\n\nYeah, I've build PGSQL without the ports, because there was no port\nof 6.4.2 to the time I needed it, and I've got it finally working.\n\nAnother question:\n\nWatfor is the patches directory on ftp.postgresql.org/pub ?\nIt contains almost nothing of the actual patches from\nthe -patches mailinglist.\nThis makes it really not simple to get a halfways actual setup\nwithout sup/cvsup or so...\n\nHolm\n-- \nFreibergNet Systemhaus GbR Holm Tiffe * Administration, Development\nSystemhaus f�r Daten- und Netzwerktechnik phone +49 3731 781279\nUnternehmensgruppe Liebscher & Partner fax +49 3731 781377\nD-09599 Freiberg * Am St. Niclas Schacht 13 http://www.freibergnet.de/\n\n", "msg_date": "Wed, 3 Mar 1999 13:38:25 +0100", "msg_from": "Holm Tiffe <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "On Wed, 3 Mar 1999, Holm Tiffe wrote:\n\n> The Hermit Hacker wrote:\n> \n> > On Wed, 3 Mar 1999, Holm Tiffe wrote:\n> > \n> > > Hi all,\n> > > \n> [..]\n> > > \n> > > BTW: why the configure script think's that tcl and tk includes\n> > > (tclConfig.sh,tkConfig.sh) must reside in the same directory ?\n> > > That's really an odd assumtion.\n> > \n> > Its only an odd assumption to those on FreeBSD...we tend to be the only\n> > ones that pervert most of the \"standards\" as far as installation\n> > directories are concerned, and require specific handholding to get things\n> > to install :(\n> \n> ..but it is possible on FreeBSD to run several different tcl/tk pairs\n> in parallel, so it turns out that it is the better way and I can't\n> remeber that I have had such problems with an other package.\n\nAgreed, but, as far as I've seen so far (and I use FreeBSD myself, and\nhate it), FreeBSD is the only one that confuses the whole mess by doing\nthis :(\n\n> Personally I HATE GNU's autoconf, it is really a pain to get things\n> working on \"non standard environments\". I prefer a Makefile in which\n> I can set options directly. But this is my own opinion and I can \n> understand why others doesn't think so.\n\nAs far as tk/tcl is concerned, FreeBSD is the only \"non-standard\nenvironment\" that I'm aware of...the rest all install to standard\nlocations...\n\n> Watfor is the patches directory on ftp.postgresql.org/pub ?\n> It contains almost nothing of the actual patches from\n> the -patches mailinglist.\n> This makes it really not simple to get a halfways actual setup\n> without sup/cvsup or so...\n\n-patches mailing list is primarily -CURRENT tree stuff, untested. The\npatches directory is pretty much tested stuff that we feel confident in\nppl using...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 3 Mar 1999 09:38:58 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "The Hermit Hacker wrote:\n\n> On Wed, 3 Mar 1999, Holm Tiffe wrote:\n> \n> > The Hermit Hacker wrote:\n> > \n> > > On Wed, 3 Mar 1999, Holm Tiffe wrote:\n> > > \n> > > > Hi all,\n> > > > \n> > [..]\n> > > > \n> > > > BTW: why the configure script think's that tcl and tk includes\n> > > > (tclConfig.sh,tkConfig.sh) must reside in the same directory ?\n> > > > That's really an odd assumtion.\n> > > \n> > > Its only an odd assumption to those on FreeBSD...we tend to be the only\n> > > ones that pervert most of the \"standards\" as far as installation\n> > > directories are concerned, and require specific handholding to get things\n> > > to install :(\n> > \n> > ..but it is possible on FreeBSD to run several different tcl/tk pairs\n> > in parallel, so it turns out that it is the better way and I can't\n> > remeber that I have had such problems with an other package.\n> \n> Agreed, but, as far as I've seen so far (and I use FreeBSD myself, and\n> hate it),\n\nWhy you use it ?\n\n> FreeBSD is the only one that confuses the whole mess by doing\n> this :(\n\nI think, this is because FreeBSD has learned. There was a time who Tcl/Tk\nversions where released nearly every day. FreeBSD has had an \"imported\"\ntcl version in it's source dist. All Tcl versions where incomatible\nto each other this time, and I've had to work around this several times.\nThe current behavior of FreeBSD is the only one solution for this.\nThink of Linux, which libc today ?\n\n> \n> > Personally I HATE GNU's autoconf, it is really a pain to get things\n> > working on \"non standard environments\". I prefer a Makefile in which\n> > I can set options directly. But this is my own opinion and I can \n> > understand why others doesn't think so.\n> \n> As far as tk/tcl is concerned, FreeBSD is the only \"non-standard\n> environment\" that I'm aware of...the rest all install to standard\n> locations...\n\nNo, that was not only relating to Tcl/Tk, there are many other things\nlike configuring a gcc-cross build environment and so on.\n\n> \n> > Watfor is the patches directory on ftp.postgresql.org/pub ?\n> > It contains almost nothing of the actual patches from\n> > the -patches mailinglist.\n> > This makes it really not simple to get a halfways actual setup\n> > without sup/cvsup or so...\n> \n> -patches mailing list is primarily -CURRENT tree stuff, untested. The\n> patches directory is pretty much tested stuff that we feel confident in\n> ppl using...\n\nHmm, there must be very few confident things, when I take a look to this\npatch directory....\n\nA \"Migration Path\" (deltas) from on release to the next is on my wishlist,\nto get a customers installation up to date without the full net access.\n\nHolm\nPs: sorry for my poor english\n-- \nFreibergNet Systemhaus GbR Holm Tiffe * Administration, Development\nSystemhaus f�r Daten- und Netzwerktechnik phone +49 3731 781279\nUnternehmensgruppe Liebscher & Partner fax +49 3731 781377\nD-09599 Freiberg * Am St. Niclas Schacht 13 http://www.freibergnet.de/\n\n", "msg_date": "Wed, 3 Mar 1999 15:24:57 +0100", "msg_from": "Holm Tiffe <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "On Wed, 3 Mar 1999, Holm Tiffe wrote:\n\n> The Hermit Hacker wrote:\n> \n> > Agreed, but, as far as I've seen so far (and I use FreeBSD myself, and\n> > hate it),\n> \n> Why you use it ?\n\nI knew after I typed that that I had worded it wrong...I hate the way it\ninstalls various ports, like tk/tcl ... I don't hate FreeBSD itself...\n\n> > FreeBSD is the only one that confuses the whole mess by doing\n> > this :(\n> \n> I think, this is because FreeBSD has learned. There was a time who Tcl/Tk\n> versions where released nearly every day. FreeBSD has had an \"imported\"\n> tcl version in it's source dist. All Tcl versions where incomatible\n> to each other this time, and I've had to work around this several times.\n> The current behavior of FreeBSD is the only one solution for this.\n> Think of Linux, which libc today ?\n\nThe \"only one solution for this\" is the standard/accepted one by everyone\n*except* for FreeBSD, that is *why* we have standards...\n\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 3 Mar 1999 10:39:59 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "> > While compiling of 6.4.2 I've realized that int8's are not supported\n> > on FreeBSD by default. The Configure script looks if %ld or %lld\n> > will do the job but FreeBSD is using %qd as the long long int \n> > format.\n> > After a little fiddeling around with the configure script and\n> > include/utils/int8.h, I've got the int8 type working.\n> > Since I've read somtimes in this Mailinglists, that FreeBSD is one\n> > of the development platforms for PostgreSQL, I wonder why PGSQL\n> > doesn't support FreeBSD's own format ...\n\nI had originally coded in %qd for some platforms, but *strongly* recall\nthat those platforms which supported %qd also supported %lld, so we\nremoved the %qd variant.\n\nIf int8 really doesn't work on FreeBSD, shame on the FreeBSDers for not\nbringing it up earlier. But, it can't be that simple. If it *really*\ndoesn't work on FreeBSD, could it be related to relatively recent\nchanges in compilers or libraries (or are you running an old version of\nFreeBSD)?\n\nscrappy, can you take a look at this on your platform?\n\n - Tom\n", "msg_date": "Wed, 03 Mar 1999 14:49:20 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "The Hermit Hacker wrote:\n\n> On Wed, 3 Mar 1999, Holm Tiffe wrote:\n> \n> > The Hermit Hacker wrote:\n> > \n> > > Agreed, but, as far as I've seen so far (and I use FreeBSD myself, and\n> > > hate it),\n> > \n> > Why you use it ?\n> \n> I knew after I typed that that I had worded it wrong...I hate the way it\n> installs various ports, like tk/tcl ... I don't hate FreeBSD itself...\n> \n> > > FreeBSD is the only one that confuses the whole mess by doing\n> > > this :(\n> > \n> > I think, this is because FreeBSD has learned. There was a time who Tcl/Tk\n> > versions where released nearly every day. FreeBSD has had an \"imported\"\n> > tcl version in it's source dist. All Tcl versions where incomatible\n> > to each other this time, and I've had to work around this several times.\n> > The current behavior of FreeBSD is the only one solution for this.\n> > Think of Linux, which libc today ?\n> \n> The \"only one solution for this\" is the standard/accepted one by everyone\n> *except* for FreeBSD, that is *why* we have standards...\n\nI meant: the only one solution to install more than one such package on one\nmachine at a time.\n(Why we have problems to express our self today ? :-))\n[ok, I have languages related problems too...]\n\nback to standards:\nWhy a PC uses metric AND inch screws ?\nWhy we have MS Windows (grr), MacOS, OS/2, *BSD, Linux .... ?\n(order isn't important here)\nIt's because we all have different tastes ...\n\nYou hate it, an I simply accept it.\nI have had to use tcl3 and tcl8 on a customers machine. \n\nHolm\n-- \nFreibergNet Systemhaus GbR Holm Tiffe * Administration, Development\nSystemhaus f�r Daten- und Netzwerktechnik phone +49 3731 781279\nUnternehmensgruppe Liebscher & Partner fax +49 3731 781377\nD-09599 Freiberg * Am St. Niclas Schacht 13 http://www.freibergnet.de/\n\n", "msg_date": "Wed, 3 Mar 1999 15:54:38 +0100", "msg_from": "Holm Tiffe <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "Thomas G. Lockhart wrote:\n\n> > > While compiling of 6.4.2 I've realized that int8's are not supported\n> > > on FreeBSD by default. The Configure script looks if %ld or %lld\n> > > will do the job but FreeBSD is using %qd as the long long int \n> > > format.\n> > > After a little fiddeling around with the configure script and\n> > > include/utils/int8.h, I've got the int8 type working.\n> > > Since I've read somtimes in this Mailinglists, that FreeBSD is one\n> > > of the development platforms for PostgreSQL, I wonder why PGSQL\n> > > doesn't support FreeBSD's own format ...\n> \n> I had originally coded in %qd for some platforms, but *strongly* recall\n> that those platforms which supported %qd also supported %lld, so we\n> removed the %qd variant.\n> \n> If int8 really doesn't work on FreeBSD, shame on the FreeBSDers for not\n> bringing it up earlier.\n\n:-)\n> But, it can't be that simple.\n\nThe regression test means it is, and this is possibly the reason why it\nisn't fixed jet :-)\n\n> If it *really*\n> doesn't work on FreeBSD, could it be related to relatively recent\n> changes in compilers or libraries (or are you running an old version of\n> FreeBSD)?\n\nNo, this is my development box:\n\n4.0-CURRENT FreeBSD 4.0-CURRENT #6: Mon Mar 1 12:55:03 CET 1999\n\nIt is a very current system, but I've stumbled over this some time\nbevore (int8 was a contrib package this times)\n\nHolm\n-- \nFreibergNet Systemhaus GbR Holm Tiffe * Administration, Development\nSystemhaus f�r Daten- und Netzwerktechnik phone +49 3731 781279\nUnternehmensgruppe Liebscher & Partner fax +49 3731 781377\nD-09599 Freiberg * Am St. Niclas Schacht 13 http://www.freibergnet.de/\n\n", "msg_date": "Wed, 3 Mar 1999 16:05:24 +0100", "msg_from": "Holm Tiffe <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": ">BTW: why the configure script think's that tcl and tk includes\n>(tclConfig.sh,tkConfig.sh) must reside in the same directory ?\n>That's really an odd assumtion.\n\nI have to agree here. We have an localized version of Tcl/Tk package\nfor FreeBSD having TkConfig.sh under /usr/local/lib/tk4.2jp/. I always\nhand-edit configure to let it recognize tkConfigure.sh.\n---\nTatsuo Ishii\n\n", "msg_date": "Thu, 04 Mar 1999 10:56:52 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD " }, { "msg_contents": "-----BEGIN PGP SIGNED MESSAGE-----\n\n>>>>> \"Tatsuo\" == Tatsuo Ishii <[email protected]> writes:\n\n >> BTW: why the configure script think's that tcl and tk includes\n >> (tclConfig.sh,tkConfig.sh) must reside in the same directory ?\n >> That's really an odd assumtion.\n\n Tatsuo> I have to agree here. We have an localized version of\n Tatsuo> Tcl/Tk package for FreeBSD having TkConfig.sh under\n Tatsuo> /usr/local/lib/tk4.2jp/. I always hand-edit configure to\n Tatsuo> let it recognize tkConfigure.sh. --- Tatsuo Ishii\n\nI don't understand the complaint; doesn't \n\n --with-tclconfig=/usr/local/lib/tk4.2jp\n\nwork for you? or is you tkConfig.sh really called tkConfigure.sh\n(maybe that is what is breaking it...).\n\nroland\n- -- \n\t\t PGP Key ID: 66 BC 3B CD\nRoland B. Roberts, PhD Custom Software Solutions\[email protected] 101 West 15th St #4NN\[email protected] New York, NY 10011\n\n-----BEGIN PGP SIGNATURE-----\nVersion: 2.6.2\nComment: Processed by Mailcrypt 3.4, an Emacs/PGP interface\n\niQCVAwUBNt4AouoW38lmvDvNAQGvpgQAgiJBjhnCL62HlKJm3oVAz+ew6AABNMzv\ni9HGhOOaDU072xgSt3ay/y7vr25TG5XBrhNvuo7zLkrr/RlHV4iSEXM3Nrzga9CV\n/k4teT7IfJ6ZVHOLsEoMPhiEeKimD3tlzIuZSSvc3SDOZHOreaV15V43xYYWtV79\nMFvFCjDJGUo=\n=QDBy\n-----END PGP SIGNATURE-----\n", "msg_date": "03 Mar 1999 22:40:20 -0500", "msg_from": "Roland Roberts <[email protected]>", "msg_from_op": false, "msg_subject": "Re: Tcl/Tk config (was: int 8 on FreeBSD)" }, { "msg_contents": "> Tatsuo> I have to agree here. We have an localized version of\n> Tatsuo> Tcl/Tk package for FreeBSD having TkConfig.sh under\n> Tatsuo> /usr/local/lib/tk4.2jp/. I always hand-edit configure to\n> Tatsuo> let it recognize tkConfigure.sh. --- Tatsuo Ishii\n>\n>I don't understand the complaint; doesn't \n>\n> --with-tclconfig=/usr/local/lib/tk4.2jp\n>\n>work for you? or is you tkConfig.sh really called tkConfigure.sh\n>(maybe that is what is breaking it...).\n\nI don't think that works. Since tclConfig.sh is under\n/usr/local/lib/tcl7.6jp/ in my configuration. As already pointed out,\nthe problem is configure doesn't allow tclConfig.sh and tkConfig.sh\nlives in separate directory.\n--\nTatsuo Ishii\n", "msg_date": "Thu, 04 Mar 1999 12:51:15 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Tcl/Tk config (was: int 8 on FreeBSD) " }, { "msg_contents": "On Thu, 4 Mar 1999, Tatsuo Ishii wrote:\n\n> > Tatsuo> I have to agree here. We have an localized version of\n> > Tatsuo> Tcl/Tk package for FreeBSD having TkConfig.sh under\n> > Tatsuo> /usr/local/lib/tk4.2jp/. I always hand-edit configure to\n> > Tatsuo> let it recognize tkConfigure.sh. --- Tatsuo Ishii\n> >\n> >I don't understand the complaint; doesn't \n> >\n> > --with-tclconfig=/usr/local/lib/tk4.2jp\n> >\n> >work for you? or is you tkConfig.sh really called tkConfigure.sh\n> >(maybe that is what is breaking it...).\n> \n> I don't think that works. Since tclConfig.sh is under\n> /usr/local/lib/tcl7.6jp/ in my configuration. As already pointed out,\n> the problem is configure doesn't allow tclConfig.sh and tkConfig.sh\n> lives in separate directory.\n\nFreeeBSD does it as:\n\n--with-tclconfig\"/usr/local/lib/tk4.2jp /usr/local/lib/tcl7.6jp\"\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 4 Mar 1999 09:22:38 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Tcl/Tk config (was: int 8 on FreeBSD) " }, { "msg_contents": "On Thu, 4 Mar 1999, The Hermit Hacker wrote:\n\n> FreeeBSD does it as:\n> \n> --with-tclconfig\"/usr/local/lib/tk4.2jp /usr/local/lib/tcl7.6jp\"\n\nThat Works?!?! Damn. One thing I never thought to try.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Thu, 4 Mar 1999 11:11:03 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Tcl/Tk config (was: int 8 on FreeBSD) " }, { "msg_contents": "On Thu, 4 Mar 1999, Vince Vielhaber wrote:\n\n> On Thu, 4 Mar 1999, The Hermit Hacker wrote:\n> \n> > FreeeBSD does it as:\n> > \n> > --with-tclconfig\"/usr/local/lib/tk4.2jp /usr/local/lib/tcl7.6jp\"\n> \n> That Works?!?! Damn. One thing I never thought to try.\n\nActually, just tried it here, and if it works in our ports collection, it\ndoesn't work outside of it :(\n\nAm going to play aroudn with it from home tonight...would like to get\nthat, at least, fixed for v6.4.3 also...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 4 Mar 1999 13:57:49 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Tcl/Tk config (was: int 8 on FreeBSD) " }, { "msg_contents": "On Thu, 4 Mar 1999, The Hermit Hacker wrote:\n\n> On Thu, 4 Mar 1999, Vince Vielhaber wrote:\n> \n> > On Thu, 4 Mar 1999, The Hermit Hacker wrote:\n> > \n> > > FreeeBSD does it as:\n> > > \n> > > --with-tclconfig\"/usr/local/lib/tk4.2jp /usr/local/lib/tcl7.6jp\"\n> > \n> > That Works?!?! Damn. One thing I never thought to try.\n> \n> Actually, just tried it here, and if it works in our ports collection, it\n> doesn't work outside of it :(\n> \n> Am going to play aroudn with it from home tonight...would like to get\n> that, at least, fixed for v6.4.3 also...\n\nHow 'bout adding a --with-tkconfig=DIR to configure with it defaulting\nto tclconfig's DIR?\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Thu, 4 Mar 1999 13:08:24 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Tcl/Tk config (was: int 8 on FreeBSD) " }, { "msg_contents": "\nSounds reasonable to me...want to do it up and submit a quick patch? I\nwon't get a chance til later tonight...\n\n\nOn Thu, 4 Mar 1999, Vince Vielhaber wrote:\n\n> On Thu, 4 Mar 1999, The Hermit Hacker wrote:\n> \n> > On Thu, 4 Mar 1999, Vince Vielhaber wrote:\n> > \n> > > On Thu, 4 Mar 1999, The Hermit Hacker wrote:\n> > > \n> > > > FreeeBSD does it as:\n> > > > \n> > > > --with-tclconfig\"/usr/local/lib/tk4.2jp /usr/local/lib/tcl7.6jp\"\n> > > \n> > > That Works?!?! Damn. One thing I never thought to try.\n> > \n> > Actually, just tried it here, and if it works in our ports collection, it\n> > doesn't work outside of it :(\n> > \n> > Am going to play aroudn with it from home tonight...would like to get\n> > that, at least, fixed for v6.4.3 also...\n> \n> How 'bout adding a --with-tkconfig=DIR to configure with it defaulting\n> to tclconfig's DIR?\n> \n> Vince.\n> -- \n> ==========================================================================\n> Vince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n> # include <std/disclaimers.h> TEAM-OS2\n> Online Campground Directory http://www.camping-usa.com\n> Online Giftshop Superstore http://www.cloudninegifts.com\n> ==========================================================================\n> \n> \n> \n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 4 Mar 1999 14:56:36 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Tcl/Tk config (was: int 8 on FreeBSD) " }, { "msg_contents": "On 04-Mar-99 The Hermit Hacker wrote:\n> \n> Sounds reasonable to me...want to do it up and submit a quick patch? I\n> won't get a chance til later tonight...\n\nOk. Having never played with autoconf, I think I got it anyway. I do\nhave a minor problem, tho, as I've never gotten tk/tcl to compile cleanly\nwith PostgreSQL. It can never find tcl.h and tk.h unless I play with the\nmakefiles. Whatever the case, attached is a diff for configure.in I hope\nwe can get it tested by a few different platforms. Also would someone \ntake a look at the --help text for it? I guess it's ok but not exactly\nthe clearest. I'm also going to take another look at why the header\nfiles for tk/tck are getting missed.\n\n\n\n> \n> \n> On Thu, 4 Mar 1999, Vince Vielhaber wrote:\n> \n>> On Thu, 4 Mar 1999, The Hermit Hacker wrote:\n>> \n>> > On Thu, 4 Mar 1999, Vince Vielhaber wrote:\n>> > \n>> > > On Thu, 4 Mar 1999, The Hermit Hacker wrote:\n>> > > \n>> > > > FreeeBSD does it as:\n>> > > > \n>> > > > --with-tclconfig\"/usr/local/lib/tk4.2jp /usr/local/lib/tcl7.6jp\"\n>> > > \n>> > > That Works?!?! Damn. One thing I never thought to try.\n>> > \n>> > Actually, just tried it here, and if it works in our ports collection, it\n>> > doesn't work outside of it :(\n>> > \n>> > Am going to play aroudn with it from home tonight...would like to get\n>> > that, at least, fixed for v6.4.3 also...\n>> \n>> How 'bout adding a --with-tkconfig=DIR to configure with it defaulting\n>> to tclconfig's DIR?\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================", "msg_date": "Thu, 04 Mar 1999 15:33:20 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Tcl/Tk config (was: int 8 on FreeBSD)" }, { "msg_contents": "\nOn 04-Mar-99 Vince Vielhaber wrote:\n> \n> On 04-Mar-99 The Hermit Hacker wrote:\n>> \n>> Sounds reasonable to me...want to do it up and submit a quick patch? I\n>> won't get a chance til later tonight...\n> \n> Ok. Having never played with autoconf, I think I got it anyway. I do\n> have a minor problem, tho, as I've never gotten tk/tcl to compile cleanly\n> with PostgreSQL. It can never find tcl.h and tk.h unless I play with the\n> makefiles. Whatever the case, attached is a diff for configure.in I hope\n> we can get it tested by a few different platforms. Also would someone \n> take a look at the --help text for it? I guess it's ok but not exactly\n> the clearest. I'm also going to take another look at why the header\n> files for tk/tck are getting missed.\n\nAn update here, the include dirs need to be added to --with-includes and I\nhad a small typo in that line. (stupid computers, they should know what I\nmean no matter what I type!). It all compiled without incident :)\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Thu, 04 Mar 1999 16:44:40 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Tcl/Tk config (was: int 8 on FreeBSD)" }, { "msg_contents": "Holm Tiffe wrote:\n\n> Hi all,\n\nHi, Holm,\n\n> While compiling of 6.4.2 I've realized that int8's are not supported\n> on FreeBSD by default. The Configure script looks if %ld or %lld\n> will do the job but FreeBSD is using %qd as the long long int format.\n> After a little fiddeling around with the configure script and\n> include/utils/int8.h, I've got the int8 type working.\n> Since I've read somtimes in this Mailinglists, that FreeBSD is one\n> of the development platforms for PostgreSQL, I wonder why PGSQL\n> doesn't support FreeBSD's own format ...\n\nHere are the patches I've used to enable int8 support.\nSince my patches modify configure.in, you'll need to install\nautoconf and regenerate the configure script.\n\nHope this helps,\nKevin.\n\n==============================================================\n\n--- ./src/backend/port/snprintf.c 1998/12/25 02:20:41 64.2\n+++ ./src/backend/port/snprintf.c 1999/01/19 00:37:40 64.2.1.1\n@@ -49,7 +49,7 @@\n #include <sys/param.h>\n\n /* IRIX doesn't do 'long long' in va_arg(), so use a typedef */\n-#ifdef HAVE_LONG_LONG_INT_64\n+#if defined(HAVE_LONG_INT_64) || defined(HAVE_LONG_LONG_INT) ||\ndefined(HAVE_QUAD_INT_64)\n typedef long long long_long;\n typedef unsigned long long ulong_long;\n #endif\n--- ./src/include/utils/int8.h 1998/09/11 17:16:11 64.2\n+++ ./src/include/utils/int8.h 1999/01/19 00:37:41 64.2.1.1\n@@ -35,11 +35,18 @@\n\n #define INT64_FORMAT \"%lld\"\n #else\n+#ifdef HAVE_QUAD_INT_64\n+/* We have working support for \"long long\", use that */\n+typedef long long int64;\n+\n+#define INT64_FORMAT \"%qd\"\n+#else\n /* Won't actually work, but fall back to long int so that int8.c compiles\n*/\n typedef long int int64;\n\n #define INT64_FORMAT \"%ld\"\n #define INT64_IS_BUSTED\n+#endif\n #endif\n #endif\n\n--- ./src/include/config.h.in 1998/12/13 20:08:24 64.2\n+++ ./src/include/config.h.in 1999/01/19 00:37:41 64.2.1.1\n@@ -267,6 +267,9 @@\n /* Set to 1 if type \"long long int\" works and is 64 bits */\n #undef HAVE_LONG_LONG_INT_64\n\n+/* Set to 1 if type \"long long\" works and is 64 bits */\n+#undef HAVE_QUAD_INT_64\n+\n /* Define as the base type of the last arg to accept */\n #undef SOCKET_SIZE_TYPE\n\n--- ./src/configure.in 1998/12/13 20:08:20 64.2\n+++ ./src/configure.in 1999/01/19 00:37:39 64.2.1.1\n@@ -688,6 +688,43 @@\n AC_MSG_RESULT(no),\n AC_MSG_RESULT(assuming not on target machine))\n\n+AC_MSG_CHECKING(whether 'long long' is 64 bits)\n+AC_TRY_RUN([#include <stdio.h>\n+typedef long long int64;\n+#define INT64_FORMAT \"%qd\"\n+\n+int64 a = 20000001;\n+int64 b = 40000005;\n+\n+int does_int64_work()\n+{\n+ int64 c,d,e;\n+ char buf[100];\n+\n+ if (sizeof(int64) != 8)\n+ return 0; /* doesn't look like the right size */\n+\n+ /* we do perfunctory checks on multiply, divide, sprintf, sscanf */\n+ c = a * b;\n+ sprintf(buf, INT64_FORMAT, c);\n+ if (strcmp(buf, \"800000140000005\") != 0)\n+ return 0; /* either multiply or sprintf is busted */\n+ if (sscanf(buf, INT64_FORMAT, &d) != 1)\n+ return 0;\n+ if (d != c)\n+ return 0;\n+ e = d / b;\n+ if (e != a)\n+ return 0;\n+ return 1;\n+}\n+main() {\n+ exit(! does_int64_work());\n+}],\n+ [AC_DEFINE(HAVE_QUAD_INT_64) AC_MSG_RESULT(yes)],\n+ AC_MSG_RESULT(no),\n+ AC_MSG_RESULT(assuming not on target machine))\n+\n dnl Checks for library functions.\n AC_FUNC_MEMCMP\n AC_TYPE_SIGNAL\n\n", "msg_date": "Sun, 07 Mar 1999 12:47:49 +0800", "msg_from": "Kevin Lo <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" } ]
[ { "msg_contents": "I think this is another one that falls under the general fix all:\nrun the \"vacuum\" command.\n\n-----Original Message-----\nFrom: Karl Auer <[email protected]>\nTo: [email protected] <[email protected]>\nDate: Wednesday, March 03, 1999 3:56 AM\nSubject: [GENERAL] what on earth does this mean?!?\n\n\n>Hi all.\n>\n>Apologies to those who have seen this in the novice list already,\n>someone suggested here might be a better spot for my question.\n>\n>Here 'tis.\n>\n>I have just started getting the following message, by the ton, in my\n>postgres log:\n>\n> ERROR: RelationCatalogInformation: Relation 21400 not found\n>\n>I've traced it to the point that I know exactly what statement is producing\n>it, namely:\n>\n> select attribute_id from dictionary where name = 'USR-MP-EDO'\n>\n>Note that this staement is representative - the same statement is executed\n>many times but with different names; each statement, regardles of the\n>name in it, causes exactly the same error message, including the same\n>number (21400). Needless to say, the job that needs doing, of which this\n>little select is a part, is not getting done :-(\n>\n>When I execute this statement by hand in the same database, as the same\nuser\n>as is executing the script (where these statements are coming from), it\nworks\n>fine.\n>\n>Any ideas most gratefully accepted. Mostly I'd like to know what the error\n>message means...\n>\n>Regards, K.\n>\n>---\n>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\n~~\n>Karl Auer ([email protected]) Geschaeft/work\n+41-1-6327531\n>Kommunikation, ETHZ RZ Privat/home\n+41-1-4517941\n>Clausiusstrasse 59 Fax\n+41-1-6321225\n>CH-8092 ZUERICH Switzerland\n>\n\n\n", "msg_date": "Wed, 3 Mar 1999 04:55:48 -0500", "msg_from": "\"K.T.\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [GENERAL] what on earth does this mean?!?" }, { "msg_contents": "Thanks KT\n\nVacuum sadly does not help.\n\nIs there anyone out there who actually knows what this message *means*? Not\nnecessarily specifically for 21400 of course, but in general?\n\n ERROR: RelationCatalogInformation: Relation 21400 not found\n\nRegards, K.\n\n---\n~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\nKarl Auer ([email protected]) Geschaeft/work +41-1-6327531\nKommunikation, ETHZ RZ Privat/home +41-1-4517941\nClausiusstrasse 59 Fax +41-1-6321225\nCH-8092 ZUERICH Switzerland\n", "msg_date": "Wed, 03 Mar 1999 11:58:04 +0100 (MET)", "msg_from": "Karl Auer <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [GENERAL] what on earth does this mean?!?" } ]
[ { "msg_contents": "\nI probably already know this answer, I just wanna make sure. If I do\na pg_dumpall from one machine and restore it to another, are the existing\ndatabases on the new machine affected if they're not of the same name as\nwhat's being restored? IOW, I have a database on the new machine called\nclassifieds. I'm about to move camping-usa's databases (there isn't one\ncalled classifieds) to this machine. Is the existing \"classifieds\" safe?\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Wed, 3 Mar 1999 08:44:59 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": true, "msg_subject": "pg_dumpall and restoring" } ]
[ { "msg_contents": "I have a problem with Access97 not working properly when entering new\nrecords using a sub form, i.e. entering a new order/orderlines or master and\ndetail tables. The problem is caused by a SQL statement that Access97 makes\ninvolving NULL. The syntax that fails is \"column_name\" = NULL. The\nfollowing attachment was provided by -Jose'-. It contains a very small\nenhancement to gram.y that will allow Access97 to work properly with sub\nforms. Can this enhancement be added to release 6.5?\n\n <<gram.patch>> \nThanks, Michael", "msg_date": "Wed, 3 Mar 1999 13:50:12 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "parser enhancement request for 6.5" }, { "msg_contents": "> I have a problem with Access97 not working properly when entering new\n> records using a sub form, i.e. entering a new order/orderlines or \n> master and detail tables. The problem is caused by a SQL statement \n> that Access97 makes involving NULL. The syntax that fails is \n> \"column_name\" = NULL. The following attachment was provided by \n> -Jose'-. It contains a very small enhancement to gram.y that will \n> allow Access97 to work properly with sub forms. Can this enhancement \n> be added to release 6.5?\n\nI recall discussing this earlier, and after the obligatory objection\nthat this is non-standard (a la everything M$) it should be OK to do it.\nI could have sworn that I'd made this fix, but can't see it in current\nsources and can't see any evidence that it ever was :(\n\nWill enable this for v6.5, with the proviso that it is fair game for\nremoval if it ever conflicts with other SQL parsing needs.\n\n - Tom\n", "msg_date": "Thu, 04 Mar 1999 02:41:01 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] parser enhancement request for 6.5" }, { "msg_contents": "Applied.\n\n\n[Charset iso-8859-1 unsupported, filtering to ASCII...]\n> I have a problem with Access97 not working properly when entering new\n> records using a sub form, i.e. entering a new order/orderlines or master and\n> detail tables. The problem is caused by a SQL statement that Access97 makes\n> involving NULL. The syntax that fails is \"column_name\" = NULL. The\n> following attachment was provided by -Jose'-. It contains a very small\n> enhancement to gram.y that will allow Access97 to work properly with sub\n> forms. Can this enhancement be added to release 6.5?\n> \n> <<gram.patch>> \n> Thanks, Michael\n> \n\n[Attachment, skipping...]\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 14 Mar 1999 00:14:18 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] parser enhancement request for 6.5" } ]
[ { "msg_contents": "The following function does not work:\n\nCREATE FUNCTION InstallPermissions(varchar, varchar) RETURNS int2 AS '\n\tDECLARE \n\t\toptions ALIAS FOR $1;\n\t\tusername ALIAS FOR $2;\nBEGIN\n\tFOR row IN select * from pg_tables LOOP\n \t \t REVOKE ALL ON row.tablename FROM username;\n\t\t GRANT options on row.tablename TO username;\n\tEND LOOP;\n\treturn 0;\n END; ' LANGUAGE 'plpgsql';\n\n\"select InstallPermissions('test1', 'test2);\" returns an error at \"select\".\nAny suggestions on how to eliminate the error?\n\nThanks, Michael\n", "msg_date": "Wed, 3 Mar 1999 16:12:54 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "for row in select loop question" }, { "msg_contents": ">\n> The following function does not work:\n>\n> CREATE FUNCTION InstallPermissions(varchar, varchar) RETURNS int2 AS '\n> DECLARE\n> options ALIAS FOR $1;\n> username ALIAS FOR $2;\n> BEGIN\n> FOR row IN select * from pg_tables LOOP\n> REVOKE ALL ON row.tablename FROM username;\n> GRANT options on row.tablename TO username;\n> END LOOP;\n> return 0;\n> END; ' LANGUAGE 'plpgsql';\n>\n> \"select InstallPermissions('test1', 'test2);\" returns an error at \"select\".\n> Any suggestions on how to eliminate the error?\n\n First you forgot to declare 'row'. Could be type record.\n\n Second REVOKE and GRANT are utility statements not supported\n for prepared SPI plans and thus PL/pgSQL currently cannot\n execute them.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 5 Mar 1999 10:59:48 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] for row in select loop question" } ]
[ { "msg_contents": ">From Tom Lane's horror story...\n\n>I spent an hour tracing through startup of 6.4.x, and I now understand\n>why the thing doesn't crash despite the horrible bugs in ShmemInitHash.\n>Read on, if you have a strong stomach.\n\nAre Tom Lane's fixes included in 6.4.3 beta? I think his findings are\nso important.\n---\nTatsuo Ishii\n", "msg_date": "Thu, 04 Mar 1999 07:24:26 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": true, "msg_subject": "Tom Lane's fixes in v6.4.3" }, { "msg_contents": "Tatsuo Ishii <[email protected]> writes:\n>> From Tom Lane's horror story...\n>> I spent an hour tracing through startup of 6.4.x, and I now understand\n>> why the thing doesn't crash despite the horrible bugs in ShmemInitHash.\n>> Read on, if you have a strong stomach.\n\n> Are Tom Lane's fixes included in 6.4.3 beta? I think his findings are\n> so important.\n\nI have made a patch for 6.4.x which I intend to commit into the REL6_4\ntree, but it hasn't gotten as much testing as I would like. The patch\nis attached if you care to try it for a while first. (These changes\nare already in the development tree, BTW.)\n\n\t\t\tregards, tom lane", "msg_date": "Wed, 03 Mar 1999 18:55:01 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Tom Lane's fixes in v6.4.3 " }, { "msg_contents": ">Tatsuo Ishii <[email protected]> writes:\n>>> From Tom Lane's horror story...\n>>> I spent an hour tracing through startup of 6.4.x, and I now understand\n>>> why the thing doesn't crash despite the horrible bugs in ShmemInitHash.\n>>> Read on, if you have a strong stomach.\n>\n>> Are Tom Lane's fixes included in 6.4.3 beta? I think his findings are\n>> so important.\n>\n>I have made a patch for 6.4.x which I intend to commit into the REL6_4\n>tree, but it hasn't gotten as much testing as I would like. The patch\n>is attached if you care to try it for a while first. (These changes\n>are already in the development tree, BTW.)\n\nThanks. Your patches work fine with fresh REL6_4 sources I got this\nmorning. One thing I noticed: when the backend runs out the\nsemaphores, postmatser dies with following messages:\n\nIpcSemaphoreCreate: semget failed (No space left on device) key=5432017, num=16, permission=600\nNOTICE: Message from PostgreSQL backend:\n\tThe Postmaster has informed me that some other backend died abnormally and possibly corrupted shared memory.\n\tI have rolled back the current transaction and am going to terminate your database system connection and exit.\n\tPlease reconnect to the database system and repeat your query.\n\nIs this normal?\n\nI thought postmatser tried to re-initialize the shared buffer and\nresume to the normal operation in this case.\n\nBTW, 6.4 tree does not have the max backend patch I posted. So even if \nthere are enough resources, the backend will crash if connections >\nMaxBackends.\n--\nTatsuo Ishii\n", "msg_date": "Thu, 04 Mar 1999 11:47:52 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Tom Lane's fixes in v6.4.3 " }, { "msg_contents": "When I tried to start postmaster as:\n\npostmaster -d 3 -B 1024\n\nI got a core dump:\n\nFindExec: searching PATH ...\nValidateBinary: can't stat \"/home/httpd/html/users/t-ishii/bin/postgres\"\nValidateBinary: can't stat \"/usr/local/bin/postgres\"\nValidateBinary: can't stat \"/bin/postgres\"\nValidateBinary: can't stat \"/usr/bin/postgres\"\nValidateBinary: can't stat \"/home/httpd/html/users/t-ishii/src/pgsql/postgresql-6.4.2/src/backend./postgres\"\nValidateBinary: can't stat \"/usr/X11R6/bin/postgres\"\nFindExec: found \"/usr/local/pgsql/bin/postgres\" using PATH\nbinding ShmemCreate(key=52e2c1, size=9859300)\nERROR: InitMultiLocks: couldnt initialize lock table\nQuit (core dumped)\n\nInitMultiLocks calls LockMethodTableInit. So I inspected\nLockMethodTableInit and found that it returned \nlockMethodTable->ctl->lockmethod with value 0 which made\nInitMultiLocks judge something went wrong.\nNote that ipcs -m -l sais:\n\nmax number of segments = 128\nmax seg size (kbytes) = 16384\nmax total shared memory (kbytes) = 16777216\nmin seg size (bytes) = 1\n\nSo there should be enogh shared mems. Also note that -B 1023 runs\nfine, but -B 1024 does not.\n\nAny idea?\n\nThis is 6.4.2 + Tom Lanes fix running Linux/Mips (kernel 2.0.33) with\n32MB memories.\n--\nTatsuo Ishii\n", "msg_date": "Thu, 04 Mar 1999 17:43:38 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Tom Lane's fixes in v6.4.3 " }, { "msg_contents": "Tatsuo Ishii <[email protected]> writes:\n> One thing I noticed: when the backend runs out the\n> semaphores, postmatser dies with following messages:\n\n> IpcSemaphoreCreate: semget failed (No space left on device) key=5432017, num=16, permission=600\n> NOTICE: Message from PostgreSQL backend:\n> \tThe Postmaster has informed me that some other backend died abnormally and possibly corrupted shared memory.\n> \tI have rolled back the current transaction and am going to terminate your database system connection and exit.\n> \tPlease reconnect to the database system and repeat your query.\n\n> Is this normal?\n\nYes, that's the behavior that we decided we'd better fix for 6.5.\n\nI think retrofitting the various MaxBackends-related changes into 6.4.x\nwould be risky --- the changes are fairly widespread and have not gotten\nall that much testing so far.\n\n\t\t\tregards, tom lane\n", "msg_date": "Thu, 04 Mar 1999 09:43:03 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Tom Lane's fixes in v6.4.3 " }, { "msg_contents": "> Tatsuo Ishii <[email protected]> writes:\n> > One thing I noticed: when the backend runs out the\n> > semaphores, postmatser dies with following messages:\n> \n> > IpcSemaphoreCreate: semget failed (No space left on device) key=5432017, num=16, permission=600\n> > NOTICE: Message from PostgreSQL backend:\n> > \tThe Postmaster has informed me that some other backend died abnormally and possibly corrupted shared memory.\n> > \tI have rolled back the current transaction and am going to terminate your database system connection and exit.\n> > \tPlease reconnect to the database system and repeat your query.\n> \n> > Is this normal?\n> \n> Yes, that's the behavior that we decided we'd better fix for 6.5.\n\nGlad to hear that.\n\n> I think retrofitting the various MaxBackends-related changes into 6.4.x\n> would be risky --- the changes are fairly widespread and have not gotten\n> all that much testing so far.\n\nOk. I will keep your patches for the case of having trouble with many\nbackends. I think it should be noted somewhere that 6.4.3 is not very\nstable with many backends (known bugs section?).\n--\nTatsuo Ishii\n", "msg_date": "Fri, 05 Mar 1999 10:08:33 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Tom Lane's fixes in v6.4.3 " }, { "msg_contents": "Tatsuo Ishii <[email protected]> writes:\n> When I tried to start postmaster as:\n> postmaster -d 3 -B 1024\n> I got a core dump:\n\nCan't duplicate that here, using either 6.4+fixes or current source.\nSome platform dependency involved perhaps??\n\nIt seems possible that this indicates some further bugs in the shared\nmemory allocation stuff, so I think it needs to be pursued.\n\n\t\t\tregards, tom lane\n", "msg_date": "Sat, 06 Mar 1999 18:52:34 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Tom Lane's fixes in v6.4.3 " } ]
[ { "msg_contents": "Hi,\n\n Just in case you'd like to see what I was talking about, I am attaching\nmy patch to src/interfaces/libpq/fe-exec.c to prevent utility functions\ncalled from SPI from locking up the client.\n\nJerry", "msg_date": "Thu, 4 Mar 1999 12:37:34 +0900", "msg_from": "\"Gerald L. Gay\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] libpq and SPI " }, { "msg_contents": "Applied.\n\n\n[Charset iso-8859-1 unsupported, filtering to ASCII...]\n> Hi,\n> \n> Just in case you'd like to see what I was talking about, I am attaching\n> my patch to src/interfaces/libpq/fe-exec.c to prevent utility functions\n> called from SPI from locking up the client.\n> \n> Jerry\n> \n\n[Attachment, skipping...]\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 14 Mar 1999 11:42:13 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI" }, { "msg_contents": "Bruce Momjian <[email protected]> writes:\n> Applied.\n>\n>> Just in case you'd like to see what I was talking about, I am attaching\n>> my patch to src/interfaces/libpq/fe-exec.c to prevent utility functions\n>> called from SPI from locking up the client.\n\nUh, I didn't actually believe that that patch was a good idea. Hacking\nlibpq to survive a protocol violation committed by the backend is *not*\na solution; the correct answer is to fix the backend. Otherwise we will\nhave to discover similar workarounds for other clients that do not\nuse libpq (ODBC, for example).\n\nPlease reverse out that patch until someone can find some time to look\nat the issue. (I will, if no one else does, but it would probably be\nmore efficient for someone who already knows something about SPI to\nfix it...)\n\n\t\t\tregards, tom lane\n", "msg_date": "Sun, 14 Mar 1999 12:31:17 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI " }, { "msg_contents": "Tom Lane wrote:\n> Please reverse out that patch until someone can find some time to look\n> at the issue. (I will, if no one else does, but it would probably be\n> more efficient for someone who already knows something about SPI to\n> fix it...)\n\nWhat is the problem? I'll research a SPI patch.\n\n:) Clark\n", "msg_date": "Sun, 14 Mar 1999 17:42:47 +0000", "msg_from": "Clark Evans <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI" }, { "msg_contents": "Reversed out. Attached is the patch.\n\n> Bruce Momjian <[email protected]> writes:\n> > Applied.\n> >\n> >> Just in case you'd like to see what I was talking about, I am attaching\n> >> my patch to src/interfaces/libpq/fe-exec.c to prevent utility functions\n> >> called from SPI from locking up the client.\n> \n> Uh, I didn't actually believe that that patch was a good idea. Hacking\n> libpq to survive a protocol violation committed by the backend is *not*\n> a solution; the correct answer is to fix the backend. Otherwise we will\n> have to discover similar workarounds for other clients that do not\n> use libpq (ODBC, for example).\n> \n> Please reverse out that patch until someone can find some time to look\n> at the issue. (I will, if no one else does, but it would probably be\n> more efficient for someone who already knows something about SPI to\n> fix it...)\n> \n> \t\t\tregards, tom lane\n> \n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n\n--- fe-exec.c.orig\tMon Feb 22 05:03:04 1999\n+++ fe-exec.c\tFri Mar 5 06:03:43 1999\n@@ -338,6 +338,7 @@\n parseInput(PGconn *conn)\n {\n \tchar\t\tid;\n+ static int pendingT = 0;\n \n \t/*\n \t * Loop to parse successive complete messages available in the buffer.\n@@ -406,7 +407,16 @@\n \t\t\t\t\t\t\t\t\t\t\t\t\t\t PGRES_COMMAND_OK);\n \t\t\t\t\tif (pqGets(conn->result->cmdStatus, CMDSTATUS_LEN, conn))\n \t\t\t\t\t\treturn;\n-\t\t\t\t\tconn->asyncStatus = PGASYNC_READY;\n+ if (pendingT) {\n+ /* Check the returned message */\n+ /* if it's a SELECT or FETCH in a pendingT case, */\n+ /* then it probably means no rows returned. */\n+ /* We clear pendingT in that case. */\n+ if ((strncmp(conn->result->cmdStatus, \"SELECT\", 6) == 0) ||\n+ (strncmp(conn->result->cmdStatus, \"FETCH\", 5) == 0))\n+ pendingT = 0;\n+ }\n+\t\t\t\t\tif (!pendingT) conn->asyncStatus = PGASYNC_READY;\n \t\t\t\t\tbreak;\n \t\t\t\tcase 'E':\t\t/* error return */\n \t\t\t\t\tif (pqGets(conn->errorMessage, ERROR_MSG_LENGTH, conn))\n@@ -416,10 +426,11 @@\n \t\t\t\t\t/* and build an error result holding the error message */\n \t\t\t\t\tconn->result = PQmakeEmptyPGresult(conn,\n \t\t\t\t\t\t\t\t\t\t\t\t\t PGRES_FATAL_ERROR);\n-\t\t\t\t\tconn->asyncStatus = PGASYNC_READY;\n+\t\t\t\t\tif (!pendingT) conn->asyncStatus = PGASYNC_READY;\n \t\t\t\t\tbreak;\n \t\t\t\tcase 'Z':\t\t/* backend is ready for new query */\n \t\t\t\t\tconn->asyncStatus = PGASYNC_IDLE;\n+ pendingT = 0;\n \t\t\t\t\tbreak;\n \t\t\t\tcase 'I':\t\t/* empty query */\n \t\t\t\t\t/* read and throw away the closing '\\0' */\n@@ -434,7 +445,7 @@\n \t\t\t\t\tif (conn->result == NULL)\n \t\t\t\t\t\tconn->result = PQmakeEmptyPGresult(conn,\n \t\t\t\t\t\t\t\t\t\t\t\t\t\t PGRES_EMPTY_QUERY);\n-\t\t\t\t\tconn->asyncStatus = PGASYNC_READY;\n+\t\t\t\t\tif (!pendingT) conn->asyncStatus = PGASYNC_READY;\n \t\t\t\t\tbreak;\n \t\t\t\tcase 'K':\t\t/* secret key data from the backend */\n \n@@ -455,11 +466,15 @@\n \t\t\t\t\tbreak;\n \t\t\t\tcase 'T':\t\t/* row descriptions (start of query\n \t\t\t\t\t\t\t\t * results) */\n+ if (pendingT) {\n+ DONOTICE(conn, \"Got second 'T' message!\\n\");\n+ }\n \t\t\t\t\tif (conn->result == NULL)\n \t\t\t\t\t{\n \t\t\t\t\t\t/* First 'T' in a query sequence */\n \t\t\t\t\t\tif (getRowDescriptions(conn))\n \t\t\t\t\t\t\treturn;\n+ pendingT = 1;\n \t\t\t\t\t}\n \t\t\t\t\telse\n \t\t\t\t\t{\n@@ -471,11 +486,13 @@\n \t\t\t\t\t\t * We stop parsing until the application accepts\n \t\t\t\t\t\t * the current result.\n \t\t\t\t\t\t */\n+ pendingT = 0;\n \t\t\t\t\t\tconn->asyncStatus = PGASYNC_READY;\n \t\t\t\t\t\treturn;\n \t\t\t\t\t}\n \t\t\t\t\tbreak;\n \t\t\t\tcase 'D':\t\t/* ASCII data tuple */\n+ pendingT = 0;\n \t\t\t\t\tif (conn->result != NULL)\n \t\t\t\t\t{\n \t\t\t\t\t\t/* Read another tuple of a normal query response */\n@@ -493,6 +510,7 @@\n \t\t\t\t\t}\n \t\t\t\t\tbreak;\n \t\t\t\tcase 'B':\t\t/* Binary data tuple */\n+ pendingT = 0;\n \t\t\t\t\tif (conn->result != NULL)\n \t\t\t\t\t{\n \t\t\t\t\t\t/* Read another tuple of a normal query response */\n@@ -510,12 +528,15 @@\n \t\t\t\t\t}\n \t\t\t\t\tbreak;\n \t\t\t\tcase 'G':\t\t/* Start Copy In */\n+ pendingT = 0;\n \t\t\t\t\tconn->asyncStatus = PGASYNC_COPY_IN;\n \t\t\t\t\tbreak;\n \t\t\t\tcase 'H':\t\t/* Start Copy Out */\n+ pendingT = 0;\n \t\t\t\t\tconn->asyncStatus = PGASYNC_COPY_OUT;\n \t\t\t\t\tbreak;\n \t\t\t\tdefault:\n+ pendingT = 0;\n \t\t\t\t\tsprintf(conn->errorMessage,\n \t\t\t\t\t\"unknown protocol character '%c' read from backend. \"\n \t\t\t\t\t\"(The protocol character is the first character the \"", "msg_date": "Sun, 14 Mar 1999 13:10:32 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI" }, { "msg_contents": "> What is the problem? I'll research a SPI patch.\n\nCheck the hackers thread (last month I think) titled \"libpq and SPI\";\nthe problem occurs when one submits a utility statement rather than\na plannable query via SPI. Apparently what is happening is that the\nbackend emits a 'T' message before it invokes the called statement\nand then emits a 'D' message afterwards --- so if the called statement\ncauses a 'C' message to come out, libpq gets unhappy. This seems\nto be clearly a violation of the FE/BE protocol to me, so I don't think\nit's libpq's fault.\n\nReasonable fixes might be to postpone the sending of 'T' till after\nthe invoked statement is executed, or to modify the traffic cop so\nthat a utility statement invoked from SPI doesn't send 'C'.\n\nI know a little bit about the parts of the backend that communicate with\nthe frontend, but nothing about SPI, so I'm not well prepared to solve\nthe problem by myself.\n\n\t\t\tregards, tom lane\n", "msg_date": "Sun, 14 Mar 1999 13:20:59 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI " }, { "msg_contents": "> Tom Lane wrote:\n> > Please reverse out that patch until someone can find some time to look\n> > at the issue. (I will, if no one else does, but it would probably be\n> > more efficient for someone who already knows something about SPI to\n> > fix it...)\n> \n> What is the problem? I'll research a SPI patch.\n> \n> :) Clark\n> \n\nI think someone already supplied such a patch. Update via cvs and try\nit.\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 14 Mar 1999 13:49:00 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI" }, { "msg_contents": "Hello all,\n\n> -----Original Message-----\n> From: [email protected]\n> [mailto:[email protected]]On Behalf Of Tom Lane\n> Sent: Monday, March 15, 1999 3:21 AM\n> To: Clark Evans\n> Cc: [email protected]\n> Subject: Re: [HACKERS] libpq and SPI \n> \n> \n> > What is the problem? I'll research a SPI patch.\n> \n> Check the hackers thread (last month I think) titled \"libpq and SPI\";\n> the problem occurs when one submits a utility statement rather than\n> a plannable query via SPI. Apparently what is happening is that the\n> backend emits a 'T' message before it invokes the called statement\n> and then emits a 'D' message afterwards --- so if the called statement\n> causes a 'C' message to come out, libpq gets unhappy. This seems\n> to be clearly a violation of the FE/BE protocol to me, so I don't think\n> it's libpq's fault.\n> \n> Reasonable fixes might be to postpone the sending of 'T' till after\n> the invoked statement is executed, or to modify the traffic cop so\n> that a utility statement invoked from SPI doesn't send 'C'.\n> \n> I know a little bit about the parts of the backend that communicate with\n> the frontend, but nothing about SPI, so I'm not well prepared to solve\n> the problem by myself.\n>\n\nProbably it's not the problem of SPI.\nSpecific utility commands(CREATE USER/ALTER USER/DROP USER\n/CREATE DATABASE/DROP DATABASE) break FE/BE protocol \nwhen they are executed inside the PostgreSQL function call.\n\nThey call pg_exec_query() (tcop/postgres.c) in their implementation. \nIn many cases the destination of pg_exec_query() is \"Remote\" \nwhich means that results are sent to frontend process. But the \nresults shouldn't be sent to the frontend directly inside the execution \nof PostgreSQL function.\n\nProcessUtility() (tcop/utility.c) function processes utility commands \nand takes the destination of command results as its second parameter. \nSo should we call pg_exec_query_dest(query_string, the destination \nparameter of ProcessUtility(), FALSE) instead of pg_exec_query(),\nshouldn't we ? \n\nThanks.\n\nHiroshi Inoue\[email protected]\n", "msg_date": "Mon, 15 Mar 1999 19:00:10 +0900", "msg_from": "\"Hiroshi Inoue\" <[email protected]>", "msg_from_op": false, "msg_subject": "RE: [HACKERS] libpq and SPI " }, { "msg_contents": ">\n> > What is the problem? I'll research a SPI patch.\n>\n> Check the hackers thread (last month I think) titled \"libpq and SPI\";\n> the problem occurs when one submits a utility statement rather than\n> a plannable query via SPI. Apparently what is happening is that the\n> backend emits a 'T' message before it invokes the called statement\n> and then emits a 'D' message afterwards --- so if the called statement\n> causes a 'C' message to come out, libpq gets unhappy. This seems\n> to be clearly a violation of the FE/BE protocol to me, so I don't think\n> it's libpq's fault.\n>\n> Reasonable fixes might be to postpone the sending of 'T' till after\n> the invoked statement is executed, or to modify the traffic cop so\n> that a utility statement invoked from SPI doesn't send 'C'.\n>\n> I know a little bit about the parts of the backend that communicate with\n> the frontend, but nothing about SPI, so I'm not well prepared to solve\n> the problem by myself.\n\n Another major problem of SPI on utilities must get fixed\n prior.\n\n SPI cannot save a prepared plan for utilities because\n copyObject() doesn't handle the utility part of Query. So it\n get's NULL and invoking such a saved plan will cause the\n backend to crash.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Mon, 15 Mar 1999 17:04:21 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI" }, { "msg_contents": "\nAre you ready to install the blinking map in the main site?\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 15 Mar 1999 11:53:56 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "developers globe" }, { "msg_contents": ">\n>\n> Are you ready to install the blinking map in the main site?\n\n Not yet.\n\n I still have problems with the incompatibilities between NS4\n and IE4. It's terrible.\n\n I have the whole map inside a centered table entry.\n\n If I don't put the main image (the map itself) into a\n diversion, I can get the x,y location of the image after\n loading in NS via the undocumented non-standard properties x\n and y. But then I'm unable to get the position in IE4.\n\n OTOH, if I put the image into a diversion with a style, then\n I can get the position in both browsers. But this time, NS4\n does very strange things if I resize the browser window (the\n image jumps out of table into the upper left corner of the\n window :-}).\n\n I'm still playing around with it locally. Hope to have a\n solution later today.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Mon, 15 Mar 1999 18:21:51 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: developers globe" }, { "msg_contents": "Hello all,\n\n> -----Original Message-----\n> From: [email protected]\n> [mailto:[email protected]]On Behalf Of Hiroshi Inoue\n> Sent: Monday, March 15, 1999 7:00 PM\n> To: Tom Lane; Clark Evans\n> Cc: [email protected]\n> Subject: RE: [HACKERS] libpq and SPI\n>\n> >\n> >\n> > > What is the problem? I'll research a SPI patch.\n> >\n> > Check the hackers thread (last month I think) titled \"libpq and SPI\";\n> > the problem occurs when one submits a utility statement rather than\n> > a plannable query via SPI. Apparently what is happening is that the\n> > backend emits a 'T' message before it invokes the called statement\n> > and then emits a 'D' message afterwards --- so if the called statement\n> > causes a 'C' message to come out, libpq gets unhappy. This seems\n> > to be clearly a violation of the FE/BE protocol to me, so I don't think\n> > it's libpq's fault.\n> >\n> > Reasonable fixes might be to postpone the sending of 'T' till after\n> > the invoked statement is executed, or to modify the traffic cop so\n> > that a utility statement invoked from SPI doesn't send 'C'.\n> >\n> > I know a little bit about the parts of the backend that communicate with\n> > the frontend, but nothing about SPI, so I'm not well prepared to solve\n> > the problem by myself.\n> >\n>\n> Probably it's not the problem of SPI.\n> Specific utility commands(CREATE USER/ALTER USER/DROP USER\n> /CREATE DATABASE/DROP DATABASE) break FE/BE protocol\n> when they are executed inside the PostgreSQL function call.\n>\n\nHere is a patch.\nI have changed to call pg_exec_query_dest() instead of pg_exec_query().\n\nThanks.\n\nHiroshi Inoue\[email protected]\n\n\n*** backend/tcop/utility.c.orig\tThu Feb 18 17:01:27 1999\n--- backend/tcop/utility.c\tTue Mar 16 09:25:07 1999\n***************\n*** 57,65 ****\n #include \"utils/syscache.h\"\n #endif\n\n! void\t\tDefineUser(CreateUserStmt *stmt);\n! void\t\tAlterUser(AlterUserStmt *stmt);\n! void\t\tRemoveUser(char *username);\n\n /* ----------------\n *\t\tCHECK_IF_ABORTED() is used to avoid doing unnecessary\n--- 57,65 ----\n #include \"utils/syscache.h\"\n #endif\n\n! void\t\tDefineUser(CreateUserStmt *stmt, CommandDest);\n! void\t\tAlterUser(AlterUserStmt *stmt, CommandDest);\n! void\t\tRemoveUser(char *username, CommandDest);\n\n /* ----------------\n *\t\tCHECK_IF_ABORTED() is used to avoid doing unnecessary\n***************\n*** 558,564 ****\n\n \t\t\t\tPS_SET_STATUS(commandTag = \"CREATEDB\");\n \t\t\t\tCHECK_IF_ABORTED();\n! \t\t\t\tcreatedb(stmt->dbname, stmt->dbpath, stmt->encoding);\n \t\t\t}\n \t\t\tbreak;\n\n--- 558,564 ----\n\n \t\t\t\tPS_SET_STATUS(commandTag = \"CREATEDB\");\n \t\t\t\tCHECK_IF_ABORTED();\n! \t\t\t\tcreatedb(stmt->dbname, stmt->dbpath, stmt->encoding, dest);\n \t\t\t}\n \t\t\tbreak;\n\n***************\n*** 568,574 ****\n\n \t\t\t\tPS_SET_STATUS(commandTag = \"DESTROYDB\");\n \t\t\t\tCHECK_IF_ABORTED();\n! \t\t\t\tdestroydb(stmt->dbname);\n \t\t\t}\n \t\t\tbreak;\n\n--- 568,574 ----\n\n \t\t\t\tPS_SET_STATUS(commandTag = \"DESTROYDB\");\n \t\t\t\tCHECK_IF_ABORTED();\n! \t\t\t\tdestroydb(stmt->dbname, dest);\n \t\t\t}\n \t\t\tbreak;\n\n***************\n*** 748,754 ****\n \t\t\tPS_SET_STATUS(commandTag = \"CREATE USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tDefineUser((CreateUserStmt *) parsetree);\n \t\t\tbreak;\n\n \t\tcase T_AlterUserStmt:\n--- 748,754 ----\n \t\t\tPS_SET_STATUS(commandTag = \"CREATE USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tDefineUser((CreateUserStmt *) parsetree, dest);\n \t\t\tbreak;\n\n \t\tcase T_AlterUserStmt:\n***************\n*** 755,761 ****\n \t\t\tPS_SET_STATUS(commandTag = \"ALTER USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tAlterUser((AlterUserStmt *) parsetree);\n \t\t\tbreak;\n\n \t\tcase T_DropUserStmt:\n--- 755,761 ----\n \t\t\tPS_SET_STATUS(commandTag = \"ALTER USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tAlterUser((AlterUserStmt *) parsetree, dest);\n \t\t\tbreak;\n\n \t\tcase T_DropUserStmt:\n***************\n*** 762,768 ****\n \t\t\tPS_SET_STATUS(commandTag = \"DROP USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tRemoveUser(((DropUserStmt *) parsetree)->user);\n \t\t\tbreak;\n\n \t\tcase T_LockStmt:\n--- 762,768 ----\n \t\t\tPS_SET_STATUS(commandTag = \"DROP USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tRemoveUser(((DropUserStmt *) parsetree)->user, dest);\n \t\t\tbreak;\n\n \t\tcase T_LockStmt:\n*** backend/commands/user.c.orig\tThu Feb 18 17:00:38 1999\n--- backend/commands/user.c\tTue Mar 16 09:50:09 1999\n***************\n*** 46,52 ****\n */\n static\n void\n! UpdatePgPwdFile(char *sql)\n {\n\n \tchar\t*filename,\n--- 46,52 ----\n */\n static\n void\n! UpdatePgPwdFile(char *sql, CommandDest dest)\n {\n\n \tchar\t*filename,\n***************\n*** 71,77 ****\n \tsnprintf(sql, SQL_LENGTH,\n \t\t\t\"copy %s to '%s' using delimiters %s\",\n \t\t\tShadowRelationName, tempname, CRYPT_PWD_FILE_SEPCHAR);\n! \tpg_exec_query(sql);\n \trename(tempname, filename);\n \tpfree((void *) tempname);\n\n--- 71,77 ----\n \tsnprintf(sql, SQL_LENGTH,\n \t\t\t\"copy %s to '%s' using delimiters %s\",\n \t\t\tShadowRelationName, tempname, CRYPT_PWD_FILE_SEPCHAR);\n! \tpg_exec_query_dest(sql, dest, false);\n \trename(tempname, filename);\n \tpfree((void *) tempname);\n\n***************\n*** 92,98 ****\n *---------------------------------------------------------------------\n */\n void\n! DefineUser(CreateUserStmt *stmt)\n {\n\n \tchar\t\t\t\t\t*pg_shadow,\n--- 92,98 ----\n *---------------------------------------------------------------------\n */\n void\n! DefineUser(CreateUserStmt *stmt, CommandDest dest)\n {\n\n \tchar\t\t\t\t\t*pg_shadow,\n***************\n*** 175,187 ****\n \t\t\tstmt->password ? stmt->password : \"''\",\n \t\t\tstmt->validUntil ? stmt->validUntil : \"\");\n\n! \tpg_exec_query(sql);\n\n \t/*\n \t * Add the stuff here for groups.\n \t */\n\n! \tUpdatePgPwdFile(sql);\n\n \t/*\n \t * This goes after the UpdatePgPwdFile to be certain that two backends\n--- 175,187 ----\n \t\t\tstmt->password ? stmt->password : \"''\",\n \t\t\tstmt->validUntil ? stmt->validUntil : \"\");\n\n! \tpg_exec_query_dest(sql, dest, false);\n\n \t/*\n \t * Add the stuff here for groups.\n \t */\n\n! \tUpdatePgPwdFile(sql, dest);\n\n \t/*\n \t * This goes after the UpdatePgPwdFile to be certain that two backends\n***************\n*** 196,202 ****\n\n\n extern void\n! AlterUser(AlterUserStmt *stmt)\n {\n\n \tchar\t\t\t*pg_shadow,\n--- 196,202 ----\n\n\n extern void\n! AlterUser(AlterUserStmt *stmt, CommandDest dest)\n {\n\n \tchar\t\t\t*pg_shadow,\n***************\n*** 282,292 ****\n\n \tsnprintf(sql, SQL_LENGTH, \"%s where usename = '%s'\", sql, stmt->user);\n\n! \tpg_exec_query(sql);\n\n \t/* do the pg_group stuff here */\n\n! \tUpdatePgPwdFile(sql);\n\n \tUnlockRelation(pg_shadow_rel, AccessExclusiveLock);\n \theap_close(pg_shadow_rel);\n--- 282,292 ----\n\n \tsnprintf(sql, SQL_LENGTH, \"%s where usename = '%s'\", sql, stmt->user);\n\n! \tpg_exec_query_dest(sql, dest, false);\n\n \t/* do the pg_group stuff here */\n\n! \tUpdatePgPwdFile(sql, dest);\n\n \tUnlockRelation(pg_shadow_rel, AccessExclusiveLock);\n \theap_close(pg_shadow_rel);\n***************\n*** 297,303 ****\n\n\n extern void\n! RemoveUser(char *user)\n {\n\n \tchar\t *pg_shadow;\n--- 297,303 ----\n\n\n extern void\n! RemoveUser(char *user, CommandDest dest)\n {\n\n \tchar\t *pg_shadow;\n***************\n*** 390,396 ****\n \t\telog(NOTICE, \"Dropping database %s\", dbase[ndbase]);\n \t\tsnprintf(sql, SQL_LENGTH, \"drop database %s\", dbase[ndbase]);\n \t\tpfree((void *) dbase[ndbase]);\n! \t\tpg_exec_query(sql);\n \t}\n \tif (dbase)\n \t\tpfree((void *) dbase);\n--- 390,396 ----\n \t\telog(NOTICE, \"Dropping database %s\", dbase[ndbase]);\n \t\tsnprintf(sql, SQL_LENGTH, \"drop database %s\", dbase[ndbase]);\n \t\tpfree((void *) dbase[ndbase]);\n! \t\tpg_exec_query_dest(sql, dest, false);\n \t}\n \tif (dbase)\n \t\tpfree((void *) dbase);\n***************\n*** 418,426 ****\n \t */\n \tsnprintf(sql, SQL_LENGTH,\n \t\t\t\"delete from %s where usename = '%s'\", ShadowRelationName, user);\n! \tpg_exec_query(sql);\n\n! \tUpdatePgPwdFile(sql);\n\n \tUnlockRelation(pg_shadow_rel, AccessExclusiveLock);\n \theap_close(pg_shadow_rel);\n--- 418,426 ----\n \t */\n \tsnprintf(sql, SQL_LENGTH,\n \t\t\t\"delete from %s where usename = '%s'\", ShadowRelationName, user);\n! \tpg_exec_query_dest(sql, dest, false);\n\n! \tUpdatePgPwdFile(sql, dest);\n\n \tUnlockRelation(pg_shadow_rel, AccessExclusiveLock);\n \theap_close(pg_shadow_rel);\n*** backend/commands/dbcommands.c.orig\tThu Feb 18 17:00:36 1999\n--- backend/commands/dbcommands.c\tTue Mar 16 09:36:33 1999\n***************\n*** 24,30 ****\n #include \"catalog/catname.h\"\n #include \"catalog/pg_database.h\"\n #include \"catalog/pg_shadow.h\"\n- #include \"commands/dbcommands.h\"\n #include \"fmgr.h\"\n #include \"miscadmin.h\"\t\t\t/* for DataDir */\n #include \"storage/bufmgr.h\"\n--- 24,29 ----\n***************\n*** 31,36 ****\n--- 30,36 ----\n #include \"storage/fd.h\"\n #include \"storage/lmgr.h\"\n #include \"tcop/tcopprot.h\"\n+ #include \"commands/dbcommands.h\"\n #include \"utils/rel.h\"\n #include \"utils/syscache.h\"\n\n***************\n*** 42,48 ****\n static void stop_vacuum(char *dbpath, char *dbname);\n\n void\n! createdb(char *dbname, char *dbpath, int encoding)\n {\n \tOid\t\t\tdb_id;\n \tint4\t\tuser_id;\n--- 42,48 ----\n static void stop_vacuum(char *dbpath, char *dbname);\n\n void\n! createdb(char *dbname, char *dbpath, int encoding, CommandDest dest)\n {\n \tOid\t\t\tdb_id;\n \tint4\t\tuser_id;\n***************\n*** 87,97 ****\n \t\t\t\"insert into pg_database (datname, datdba, encoding, datpath)\"\n \t\t\t\" values ('%s', '%d', '%d', '%s');\", dbname, user_id, encoding,\nloc);\n\n! \tpg_exec_query(buf);\n }\n\n void\n! destroydb(char *dbname)\n {\n \tint4\t\tuser_id;\n \tOid\t\t\tdb_id;\n--- 87,97 ----\n \t\t\t\"insert into pg_database (datname, datdba, encoding, datpath)\"\n \t\t\t\" values ('%s', '%d', '%d', '%s');\", dbname, user_id, encoding,\nloc);\n\n! \tpg_exec_query_dest(buf, dest, false);\n }\n\n void\n! destroydb(char *dbname, CommandDest dest)\n {\n \tint4\t\tuser_id;\n \tOid\t\t\tdb_id;\n***************\n*** 123,129 ****\n \t */\n \tsnprintf(buf, 512,\n \t\t\t\"delete from pg_database where pg_database.oid = \\'%d\\'::oid\",\ndb_id);\n! \tpg_exec_query(buf);\n\n \t/*\n \t * remove the data directory. If the DELETE above failed, this will\n--- 123,129 ----\n \t */\n \tsnprintf(buf, 512,\n \t\t\t\"delete from pg_database where pg_database.oid = \\'%d\\'::oid\",\ndb_id);\n! \tpg_exec_query_dest(buf ,dest, false);\n\n \t/*\n \t * remove the data directory. If the DELETE above failed, this will\n*** include/commands/user.h.orig\tThu Feb 18 17:01:49 1999\n--- include/commands/user.h\tTue Mar 16 09:23:01 1999\n***************\n*** 10,17 ****\n #ifndef USER_H\n #define USER_H\n\n! extern void DefineUser(CreateUserStmt *stmt);\n! extern void AlterUser(AlterUserStmt *stmt);\n! extern void RemoveUser(char *user);\n\n #endif\t /* USER_H */\n--- 10,17 ----\n #ifndef USER_H\n #define USER_H\n\n! extern void DefineUser(CreateUserStmt *stmt, CommandDest);\n! extern void AlterUser(AlterUserStmt *stmt, CommandDest);\n! extern void RemoveUser(char *user, CommandDest);\n\n #endif\t /* USER_H */\n*** include/commands/dbcommands.h.orig\tThu Feb 18 17:01:48 1999\n--- include/commands/dbcommands.h\tTue Mar 16 09:23:52 1999\n***************\n*** 19,25 ****\n */\n #define SIGKILLDAEMON1\tSIGTERM\n\n! extern void createdb(char *dbname, char *dbpath, int encoding);\n! extern void destroydb(char *dbname);\n\n #endif\t /* DBCOMMANDS_H */\n--- 19,25 ----\n */\n #define SIGKILLDAEMON1\tSIGTERM\n\n! extern void createdb(char *dbname, char *dbpath, int encoding,\nCommandDest);\n! extern void destroydb(char *dbname, CommandDest);\n\n #endif\t /* DBCOMMANDS_H */\n\n\n\n\n\n", "msg_date": "Tue, 16 Mar 1999 11:57:37 +0900", "msg_from": "\"Hiroshi Inoue\" <[email protected]>", "msg_from_op": false, "msg_subject": "RE: [HACKERS] libpq and SPI " }, { "msg_contents": "\nApplied.\n\n\n---------------------------------------------------------------------------\nHello all,\n\n> -----Original Message-----\n> From: [email protected]\n> [mailto:[email protected]]On Behalf Of Hiroshi Inoue\n> Sent: Monday, March 15, 1999 7:00 PM\n> To: Tom Lane; Clark Evans\n> Cc: [email protected]\n> Subject: RE: [HACKERS] libpq and SPI\n>\n> >\n> >\n> > > What is the problem? I'll research a SPI patch.\n> >\n> > Check the hackers thread (last month I think) titled \"libpq and SPI\";\n> > the problem occurs when one submits a utility statement rather than\n> > a plannable query via SPI. Apparently what is happening is that the\n> > backend emits a 'T' message before it invokes the called statement\n> > and then emits a 'D' message afterwards --- so if the called statement\n> > causes a 'C' message to come out, libpq gets unhappy. This seems\n> > to be clearly a violation of the FE/BE protocol to me, so I don't think\n> > it's libpq's fault.\n> >\n> > Reasonable fixes might be to postpone the sending of 'T' till after\n> > the invoked statement is executed, or to modify the traffic cop so\n> > that a utility statement invoked from SPI doesn't send 'C'.\n> >\n> > I know a little bit about the parts of the backend that communicate with\n> > the frontend, but nothing about SPI, so I'm not well prepared to solve\n> > the problem by myself.\n> >\n>\n> Probably it's not the problem of SPI.\n> Specific utility commands(CREATE USER/ALTER USER/DROP USER\n> /CREATE DATABASE/DROP DATABASE) break FE/BE protocol\n> when they are executed inside the PostgreSQL function call.\n>\n\nHere is a patch.\nI have changed to call pg_exec_query_dest() instead of pg_exec_query().\n\nThanks.\n\nHiroshi Inoue\[email protected]\n\n\n*** backend/tcop/utility.c.orig\tThu Feb 18 17:01:27 1999\n--- backend/tcop/utility.c\tTue Mar 16 09:25:07 1999\n***************\n*** 57,65 ****\n #include \"utils/syscache.h\"\n #endif\n\n! void\t\tDefineUser(CreateUserStmt *stmt);\n! void\t\tAlterUser(AlterUserStmt *stmt);\n! void\t\tRemoveUser(char *username);\n\n /* ----------------\n *\t\tCHECK_IF_ABORTED() is used to avoid doing unnecessary\n--- 57,65 ----\n #include \"utils/syscache.h\"\n #endif\n\n! void\t\tDefineUser(CreateUserStmt *stmt, CommandDest);\n! void\t\tAlterUser(AlterUserStmt *stmt, CommandDest);\n! void\t\tRemoveUser(char *username, CommandDest);\n\n /* ----------------\n *\t\tCHECK_IF_ABORTED() is used to avoid doing unnecessary\n***************\n*** 558,564 ****\n\n \t\t\t\tPS_SET_STATUS(commandTag = \"CREATEDB\");\n \t\t\t\tCHECK_IF_ABORTED();\n! \t\t\t\tcreatedb(stmt->dbname, stmt->dbpath, stmt->encoding);\n \t\t\t}\n \t\t\tbreak;\n\n--- 558,564 ----\n\n \t\t\t\tPS_SET_STATUS(commandTag = \"CREATEDB\");\n \t\t\t\tCHECK_IF_ABORTED();\n! \t\t\t\tcreatedb(stmt->dbname, stmt->dbpath, stmt->encoding, dest);\n \t\t\t}\n \t\t\tbreak;\n\n***************\n*** 568,574 ****\n\n \t\t\t\tPS_SET_STATUS(commandTag = \"DESTROYDB\");\n \t\t\t\tCHECK_IF_ABORTED();\n! \t\t\t\tdestroydb(stmt->dbname);\n \t\t\t}\n \t\t\tbreak;\n\n--- 568,574 ----\n\n \t\t\t\tPS_SET_STATUS(commandTag = \"DESTROYDB\");\n \t\t\t\tCHECK_IF_ABORTED();\n! \t\t\t\tdestroydb(stmt->dbname, dest);\n \t\t\t}\n \t\t\tbreak;\n\n***************\n*** 748,754 ****\n \t\t\tPS_SET_STATUS(commandTag = \"CREATE USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tDefineUser((CreateUserStmt *) parsetree);\n \t\t\tbreak;\n\n \t\tcase T_AlterUserStmt:\n--- 748,754 ----\n \t\t\tPS_SET_STATUS(commandTag = \"CREATE USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tDefineUser((CreateUserStmt *) parsetree, dest);\n \t\t\tbreak;\n\n \t\tcase T_AlterUserStmt:\n***************\n*** 755,761 ****\n \t\t\tPS_SET_STATUS(commandTag = \"ALTER USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tAlterUser((AlterUserStmt *) parsetree);\n \t\t\tbreak;\n\n \t\tcase T_DropUserStmt:\n--- 755,761 ----\n \t\t\tPS_SET_STATUS(commandTag = \"ALTER USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tAlterUser((AlterUserStmt *) parsetree, dest);\n \t\t\tbreak;\n\n \t\tcase T_DropUserStmt:\n***************\n*** 762,768 ****\n \t\t\tPS_SET_STATUS(commandTag = \"DROP USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tRemoveUser(((DropUserStmt *) parsetree)->user);\n \t\t\tbreak;\n\n \t\tcase T_LockStmt:\n--- 762,768 ----\n \t\t\tPS_SET_STATUS(commandTag = \"DROP USER\");\n \t\t\tCHECK_IF_ABORTED();\n\n! \t\t\tRemoveUser(((DropUserStmt *) parsetree)->user, dest);\n \t\t\tbreak;\n\n \t\tcase T_LockStmt:\n*** backend/commands/user.c.orig\tThu Feb 18 17:00:38 1999\n--- backend/commands/user.c\tTue Mar 16 09:50:09 1999\n***************\n*** 46,52 ****\n */\n static\n void\n! UpdatePgPwdFile(char *sql)\n {\n\n \tchar\t*filename,\n--- 46,52 ----\n */\n static\n void\n! UpdatePgPwdFile(char *sql, CommandDest dest)\n {\n\n \tchar\t*filename,\n***************\n*** 71,77 ****\n \tsnprintf(sql, SQL_LENGTH,\n \t\t\t\"copy %s to '%s' using delimiters %s\",\n \t\t\tShadowRelationName, tempname, CRYPT_PWD_FILE_SEPCHAR);\n! \tpg_exec_query(sql);\n \trename(tempname, filename);\n \tpfree((void *) tempname);\n\n--- 71,77 ----\n \tsnprintf(sql, SQL_LENGTH,\n \t\t\t\"copy %s to '%s' using delimiters %s\",\n \t\t\tShadowRelationName, tempname, CRYPT_PWD_FILE_SEPCHAR);\n! \tpg_exec_query_dest(sql, dest, false);\n \trename(tempname, filename);\n \tpfree((void *) tempname);\n\n***************\n*** 92,98 ****\n *---------------------------------------------------------------------\n */\n void\n! DefineUser(CreateUserStmt *stmt)\n {\n\n \tchar\t\t\t\t\t*pg_shadow,\n--- 92,98 ----\n *---------------------------------------------------------------------\n */\n void\n! DefineUser(CreateUserStmt *stmt, CommandDest dest)\n {\n\n \tchar\t\t\t\t\t*pg_shadow,\n***************\n*** 175,187 ****\n \t\t\tstmt->password ? stmt->password : \"''\",\n \t\t\tstmt->validUntil ? stmt->validUntil : \"\");\n\n! \tpg_exec_query(sql);\n\n \t/*\n \t * Add the stuff here for groups.\n \t */\n\n! \tUpdatePgPwdFile(sql);\n\n \t/*\n \t * This goes after the UpdatePgPwdFile to be certain that two backends\n--- 175,187 ----\n \t\t\tstmt->password ? stmt->password : \"''\",\n \t\t\tstmt->validUntil ? stmt->validUntil : \"\");\n\n! \tpg_exec_query_dest(sql, dest, false);\n\n \t/*\n \t * Add the stuff here for groups.\n \t */\n\n! \tUpdatePgPwdFile(sql, dest);\n\n \t/*\n \t * This goes after the UpdatePgPwdFile to be certain that two backends\n***************\n*** 196,202 ****\n\n\n extern void\n! AlterUser(AlterUserStmt *stmt)\n {\n\n \tchar\t\t\t*pg_shadow,\n--- 196,202 ----\n\n\n extern void\n! AlterUser(AlterUserStmt *stmt, CommandDest dest)\n {\n\n \tchar\t\t\t*pg_shadow,\n***************\n*** 282,292 ****\n\n \tsnprintf(sql, SQL_LENGTH, \"%s where usename = '%s'\", sql, stmt->user);\n\n! \tpg_exec_query(sql);\n\n \t/* do the pg_group stuff here */\n\n! \tUpdatePgPwdFile(sql);\n\n \tUnlockRelation(pg_shadow_rel, AccessExclusiveLock);\n \theap_close(pg_shadow_rel);\n--- 282,292 ----\n\n \tsnprintf(sql, SQL_LENGTH, \"%s where usename = '%s'\", sql, stmt->user);\n\n! \tpg_exec_query_dest(sql, dest, false);\n\n \t/* do the pg_group stuff here */\n\n! \tUpdatePgPwdFile(sql, dest);\n\n \tUnlockRelation(pg_shadow_rel, AccessExclusiveLock);\n \theap_close(pg_shadow_rel);\n***************\n*** 297,303 ****\n\n\n extern void\n! RemoveUser(char *user)\n {\n\n \tchar\t *pg_shadow;\n--- 297,303 ----\n\n\n extern void\n! RemoveUser(char *user, CommandDest dest)\n {\n\n \tchar\t *pg_shadow;\n***************\n*** 390,396 ****\n \t\telog(NOTICE, \"Dropping database %s\", dbase[ndbase]);\n \t\tsnprintf(sql, SQL_LENGTH, \"drop database %s\", dbase[ndbase]);\n \t\tpfree((void *) dbase[ndbase]);\n! \t\tpg_exec_query(sql);\n \t}\n \tif (dbase)\n \t\tpfree((void *) dbase);\n--- 390,396 ----\n \t\telog(NOTICE, \"Dropping database %s\", dbase[ndbase]);\n \t\tsnprintf(sql, SQL_LENGTH, \"drop database %s\", dbase[ndbase]);\n \t\tpfree((void *) dbase[ndbase]);\n! \t\tpg_exec_query_dest(sql, dest, false);\n \t}\n \tif (dbase)\n \t\tpfree((void *) dbase);\n***************\n*** 418,426 ****\n \t */\n \tsnprintf(sql, SQL_LENGTH,\n \t\t\t\"delete from %s where usename = '%s'\", ShadowRelationName, user);\n! \tpg_exec_query(sql);\n\n! \tUpdatePgPwdFile(sql);\n\n \tUnlockRelation(pg_shadow_rel, AccessExclusiveLock);\n \theap_close(pg_shadow_rel);\n--- 418,426 ----\n \t */\n \tsnprintf(sql, SQL_LENGTH,\n \t\t\t\"delete from %s where usename = '%s'\", ShadowRelationName, user);\n! \tpg_exec_query_dest(sql, dest, false);\n\n! \tUpdatePgPwdFile(sql, dest);\n\n \tUnlockRelation(pg_shadow_rel, AccessExclusiveLock);\n \theap_close(pg_shadow_rel);\n*** backend/commands/dbcommands.c.orig\tThu Feb 18 17:00:36 1999\n--- backend/commands/dbcommands.c\tTue Mar 16 09:36:33 1999\n***************\n*** 24,30 ****\n #include \"catalog/catname.h\"\n #include \"catalog/pg_database.h\"\n #include \"catalog/pg_shadow.h\"\n- #include \"commands/dbcommands.h\"\n #include \"fmgr.h\"\n #include \"miscadmin.h\"\t\t\t/* for DataDir */\n #include \"storage/bufmgr.h\"\n--- 24,29 ----\n***************\n*** 31,36 ****\n--- 30,36 ----\n #include \"storage/fd.h\"\n #include \"storage/lmgr.h\"\n #include \"tcop/tcopprot.h\"\n+ #include \"commands/dbcommands.h\"\n #include \"utils/rel.h\"\n #include \"utils/syscache.h\"\n\n***************\n*** 42,48 ****\n static void stop_vacuum(char *dbpath, char *dbname);\n\n void\n! createdb(char *dbname, char *dbpath, int encoding)\n {\n \tOid\t\t\tdb_id;\n \tint4\t\tuser_id;\n--- 42,48 ----\n static void stop_vacuum(char *dbpath, char *dbname);\n\n void\n! createdb(char *dbname, char *dbpath, int encoding, CommandDest dest)\n {\n \tOid\t\t\tdb_id;\n \tint4\t\tuser_id;\n***************\n*** 87,97 ****\n \t\t\t\"insert into pg_database (datname, datdba, encoding, datpath)\"\n \t\t\t\" values ('%s', '%d', '%d', '%s');\", dbname, user_id, encoding,\nloc);\n\n! \tpg_exec_query(buf);\n }\n\n void\n! destroydb(char *dbname)\n {\n \tint4\t\tuser_id;\n \tOid\t\t\tdb_id;\n--- 87,97 ----\n \t\t\t\"insert into pg_database (datname, datdba, encoding, datpath)\"\n \t\t\t\" values ('%s', '%d', '%d', '%s');\", dbname, user_id, encoding,\nloc);\n\n! \tpg_exec_query_dest(buf, dest, false);\n }\n\n void\n! destroydb(char *dbname, CommandDest dest)\n {\n \tint4\t\tuser_id;\n \tOid\t\t\tdb_id;\n***************\n*** 123,129 ****\n \t */\n \tsnprintf(buf, 512,\n \t\t\t\"delete from pg_database where pg_database.oid = \\'%d\\'::oid\",\ndb_id);\n! \tpg_exec_query(buf);\n\n \t/*\n \t * remove the data directory. If the DELETE above failed, this will\n--- 123,129 ----\n \t */\n \tsnprintf(buf, 512,\n \t\t\t\"delete from pg_database where pg_database.oid = \\'%d\\'::oid\",\ndb_id);\n! \tpg_exec_query_dest(buf ,dest, false);\n\n \t/*\n \t * remove the data directory. If the DELETE above failed, this will\n*** include/commands/user.h.orig\tThu Feb 18 17:01:49 1999\n--- include/commands/user.h\tTue Mar 16 09:23:01 1999\n***************\n*** 10,17 ****\n #ifndef USER_H\n #define USER_H\n\n! extern void DefineUser(CreateUserStmt *stmt);\n! extern void AlterUser(AlterUserStmt *stmt);\n! extern void RemoveUser(char *user);\n\n #endif\t /* USER_H */\n--- 10,17 ----\n #ifndef USER_H\n #define USER_H\n\n! extern void DefineUser(CreateUserStmt *stmt, CommandDest);\n! extern void AlterUser(AlterUserStmt *stmt, CommandDest);\n! extern void RemoveUser(char *user, CommandDest);\n\n #endif\t /* USER_H */\n*** include/commands/dbcommands.h.orig\tThu Feb 18 17:01:48 1999\n--- include/commands/dbcommands.h\tTue Mar 16 09:23:52 1999\n***************\n*** 19,25 ****\n */\n #define SIGKILLDAEMON1\tSIGTERM\n\n! extern void createdb(char *dbname, char *dbpath, int encoding);\n! extern void destroydb(char *dbname);\n\n #endif\t /* DBCOMMANDS_H */\n--- 19,25 ----\n */\n #define SIGKILLDAEMON1\tSIGTERM\n\n! extern void createdb(char *dbname, char *dbpath, int encoding,\nCommandDest);\n! extern void destroydb(char *dbname, CommandDest);\n\n #endif\t /* DBCOMMANDS_H */\n\n\n\n\n\n\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 15 Mar 1999 22:24:25 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI" }, { "msg_contents": "\nIs this fixed?\n\n\n> > What is the problem? I'll research a SPI patch.\n> \n> Check the hackers thread (last month I think) titled \"libpq and SPI\";\n> the problem occurs when one submits a utility statement rather than\n> a plannable query via SPI. Apparently what is happening is that the\n> backend emits a 'T' message before it invokes the called statement\n> and then emits a 'D' message afterwards --- so if the called statement\n> causes a 'C' message to come out, libpq gets unhappy. This seems\n> to be clearly a violation of the FE/BE protocol to me, so I don't think\n> it's libpq's fault.\n> \n> Reasonable fixes might be to postpone the sending of 'T' till after\n> the invoked statement is executed, or to modify the traffic cop so\n> that a utility statement invoked from SPI doesn't send 'C'.\n> \n> I know a little bit about the parts of the backend that communicate with\n> the frontend, but nothing about SPI, so I'm not well prepared to solve\n> the problem by myself.\n> \n> \t\t\tregards, tom lane\n> \n> \n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 9 May 1999 16:10:52 -0400 (EDT)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] libpq and SPI" } ]
[ { "msg_contents": "i have been playing with the inet and cidr types, and i have noticed a couple\nproblems, and would entertain suggestions as to how to fix them without\nnecessarily breaking how other people might be using them.\n\n- a table with an element of type inet, will show \"0.0.0.0/0\" as \"00/0\"\n\ni suspect this can be fixed in src/backend/adt/network.c or some such.\n\n- when creating a table with either type inet or type cidr as a primary, unique\n key, the elements \"198.68.123.0/24\" and \"198.68.123.0/27\" are considered\n equal.\n\ni considered editing src/backend/adt/network.c and changing the various\ncomparison functions to take the subnet bits into account, but i wasn't sure\nwhat the proper fix is.\n\nmy feeling is to make \"/27\" to be greater than \"/24\", such that when fetching\nin order, the \"/24\" will come first.\n\nif i make such changes to the source, will it break other people's code if\nthe changes get added to the core source (6.4.3 or 6.5)?\n\n-- \n[ Jim Mercer Reptilian Research [email protected] +1 416 410-5633 ]\n[ The telephone, for those of you who have forgotten, was a commonly used ]\n[ communications technology in the days before electronic mail. ]\n[ They're still easy to find in most large cities. -- Nathaniel Borenstein ]\n", "msg_date": "Wed, 3 Mar 1999 23:40:40 -0500 (EST)", "msg_from": "[email protected] (Jim Mercer)", "msg_from_op": true, "msg_subject": "inet and cidr type problems" }, { "msg_contents": "\nThis is still an open item\n\n\n> i have been playing with the inet and cidr types, and i have noticed a couple\n> problems, and would entertain suggestions as to how to fix them without\n> necessarily breaking how other people might be using them.\n> \n> - a table with an element of type inet, will show \"0.0.0.0/0\" as \"00/0\"\n> \n> i suspect this can be fixed in src/backend/adt/network.c or some such.\n> \n> - when creating a table with either type inet or type cidr as a primary, unique\n> key, the elements \"198.68.123.0/24\" and \"198.68.123.0/27\" are considered\n> equal.\n> \n> i considered editing src/backend/adt/network.c and changing the various\n> comparison functions to take the subnet bits into account, but i wasn't sure\n> what the proper fix is.\n> \n> my feeling is to make \"/27\" to be greater than \"/24\", such that when fetching\n> in order, the \"/24\" will come first.\n> \n> if i make such changes to the source, will it break other people's code if\n> the changes get added to the core source (6.4.3 or 6.5)?\n> \n> -- \n> [ Jim Mercer Reptilian Research [email protected] +1 416 410-5633 ]\n> [ The telephone, for those of you who have forgotten, was a commonly used ]\n> [ communications technology in the days before electronic mail. ]\n> [ They're still easy to find in most large cities. -- Nathaniel Borenstein ]\n> \n> \n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 9 May 1999 11:33:06 -0400 (EDT)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [GENERAL] inet and cidr type problems" }, { "msg_contents": "\nI just ran a test in 6.5beta, and this still a problem:\n\t\n\ttest=> create table t (x inet primary key);\n\tNOTICE: CREATE TABLE/PRIMARY KEY will create implicit index t_pkey for\n\ttable t\n\tCREATE\n\ttest=> insert into t values ('198.68.123.0/24');\n\tINSERT 18604 1\n\ttest=> insert into t values ('198.68.123.0/27');\n\tERROR: Cannot insert a duplicate key into a unique index\n\n\n> i have been playing with the inet and cidr types, and i have noticed a couple\n> problems, and would entertain suggestions as to how to fix them without\n> necessarily breaking how other people might be using them.\n> \n> - a table with an element of type inet, will show \"0.0.0.0/0\" as \"00/0\"\n> \n> i suspect this can be fixed in src/backend/adt/network.c or some such.\n> \n> - when creating a table with either type inet or type cidr as a primary, unique\n> key, the elements \"198.68.123.0/24\" and \"198.68.123.0/27\" are considered\n> equal.\n> \n> i considered editing src/backend/adt/network.c and changing the various\n> comparison functions to take the subnet bits into account, but i wasn't sure\n> what the proper fix is.\n> \n> my feeling is to make \"/27\" to be greater than \"/24\", such that when fetching\n> in order, the \"/24\" will come first.\n> \n> if i make such changes to the source, will it break other people's code if\n> the changes get added to the core source (6.4.3 or 6.5)?\n> \n> -- \n> [ Jim Mercer Reptilian Research [email protected] +1 416 410-5633 ]\n> [ The telephone, for those of you who have forgotten, was a commonly used ]\n> [ communications technology in the days before electronic mail. ]\n> [ They're still easy to find in most large cities. -- Nathaniel Borenstein ]\n> \n> \n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 9 May 1999 11:51:58 -0400 (EDT)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [GENERAL] inet and cidr type problems" }, { "msg_contents": "> i have been playing with the inet and cidr types, and i have noticed a couple\n> problems, and would entertain suggestions as to how to fix them without\n> necessarily breaking how other people might be using them.\n> \n> - a table with an element of type inet, will show \"0.0.0.0/0\" as \"00/0\"\n\nThis was fixed in 6.5.\n\n> i suspect this can be fixed in src/backend/adt/network.c or some such.\n> \n> - when creating a table with either type inet or type cidr as a primary, unique\n> key, the elements \"198.68.123.0/24\" and \"198.68.123.0/27\" are considered\n> equal.\n\nThis has been addressed in 6.5 too.\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 6 Jul 1999 22:06:27 -0400 (EDT)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [GENERAL] inet and cidr type problems" }, { "msg_contents": "> > i have been playing with the inet and cidr types, and i have noticed a couple\n> > problems, and would entertain suggestions as to how to fix them without\n> > necessarily breaking how other people might be using them.\n> > \n> > - a table with an element of type inet, will show \"0.0.0.0/0\" as \"00/0\"\n> \n> This was fixed in 6.5.\n> \n> > i suspect this can be fixed in src/backend/adt/network.c or some such.\n> > \n> > - when creating a table with either type inet or type cidr as a primary, unique\n> > key, the elements \"198.68.123.0/24\" and \"198.68.123.0/27\" are considered\n> > equal.\n> \n> This has been addressed in 6.5 too.\n\ncool.\n\ni'm not sure how to query the bug list, so can you tell me if the following\nproblems have been addressed?\n\n\n- ALTER TABLE tab_name RENAME TO new_name;\n\n this command does not deal with extended tables (my name for when\n a table physically exceeds 2Gig, on unix, and creates a second (third)\n extent in the .../data/base/dbname dir.\n\n- VACUUM VERBOSE;\n\n has problems with extended tables, something like \"cannot truncate\n extended file\".\n\n\n-- \n[ Jim Mercer Reptilian Research [email protected] +1 416 410-5633 ]\n[ The telephone, for those of you who have forgotten, was a commonly used ]\n[ communications technology in the days before electronic mail. ]\n[ They're still easy to find in most large cities. -- Nathaniel Borenstein ]\n", "msg_date": "Wed, 7 Jul 1999 02:10:38 -0400 (EDT)", "msg_from": "[email protected] (Jim Mercer)", "msg_from_op": true, "msg_subject": "Re: [GENERAL] inet and cidr type problems" }, { "msg_contents": "> > This has been addressed in 6.5 too.\n> \n> cool.\n> \n> i'm not sure how to query the bug list, so can you tell me if the following\n> problems have been addressed?\n> \n> \n> - ALTER TABLE tab_name RENAME TO new_name;\n> \n> this command does not deal with extended tables (my name for when\n> a table physically exceeds 2Gig, on unix, and creates a second (third)\n> extent in the .../data/base/dbname dir.\n\nYes, I believe this is fixed.\n\n> \n> - VACUUM VERBOSE;\n> \n> has problems with extended tables, something like \"cannot truncate\n> extended file\".\n> \n\nI think this fix will be in 6.5.1, which should be next week.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 7 Jul 1999 03:44:22 -0400 (EDT)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [GENERAL] inet and cidr type problems" } ]
[ { "msg_contents": "Hello all,\n\n>\n> As to the 2nd case I reported,it was probably introduced by my\n> patch. The patch was made to solve other Aggregate bugs.\n> get_agg_tlist_references() is used to reconstruct aggs member\n> node of Agg type nodes as SS_pull_subPlan() does in CopyPlan\n> Fields(). (The name of function was set_agg_tlist_references()\n> originally. Probably Bruce changed it).\n> We may be able to patch by following related code closely.\n> But it seems strange that copying objects requires such\n> a complicated procedure.\n>\n\nI made a patch for the 2nd case I reported.\nI applied my trial implementation of copyObject() partially to\nthe relation between Agg and Aggref type nodes. This patch\nsolves both this bug and an old bug solved by my old patch\n[Subject:SPI_prepare() doesn't work well ? ].\n\nThanks in advance.\n\nHiroshi Inoue\[email protected]\n\n*** backend/nodes/copyfuncs.c.orig\tTue Feb 23 17:00:26 1999\n--- backend/nodes/copyfuncs.c\tThu Mar 4 16:32:21 1999\n***************\n*** 31,37 ****\n--- 31,78 ----\n #include \"storage/lmgr.h\"\n #include \"optimizer/planmain.h\"\n\n+ #include \"access/xact.h\"\n+\n /*\n+ *\tLocal structures for maintaining the list of\n+ *\t\t(old -> new) pairs\n+ *\t\tduring copyObject operations\n+ */\n+\n+ typedef\n+ struct\n+ {\n+ \tconst void\t*old;\n+ \tvoid\t\t*new;\n+ } old_new_pair;\n+\n+ typedef\n+ struct\n+ {\n+ \tint\tmaxcnt;\n+ \tint\tcount;\n+ \told_new_pair old_new[1];\n+ } old_new_Hash;\n+\n+ typedef\n+ struct\n+ {\n+ \tbool\t\tinuse;\n+ \tTransactionId\texecXID;\n+ \tunsigned long\talloconce;\n+ \tList\t\t*first;\n+ \tList\t\t*last;\n+ } copyContext;\n+\n+ static\tcopyContext common_context = {false, 0, 0, NIL, NIL};\n+\n+ static bool appendToCopyContext(const void *, void *, copyContext *);\n+ static void initializeCopyContext(copyContext *, int);\n+ static void releaseCopyContext(copyContext *);\n+ static void *alreadyCopiedTo(const void *, const copyContext *);\n+ static bool isContextStillAlive(const copyContext *);\n+\n+ /*\n * listCopy\n *\t this copy function only copies the \"lcons-cells\" of the list but not\n *\t its contents. (good for list of pointers as well as list of\nintegers).\n***************\n*** 472,483 ****\n static Agg *\n _copyAgg(Agg *from)\n {\n \tAgg\t\t *newnode = makeNode(Agg);\n\n \tCopyPlanFields((Plan *) from, (Plan *) newnode);\n\n! \tnewnode->aggs = get_agg_tlist_references(newnode);\n\n \treturn newnode;\n }\n\n--- 513,535 ----\n static Agg *\n _copyAgg(Agg *from)\n {\n+ \tbool\t\tinit_here = false;\n+ \tcopyContext\t*copycontext = &common_context;\n \tAgg\t\t *newnode = makeNode(Agg);\n\n+ \tif (!isContextStillAlive(copycontext))\n+ \t{\n+ \t\tinitializeCopyContext(copycontext,0);\n+ \t\tinit_here = true;\n+ \t}\n+\n \tCopyPlanFields((Plan *) from, (Plan *) newnode);\n\n! \tNode_Copy(from, newnode, aggs);\n\n+ \tif (init_here)\n+ \t\treleaseCopyContext(copycontext);\n+\n \treturn newnode;\n }\n\n***************\n*** 871,878 ****\n static Aggref *\n _copyAggref(Aggref *from)\n {\n! \tAggref\t *newnode = makeNode(Aggref);\n\n \t/* ----------------\n \t *\tcopy remainder of node\n \t * ----------------\n--- 923,941 ----\n static Aggref *\n _copyAggref(Aggref *from)\n {\n! \tcopyContext\t*copycontext = &common_context;\n! \tAggref\t \t*newnode;\n\n+ \tif (copycontext->inuse)\n+ \t{\n+ \t\tnewnode = (Aggref *)alreadyCopiedTo(from, copycontext);\n+ \t\tif (newnode)\treturn newnode;\n+ \t}\n+\n+ \tnewnode = makeNode(Aggref);\n+ \tif (copycontext->inuse)\n+ \t\tappendToCopyContext(from, newnode, copycontext);\n+\n \t/* ----------------\n \t *\tcopy remainder of node\n \t * ----------------\n***************\n*** 1868,1871 ****\n--- 1931,2032 ----\n \t\t\tbreak;\n \t}\n \treturn retval;\n+ }\n+\n+ /*\n+ *\tstatic functions for maintaining the list of\n+ *\t\t(old -> new) pairs\n+ *\t\tduring copyObject operations\n+ */\n+ #define\t_Count_Per_Elem_\t127\n+ static void initializeCopyContext(copyContext *copycontext, int alloconce)\n+ {\n+ \tcopycontext->inuse = true;\n+ \tcopycontext->execXID = GetCurrentTransactionId();\n+ \tif (alloconce>0)\n+ \t\tcopycontext->alloconce = alloconce;\n+ \telse\n+ \t\tcopycontext->alloconce = _Count_Per_Elem_;\n+ \tcopycontext->first = copycontext->last = NIL;\n+ }\n+\n+ static void releaseCopyContext(copyContext *copycontext)\n+ {\n+ \tList\t\t*l;\n+\n+ \tif (!copycontext)\treturn;\n+ \tforeach(l, copycontext->first)\n+ \t\tpfree(lfirst(l));\n+ \tfreeList(copycontext->first);\n+ \tcopycontext->inuse = false;\n+ \tcopycontext->alloconce = 0;\n+ \tStoreInvalidTransactionId(&copycontext->execXID);\n+ \tcopycontext->first = copycontext->last = NIL;\n+ }\n+\n+ static bool appendToCopyContext(const void *old, void *new, copyContext\n*copycontext)\n+ {\n+ \told_new_Hash\t*copyHash;\n+ \tList\t\t*newl;\n+ \tbool\t\tnewList = false;\n+\n+ \tif (!copycontext)\t\treturn false;\n+ \tif ((!old) || (!new))\t\treturn false;\n+ \tif (!copycontext->last)\n+ \t\tnewList = true;\n+ \telse\n+ \t{\n+ \t\tcopyHash = (old_new_Hash *)lfirst(copycontext->last);\n+ \t\tif (copyHash->count >= copyHash->maxcnt)\n+ \t\t\tnewList = true;\n+ \t}\n+ \tif (newList)\n+ \t{\n+ \t\tcopyHash = (old_new_Hash *) palloc(\n+ \t\t\tsizeof(old_new_Hash) + sizeof(old_new_pair) *\n+ \t\t\t\t(copycontext->alloconce - 1) );\n+ \t\tif (!copyHash)\t\treturn false;\n+ \t\tcopyHash->maxcnt = copycontext->alloconce;\n+ \t\tcopyHash->count = 0;\n+ \t\tif (copycontext->last)\n+ \t\t{\n+ \t\t\tnewl = lcons(copyHash, NIL);\n+ \t\t\tlnext(copycontext->last) = newl;\n+ \t\t\tcopycontext->last = newl;\n+ \t\t}\n+ \t\telse\n+ \t\t{\n+ \t\t\tnewl = lcons(copyHash, NIL);\n+ \t\t\tcopycontext->first = copycontext->last = newl;\n+ \t\t}\n+ \t}\n+\n+ \tcopyHash->old_new[copyHash->count].old = old;\n+ \tcopyHash->old_new[copyHash->count].new = new;\n+ \tcopyHash->count++;\n+ \treturn true;\n+ }\n+ static void *alreadyCopiedTo(const void *obj, const copyContext\n*copycontext)\n+ {\n+ \tList\t\t\t*l;\n+ \tint\t\t\ti;\n+ \tconst old_new_Hash\t*copyHash;\n+\n+ \tif (!copycontext)\treturn NULL;\n+ \tforeach(l, copycontext->first)\n+ \t{\n+ \t\tcopyHash = (const old_new_Hash *)lfirst(l);\n+ \t\tfor (i=0; i<copyHash->count; i++)\n+ \t\t{\n+ \t\t\tif ( copyHash->old_new[i].old == obj )\n+ \t\t\t\treturn copyHash->old_new[i].new;\n+ \t\t}\n+ \t}\n+ \treturn NULL;\n+ }\n+\n+ static bool isContextStillAlive(const copyContext *copycontext)\n+ {\n+ \treturn ( copycontext->inuse &&\n+ \tTransactionIdIsCurrentTransactionId(copycontext->execXID) ) ;\n }\n\n", "msg_date": "Thu, 4 Mar 1999 19:00:10 +0900", "msg_from": "\"Hiroshi Inoue\" <[email protected]>", "msg_from_op": true, "msg_subject": "RE: [HACKERS] copyObject() ? again " } ]
[ { "msg_contents": "I've seen several postings about this, but no resolution. I've\nfiled another bug report.\n\nIf someone has a fix for this, let me know.\n\nThanks.\n\n_______ \n\n\n============================================================================\n POSTGRESQL BUG REPORT TEMPLATE\n============================================================================\n\n\nYour name : David Favor\nYour email address : [email protected]\n\nCategory : runtime: back-end\nSeverity : critical\n\nSummary: destroydb fails with ERROR: typeidTypeRelid: Invalid type - oid = 0\n\nSystem Configuration\n--------------------\n Operating System : AIX 432\n\n PostgreSQL version : 6.4.2\n\n Compiler used : gcc version egcs-2.93.08 19990214\n\nHardware:\n---------\nAIX aasc24 3 4 00FFFFFF4C00 (Power PC)\n\nVersions of other tools:\n------------------------\ngmake - 3.77\nbison - 1.25\nflex - 2.5.4\n\n--------------------------------------------------------------------------\n\nProblem Description:\n--------------------\ntemplate1=> create database foo;\nCREATEDB\ntemplate1=> drop database foo;\nERROR: typeidTypeRelid: Invalid type - oid = 0\nERROR: typeidTypeRelid: Invalid type - oid = 0\ntemplate1=> \\q\n", "msg_date": "Thu, 04 Mar 1999 05:26:43 -0600", "msg_from": "\"David R. Favor\" <[email protected]>", "msg_from_op": true, "msg_subject": "PostgreSQL 6.4.2 on AIX 4.3.2: typeidTypeRelid error" } ]
[ { "msg_contents": "I've seen several postings about this, but no resolution. I've\nfiled another bug report.\n\nIf someone has a fix for this, let me know.\n\nThanks.\n\n_______ \n\n\n============================================================================\n POSTGRESQL BUG REPORT TEMPLATE\n============================================================================\n\n\nYour name : David Favor\nYour email address : [email protected]\n\nCategory : runtime: back-end\nSeverity : critical\n\nSummary: destroydb fails with ERROR: typeidTypeRelid: Invalid type - oid = 0\n\nSystem Configuration\n--------------------\n Operating System : AIX 432\n\n PostgreSQL version : 6.4.2\n\n Compiler used : gcc version egcs-2.93.08 19990214\n\nHardware:\n---------\nAIX aasc24 3 4 00FFFFFF4C00 (Power PC)\n\nVersions of other tools:\n------------------------\ngmake - 3.77\nbison - 1.25\nflex - 2.5.4\n\n--------------------------------------------------------------------------\n\nProblem Description:\n--------------------\ntemplate1=> create database foo;\nCREATEDB\ntemplate1=> drop database foo;\nERROR: typeidTypeRelid: Invalid type - oid = 0\nERROR: typeidTypeRelid: Invalid type - oid = 0\ntemplate1=> \\q\n\n", "msg_date": "Thu, 4 Mar 1999 05:31:04 -0600", "msg_from": "\"David R. Favor\" <[email protected]>", "msg_from_op": true, "msg_subject": "PostgreSQL 6.4.2 on AIX 4.3.2: typeidTypeRelid error" }, { "msg_contents": "> I've seen several postings about this, but no resolution. I've\n> filed another bug report.\n> Problem Description:\n> --------------------\n> template1=> create database foo;\n> CREATEDB\n> template1=> drop database foo;\n> ERROR: typeidTypeRelid: Invalid type - oid = 0\n> ERROR: typeidTypeRelid: Invalid type - oid = 0\n> template1=> \\q\n\nYou have a damaged database afaik. Can you reproduce this with a newly\ninstalled (from initdb) database? If you can shrink down a test case to\nsomething others can look at then we'd have a chance to try fixing it.\nOtherwise, you will just need to work around it.\n\nfwiw the \"several postings\" don't ring a bell (but I'm not on the\nGeneral list if that's where they appeared).\n\n - Tom\n", "msg_date": "Thu, 04 Mar 1999 14:05:03 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL 6.4.2 on AIX 4.3.2: typeidTypeRelid error" }, { "msg_contents": "(back on-list, OK?)\n\n> > > Problem Description:\n> > > --------------------\n> > > template1=> create database foo;\n> > > CREATEDB\n> > > template1=> drop database foo;\n> > > ERROR: typeidTypeRelid: Invalid type - oid = 0\n> > > ERROR: typeidTypeRelid: Invalid type - oid = 0\n> > > template1=> \\q\n> > You have a damaged database afaik. Can you reproduce this with a \n> > newly installed (from initdb) database?\n> This was done with a fresh database, meaning:\n> 1) rm -rf /usr/local/pgsql/data\n> 2) initdb -u postgres\n> 3) postmaster -i &\n> 4) psql -d template1\n> 5) createdb foo;\n> 6) drop database foo; <--- error occurs here\n> This is can be reproduced consistently. Every create/drop results in \n> the same error.\n> Right now I'm recompiling everything with xlc 3.6.4 (latest native AIX \n> compiler). I should know in a bit whether the problem relates to gcc.\n> One thing that I did have to do was #ifdef out the definitions of \n> int(8|16|32|64) as AIX 4.3.x already defines these. I've left in \n> -qhalt=w, to ensure that all warnings are cleaned up.\n\nWell, that's a simple test case all right :)\n\nWithout knowing more, my guess is that it is AIX-specific, and specific\nto your version and/or compiler. We have (had?) at least a few other\nactive developers on AIX, and afaik they are not seeing this on their\ninstallations. I'll guess that they are not as current as you are on\nversions of OS and compilers.\n\nYou say that AIX already defined int32, etc. Are those definitions\nexactly compatible with those in the Postgres sources? If so, then\nperhaps you can submit patches which do a conditional definition in the\nPostgres code (e.g. #ifndef int32...).\n\nHave you compiled with assert checking turned on? Have you compiled\nwithout optimizations (-O0)? Do other operations on the database seem to\nwork, such as table creation or row insertion? I'll be interested in\nseeing what you find when using gcc.\n\nRegards.\n\n - Tom\n", "msg_date": "Thu, 04 Mar 1999 14:53:32 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: PostgreSQL 6.4.2 on AIX 4.3.2: typeidTypeRelid error" } ]
[ { "msg_contents": "Hello!\n\n A week ago I reported this, but haven't got any good help. I am trying\nto repeat.\n\n Postgres 6.4.2 on Solaris. Query:\nSELECT p.subsec_id\n FROM positions p, central cn, shops sh, districts d\n WHERE cn.pos_id = p.pos_id AND d.city_id = 2\n AND cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n AND cn.date_i >= current_date - '7 days'::timespan\n\n I am not publishing here my tables, but the structure is obvious. There\nis central database, which only knows shop_id, and shop only knows its\ndistrict, and district knows city. I want to count distinct p.subsec_id for\none city.\n With the query, postgres eats all memory and dies.\n\n I played with the query, and found I can remove (AND d.city_id = 2). The\nquery executes pretty fast (my database is small). I tried to get\nd.city_id:\n\nSELECT p.subsec_id, d.city_id\n FROM positions p, central cn, shops sh, districts d\n WHERE cn.pos_id = p.pos_id\n AND cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n AND cn.date_i >= current_date - '7 days'::timespan\n\nbut postgres eats all memory and dies :(((\n\n Memory leak? Other bugs?\n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Thu, 4 Mar 1999 14:49:46 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Not enough memory for complex join" }, { "msg_contents": "Oleg Broytmann wrote:\n> \n> I played with the query, and found I can remove (AND d.city_id = 2). The\n> query executes pretty fast (my database is small). I tried to get\n> d.city_id:\n> \n> SELECT p.subsec_id, d.city_id\n> FROM positions p, central cn, shops sh, districts d\n> WHERE cn.pos_id = p.pos_id\n> AND cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n> AND cn.date_i >= current_date - '7 days'::timespan\n> \n> but postgres eats all memory and dies :(((\n\nWhat does EXPLAIN say ?\n\n--------------------\nHannu\n", "msg_date": "Thu, 04 Mar 1999 15:14:25 +0200", "msg_from": "Hannu Krosing <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Not enough memory for complex join" }, { "msg_contents": "On Thu, 4 Mar 1999, Hannu Krosing wrote:\n> > I played with the query, and found I can remove (AND d.city_id = 2). The\n> > query executes pretty fast (my database is small). I tried to get\n> > d.city_id:\n> > \n> > SELECT p.subsec_id, d.city_id\n> > FROM positions p, central cn, shops sh, districts d\n> > WHERE cn.pos_id = p.pos_id\n> > AND cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n> > AND cn.date_i >= current_date - '7 days'::timespan\n> > \n> > but postgres eats all memory and dies :(((\n> \n> What does EXPLAIN say ?\n\nEXPLAIN SELECT p.subsec_id, d.city_id\n FROM positions p, central cn, shops sh, districts d\n WHERE cn.pos_id = p.pos_id\n AND cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n AND cn.date_i >= current_date - '7 days'::timespan\n;\nNOTICE: QUERY PLAN:\n\nNested Loop (cost=0.00 size=1 width=18)\n -> Nested Loop (cost=0.00 size=1 width=14)\n -> Merge Join (cost=0.00 size=1 width=10)\n -> Seq Scan (cost=0.00 size=0 width=0)\n -> Sort (cost=0.00 size=0 width=0)\n -> Seq Scan on districts d (cost=0.00 size=0 width=4)\n -> Seq Scan (cost=0.00 size=0 width=0)\n -> Sort (cost=0.00 size=0 width=0)\n -> Seq Scan on shops sh (cost=0.00 size=0 width=6)\n -> Seq Scan on central cn (cost=0.00 size=0 width=4)\n -> Seq Scan on positions p (cost=0.00 size=0 width=4)\n\nEXPLAIN\n\n\n> --------------------\n> Hannu\n> \n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Thu, 4 Mar 1999 16:15:53 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Not enough memory for complex join" }, { "msg_contents": "Oleg Broytmann wrote:\n> \n> Nested Loop (cost=0.00 size=1 width=18)\n> -> Nested Loop (cost=0.00 size=1 width=14)\n> -> Merge Join (cost=0.00 size=1 width=10)\n> -> Seq Scan (cost=0.00 size=0 width=0)\n> -> Sort (cost=0.00 size=0 width=0)\n> -> Seq Scan on districts d (cost=0.00 size=0 width=4)\n> -> Seq Scan (cost=0.00 size=0 width=0)\n> -> Sort (cost=0.00 size=0 width=0)\n> -> Seq Scan on shops sh (cost=0.00 size=0 width=6)\n> -> Seq Scan on central cn (cost=0.00 size=0 width=4)\n> -> Seq Scan on positions p (cost=0.00 size=0 width=4)\n ^^^^^^\nvacuum...\n\nVadim\n", "msg_date": "Fri, 05 Mar 1999 09:03:00 +0700", "msg_from": "Vadim Mikheev <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Not enough memory for complex join" }, { "msg_contents": "Hi!\n\nOn Fri, 5 Mar 1999, Vadim Mikheev wrote:\n> Oleg Broytmann wrote:\n> > \n> > Nested Loop (cost=0.00 size=1 width=18)\n> > -> Nested Loop (cost=0.00 size=1 width=14)\n> > -> Merge Join (cost=0.00 size=1 width=10)\n> > -> Seq Scan (cost=0.00 size=0 width=0)\n> > -> Sort (cost=0.00 size=0 width=0)\n> > -> Seq Scan on districts d (cost=0.00 size=0 width=4)\n> > -> Seq Scan (cost=0.00 size=0 width=0)\n> > -> Sort (cost=0.00 size=0 width=0)\n> > -> Seq Scan on shops sh (cost=0.00 size=0 width=6)\n> > -> Seq Scan on central cn (cost=0.00 size=0 width=4)\n> > -> Seq Scan on positions p (cost=0.00 size=0 width=4)\n> ^^^^^^\n> vacuum...\n\n I didn't think it could be of any help. I have a copy of this database\non my local computer. I dump db on server and put it on local computer\nevery other day, so I thiink VACUUM is unneccessary here.\n Anyway, I tried to VACUUM the db. No, the query didn't execute -\npostgres ate all memory and died.\n\n Any other idea, anyone?\n\n> Vadim\n\nOleg.\n---- \n Oleg Broytmann National Research Surgery Centre http://sun.med.ru/~phd/\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Fri, 5 Mar 1999 12:50:49 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Not enough memory for complex join" }, { "msg_contents": "On Fri, 5 Mar 1999, Oleg Broytmann wrote:\n\n> Hi!\n> \n> On Fri, 5 Mar 1999, Vadim Mikheev wrote:\n> > Oleg Broytmann wrote:\n> > > \n> > > Nested Loop (cost=0.00 size=1 width=18)\n> > > -> Nested Loop (cost=0.00 size=1 width=14)\n> > > -> Merge Join (cost=0.00 size=1 width=10)\n> > > -> Seq Scan (cost=0.00 size=0 width=0)\n> > > -> Sort (cost=0.00 size=0 width=0)\n> > > -> Seq Scan on districts d (cost=0.00 size=0 width=4)\n> > > -> Seq Scan (cost=0.00 size=0 width=0)\n> > > -> Sort (cost=0.00 size=0 width=0)\n> > > -> Seq Scan on shops sh (cost=0.00 size=0 width=6)\n> > > -> Seq Scan on central cn (cost=0.00 size=0 width=4)\n> > > -> Seq Scan on positions p (cost=0.00 size=0 width=4)\n> > ^^^^^^\n> > vacuum...\n> \n> I didn't think it could be of any help. I have a copy of this database\n> on my local computer. I dump db on server and put it on local computer\n> every other day, so I thiink VACUUM is unneccessary here.\n\nTry 'vacuum analyze'...vacuum, rom my understanding, just cleans out the\ndatabase of old records...reloading the db from scratch effectively has\nthat already done. 'vacuum analyze' adjusts statistics that don't get\nchanged on a load, that determins, to a large extet, how the optimizaer\nruns things...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Fri, 5 Mar 1999 13:42:31 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Not enough memory for complex join" }, { "msg_contents": "Oleg Broytmann wrote:\n> \n> Hi!\n> \n> On Fri, 5 Mar 1999, Vadim Mikheev wrote:\n> > Oleg Broytmann wrote:\n> > >\n> > > Nested Loop (cost=0.00 size=1 width=18)\n> > > -> Nested Loop (cost=0.00 size=1 width=14)\n> > > -> Merge Join (cost=0.00 size=1 width=10)\n> > > -> Seq Scan (cost=0.00 size=0 width=0)\n> > > -> Sort (cost=0.00 size=0 width=0)\n> > > -> Seq Scan on districts d (cost=0.00 size=0 width=4)\n> > > -> Seq Scan (cost=0.00 size=0 width=0)\n> > > -> Sort (cost=0.00 size=0 width=0)\n> > > -> Seq Scan on shops sh (cost=0.00 size=0 width=6)\n> > > -> Seq Scan on central cn (cost=0.00 size=0 width=4)\n> > > -> Seq Scan on positions p (cost=0.00 size=0 width=4)\n> > ^^^^^^\n> > vacuum...\n> \n> I didn't think it could be of any help. I have a copy of this database\n> on my local computer. I dump db on server and put it on local computer\n> every other day, so I thiink VACUUM is unneccessary here.\n> Anyway, I tried to VACUUM the db. No, the query didn't execute -\n> postgres ate all memory and died.\n\nEXPLAIN after vacuum?\n\nThere was MergerJoin before vacuum => 2 sorts, but sorting eats\nsome memory, this is why I suggested vacuum..\n\nAnd, btw, please re-post your query/table definition...\n\nVadim\n", "msg_date": "Sat, 06 Mar 1999 16:57:10 +0700", "msg_from": "Vadim Mikheev <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Not enough memory for complex join" }, { "msg_contents": "Hello!\n\nOn Fri, 5 Mar 1999, The Hermit Hacker wrote:\n> > > vacuum...\n> > \n> > I didn't think it could be of any help. I have a copy of this database\n> > on my local computer. I dump db on server and put it on local computer\n> > every other day, so I thiink VACUUM is unneccessary here.\n> \n> Try 'vacuum analyze'...vacuum, rom my understanding, just cleans out the\n> database of old records...reloading the db from scratch effectively has\n> that already done. 'vacuum analyze' adjusts statistics that don't get\n> changed on a load, that determins, to a large extet, how the optimizaer\n> runs things...\n\n Ha, then I stuck on another bug. VACUUM ANALYZE failed on glibc2 with\n--enable-locale. I reported this in \"VACUUM ANALYZE problem\" thread. Look\ninto mail archive for February :(((\n\n> Marc G. Fournier \n> Systems Administrator @ hub.org \n> primary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Tue, 9 Mar 1999 11:59:17 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Not enough memory for complex join" } ]
[ { "msg_contents": "Does anyone know where I can get a license number for the legacy 2.0\ngenealogy software package?\n\n", "msg_date": "Thu, 04 Mar 1999 13:04:52 +0000", "msg_from": "Brad <[email protected]>", "msg_from_op": true, "msg_subject": "Legacy" } ]
[ { "msg_contents": "\nAny chance that we can get that into v6.4.3? Is/was it simply a matter of\nchanging the max value, or something completely different?\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 4 Mar 1999 09:54:22 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "2gig limit fix..." } ]
[ { "msg_contents": "\nIs now available for testing...\n\nUnder FreeBSD 3.0-STABLE, it compiles cleanly, with the fullowing failed\nregression tests:\n\nint8 .. failed\nfloat8 .. failed\ngeometry .. failed\ncreate_function_2 .. failed\ntriggers .. failed\nsanity_check .. failed\nmisc .. failed\nalter_table .. failed\nplpgsql .. failed\n\nWill be looking into those, but I believe those ones are pretty\nstandard...\n\nWould like to get v6.4.3 out the door over the next couple of days, so\nplease test it on the various other platforms ...\n\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 4 Mar 1999 10:39:57 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "v6.4.3beta3 ..." }, { "msg_contents": "On Thu, 4 Mar 1999, The Hermit Hacker wrote:\n\n> \n> Is now available for testing...\n> \n> Under FreeBSD 3.0-STABLE, it compiles cleanly, with the fullowing failed\n> regression tests:\n> \n> int8 .. failed\n\nI'm wondering if this could have something to do with the %qd vs %lld\nthing. IIRC on the FreeBSD-Hackers list there was some chat about \nthat recently and FreeBSD treats %lld as %ld.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Thu, 4 Mar 1999 11:23:30 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] v6.4.3beta3 ..." }, { "msg_contents": "\nAlready fixed :)\n\n\nOn Thu, 4 Mar 1999, Vince Vielhaber wrote:\n\n> On Thu, 4 Mar 1999, The Hermit Hacker wrote:\n> \n> > \n> > Is now available for testing...\n> > \n> > Under FreeBSD 3.0-STABLE, it compiles cleanly, with the fullowing failed\n> > regression tests:\n> > \n> > int8 .. failed\n> \n> I'm wondering if this could have something to do with the %qd vs %lld\n> thing. IIRC on the FreeBSD-Hackers list there was some chat about \n> that recently and FreeBSD treats %lld as %ld.\n> \n> Vince.\n> -- \n> ==========================================================================\n> Vince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n> # include <std/disclaimers.h> TEAM-OS2\n> Online Campground Directory http://www.camping-usa.com\n> Online Giftshop Superstore http://www.cloudninegifts.com\n> ==========================================================================\n> \n> \n> \n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 4 Mar 1999 13:58:14 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] v6.4.3beta3 ..." }, { "msg_contents": "> Is now available for testing...\n\nWhere can I get that?\nftp://ftp.postgresql.org/pub/postgresql.snapshot.tar.gz ?\n\n> Under FreeBSD 3.0-STABLE, it compiles cleanly, with the fullowing failed\n> regression tests:\n> \n> int8 .. failed\n> float8 .. failed\n> geometry .. failed\n\nI'm not sure above but...\n\n> create_function_2 .. failed\n> triggers .. failed\n> sanity_check .. failed\n> misc .. failed\n> alter_table .. failed\n> plpgsql .. failed\n> \n> Will be looking into those, but I believe those ones are pretty\n> standard...\n\nI think those ones have been already fixed by a port maintainer. If\nyou are interested, please let me know. I will manage to contact to\nhim.\n--\nTatsuo Ishii\n", "msg_date": "Fri, 05 Mar 1999 10:12:38 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] v6.4.3beta3 ... " } ]
[ { "msg_contents": "I was able to fix this problem by recompiling everything using\nxlc 3.6.4 (AIX native compiler). This leads me to believe there\nis some evil bug in gcc.\n\nI have no idea how to pursue this. For now I will just use xlc.\n\nIf anyone wants to try and solve this and needs help, I will keep\nthe gcc build tree I have and give any assistance requested.\n\nThanks.\n", "msg_date": "Thu, 4 Mar 1999 08:54:36 -0600", "msg_from": "\"David R. Favor\" <[email protected]>", "msg_from_op": true, "msg_subject": "Fixed! PostgreSQL 6.4.2 on AIX 4.3.2: typeidTypeRelid error" }, { "msg_contents": "\"David R. Favor\" <[email protected]> writes:\n> I was able to fix this problem by recompiling everything using\n> xlc 3.6.4 (AIX native compiler). This leads me to believe there\n> is some evil bug in gcc.\n\n... where apparently \"gcc\" = \"egcs-2.93.08 19990214\".\n\nI'm not sure about the stability of egcs releases; do you have a\nproduction release there or just a beta?\n\nYou might have better luck with gcc 2.8.1 or even 2.7.2. (I've\nbeen using 2.7.2.2 for several years with nary a bug, though\nadmittedly I don't do anything with templates or exceptions...)\n\n\t\t\tregards, tom lane\n", "msg_date": "Thu, 04 Mar 1999 10:06:48 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Fixed! PostgreSQL 6.4.2 on AIX 4.3.2: typeidTypeRelid\n\terror" }, { "msg_contents": "On Thu, 4 Mar 1999, David R. Favor wrote:\n\n> I was able to fix this problem by recompiling everything using\n> xlc 3.6.4 (AIX native compiler). This leads me to believe there\n> is some evil bug in gcc.\n> \n> I have no idea how to pursue this. For now I will just use xlc.\n> \n> If anyone wants to try and solve this and needs help, I will keep\n> the gcc build tree I have and give any assistance requested.\n\nWhat version of gcc, and have you tried egcs 1.1.1? \n\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 4 Mar 1999 11:13:20 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Fixed! PostgreSQL 6.4.2 on AIX 4.3.2: typeidTypeRelid\n\terror" }, { "msg_contents": "On Thu, 4 Mar 1999, Tom Lane wrote:\n\n> \"David R. Favor\" <[email protected]> writes:\n> > I was able to fix this problem by recompiling everything using\n> > xlc 3.6.4 (AIX native compiler). This leads me to believe there\n> > is some evil bug in gcc.\n> \n> ... where apparently \"gcc\" = \"egcs-2.93.08 19990214\".\n> \n> I'm not sure about the stability of egcs releases; do you have a\n> production release there or just a beta?\n\nI'm using 1.1.1 athome...I suspect that unless you start using C++ stuff,\negcs should be as stable, if not more so, then gcc ... the C++ stuff\nappears to be their primary focus, as its totally broken in gcc ..\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 4 Mar 1999 11:38:21 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Fixed! PostgreSQL 6.4.2 on AIX 4.3.2: typeidTypeRelid\n\terror" }, { "msg_contents": "> I was able to fix this problem by recompiling everything using\n> xlc 3.6.4 (AIX native compiler). This leads me to believe there\n> is some evil bug in gcc.\n> I have no idea how to pursue this. For now I will just use xlc.\n> If anyone wants to try and solve this and needs help, I will keep\n> the gcc build tree I have and give any assistance requested.\n\nIf you would like to start a doc/FAQ_AIX this would be a good thing to\nstart with...\n\nAnd if you want to submit a patch for the conditional definition of data\ntypes try something like\n\n $ diff -c file.h.orig file.h > file.h.patch\n\n - Tom\n", "msg_date": "Fri, 05 Mar 1999 02:39:57 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: Fixed! PostgreSQL 6.4.2 on AIX 4.3.2: typeidTypeRelid error" } ]
[ { "msg_contents": "I'm using egcs 1.1.1.\n", "msg_date": "Thu, 4 Mar 1999 09:17:41 -0600", "msg_from": "\"David R. Favor\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Fixed! PostgreSQL 6.4.2 on AIX 4.3.2: typeidTypeRelid\n\terror" } ]
[ { "msg_contents": "If anyone has a fix for this, let me know.\n\nThanks.\n\n_______\n\n\naasc24>gmake all\nxlc -I../../include -I../../backend -I/usr/local/include -O2 -qchars=signed -qmaxmem=8192 -qhalt=w -qsrcmsg -qcheck=divzero -qlonglong -DFIX_AIX43_TYPES -I../../interfaces/libpq -I../../include \n-c regress.c -o regress.o\n../../backend/port/aix/mkldexport.sh regress.o `pwd` > regress.exp\nMaking share library regress.so from regress.o, regress.exp, and installed postgres.imp\nld -H512 -bM:SRE -bI:/usr/local/pgsql/lib/postgres.imp -bE:regress.exp -o regress.so regress.o -L/usr/local/lib -L/lib -lPW -lcrypt -lld -lnsl -ldl -lm -lreadline -lhistory -lcurses \nld: 0711-224 WARNING: Duplicate symbol: fmgr_pl_finfo\nld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more information.\nld: 0711-317 ERROR: Undefined symbol: _iob\ngmake: *** [regress.so] Error 8\nrm regress.exp regress.o\n", "msg_date": "Thu, 04 Mar 1999 14:42:12 -0600", "msg_from": "\"David R. Favor\" <[email protected]>", "msg_from_op": true, "msg_subject": "PostgreSQL 6.4.2 on AIX 4.3.2: regression test compile error" } ]
[ { "msg_contents": "I have a problem with Access97 not working properly when entering new\nrecords using a sub form, i.e. entering a new order/orderlines or master and\ndetail tables. The problem is caused by a SQL statement that Access97 makes\ninvolving NULL. The syntax that fails is \"column_name\" = NULL. The\nfollowing attachment was provided by -Jose'-. It contains a very small\nenhancement to gram.y that will allow Access97 to work properly with sub\nforms. Can this enhancement be added to release 6.5?\n\n <<gram.patch>> \n\n\nThanks, Michael", "msg_date": "Thu, 4 Mar 1999 15:33:03 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "6.5 enhancement request" } ]
[ { "msg_contents": "\nWhat in the hell is this all about?\n\n--Michael\n\nTopica Support <[email protected]> writes:\n\n> Hi there,\n> \n> We received your request to [email protected].\n> \n> We're sorry but that list does not seem to exist.\n> \n> Please check the spelling and punctuation of the list name.\n> You can also find the list in our directory at http://demo001.topica.com\n> and join the list from there.\n> \n> If you feel that you have received this message in error or for more\n> information on using the Topica system, please visit our web site at\n> http://demo001.topica.com. You can also contact our Customer Support\n> Department at [email protected].\n> \n> If you are trying to reach a topica employee, please ensure that you \n> use the @get.topica.com domain, and not @topica.com. \n> \n> Thanks!\n> \n> Thomas\n> Topica Customer Support\n> \n> \n> \n> -------------------------------\n> +++++ Original Message ++++++++\n> -------------------------------\n> \n> From: Michael Graff <[email protected]>\n> Date: 03 Mar 1999 00:32:46 -0800\n> In-Reply-To: Holm Tiffe's message of \"Wed, 3 Mar 1999 08:28:16 +0100\"\n> Message-ID: <[email protected]>\n> Lines: 33\n> X-Mailer: Gnus v5.5/Emacs 20.3\n> Sender: [email protected]\n> Precedence: bulk\n> \n> \n> NetBSD is the same way. Any chance of someone coming up with a patch\n> if int8 isn't supported?\n> \n> --Michael\n> \n> Holm Tiffe <[email protected]> writes:\n> \n> > Hi all,\n> > \n> > While compiling of 6.4.2 I've realized that int8's are not supported\n> > on FreeBSD by default. The Configure script looks if %ld or %lld\n> > will do the job but FreeBSD is using %qd as the long long int format.\n> > After a little fiddeling around with the configure script and \n> > include/utils/int8.h, I've got the int8 type working.\n> > Since I've read somtimes in this Mailinglists, that FreeBSD is one\n> > of the development platforms for PostgreSQL, I wonder why PGSQL\n> > doesn't support FreeBSD's own format ...\n> > \n> > BTW: why the configure script think's that tcl and tk includes\n> > (tclConfig.sh,tkConfig.sh) must reside in the same directory ?\n> > That's really an odd assumtion.\n> > \n> > \n> > Holm\n> > \n> > PS:\n> > sorry for my poor english.\n> > -- \n> > FreibergNet Systemhaus GbR Holm Tiffe * Administration, Development\n> > Systemhaus f�r Daten- und Netzwerktechnik phone +49 3731 781279\n> > Unternehmensgruppe Liebscher & Partner fax +49 3731 781377\n> > D-09599 Freiberg * Am St. Niclas Schacht 13 http://www.freibergnet.de/\n", "msg_date": "04 Mar 1999 15:03:27 -0800", "msg_from": "Michael Graff <[email protected]>", "msg_from_op": true, "msg_subject": "Re: Your recent message to Topica.com" }, { "msg_contents": "Michael Graff <[email protected]> writes:\n> What in the hell is this all about?\n\nI got about a dozen of those yesterday, from both pg-hackers and\npg-interfaces posts IIRC.\n\nI would say that someone has subscribed an archiving robot to the pgsql\nlists ... only the someone is incompetent to write an archiving robot.\nBouncing robot failure messages back to the originator of a list message\nis not acceptable behavior, since the originator neither cares nor is in\na position to do anything about it.\n\nHey Marc, did these people ask permission before deciding to set up\ntheir own archiver for the pgsql lists? If not, that'd be another\nfailure to respect basic netiquette in my eyes.\n\nI'd vote for unsubscribing topica.com with extreme prejudice.\n\n\t\t\tregards, tom lane\n", "msg_date": "Fri, 05 Mar 1999 10:03:50 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Your recent message to Topica.com " } ]
[ { "msg_contents": "Hi,\n\nI have a 2 tables and in some cases joining them are very slow.\n\nHere are details.\n\ncreate table postal (\n\toldcode varchar(5),\t-- has an btree index\n\tnewcode char(7),\t-- has an btree index\n\tpid int2,\t\t-- has an btree index\n\tkana_city text,\t\t-- has an btree index\n\tkana_town text,\t\t-- has an btree index\n\tcity text,\t\t-- has an btree index\n\ttown text\t\t-- has an btree index\n);\n\n(has 119479 records)\n\ncreate table prefecture (\n\tpid int2,\t\t-- has an btree index\n\tpref char(8),\n\tkana_pref char(16)\n);\n\n(has 47 records)\n\nMy question is:\n\nThis is fast as I expected.\n\npostal=> explain select * from postal,prefecture where city = 'aaa' and postal.pid = prefecture.pid;\nNOTICE: QUERY PLAN:\n\nNested Loop (cost=4.10 size=1 width=100)\n -> Index Scan using cityindex on postal (cost=2.05 size=1 width=74)\n -> Index Scan using prefpidindex on prefecture (cost=2.05 size=47 width=26)\n\nBut:\n\npostal=> explain select * from postal,prefecture where city ~ '^aaa' and postal.pid = prefecture.pid;\nNOTICE: QUERY PLAN:\n\nNested Loop (cost=98.90 size=1 width=100)\n -> Seq Scan on prefecture (cost=2.55 size=47 width=26)\n -> Index Scan using pidindex on postal (cost=2.05 size=1 width=74)\n\nThis is so slooow. Can anybody explain this? Am I missing something?\n\nNote that 6.4.x and current show same behavior.\n---\nTatsuo Ishii\n\n", "msg_date": "Fri, 05 Mar 1999 12:50:48 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": true, "msg_subject": "Why is that so slow?" }, { "msg_contents": "I got a good suggestion from Hiroshi. The conclusion is:\n\n>\tpid int2,\t\t-- has an btree index\n\nThis is bad. I had defined a btree index on pid and it has 2000\nduplicate entries in average! After I removed the index, the query\nruns unbelievably fast! Now explain shows:\n\nNested Loop (cost=933.82 size=1 width=100)\n -> Index Scan using cityindex on postal (cost=931.77 size=1 width=74)\n -> Index Scan using prefpidindex on prefecture (cost=2.05 size=47 width=26)\n\n--\nTatsuo Ishii\n", "msg_date": "Fri, 05 Mar 1999 14:37:38 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Why is that so slow? " }, { "msg_contents": "Tatsuo Ishii <[email protected]> writes:\n> postal=> explain select * from postal,prefecture where city ~ '^aaa' and postal.pid = prefecture.pid;\n> NOTICE: QUERY PLAN:\n> \n> Nested Loop (cost=98.90 size=1 width=100)\n> -> Seq Scan on prefecture (cost=2.55 size=47 width=26)\n> -> Index Scan using pidindex on postal (cost=2.05 size=1 width=74)\n> \n> This is so slooow. Can anybody explain this? Am I missing something?\n\nand later:\n> I had defined a btree index on pid and it has 2000\n> duplicate entries in average! After I removed the index, the query\n> runs unbelievably fast! Now explain shows:\n\n> Nested Loop (cost=933.82 size=1 width=100)\n> -> Index Scan using cityindex on postal (cost=931.77 size=1 width=74)\n> -> Index Scan using prefpidindex on prefecture (cost=2.05 size=47 width=26)\n\nHmm. Removal of the index is just a hack --- the system should have\nbeen smart enough not to use it. It looks like the system chose the\nfirst plan shown above because it thought that selecting postal entries\nmatching a particular pid value would on average match only one postal\ntuple (note the \"size\" fields, which are estimates of the numbers of\nresulting tuples). But in reality, each scan produced 2000 matching\nentries on average, according to your second message --- and each of\nthose entries had to be tested to see if it had the right city name.\nSo, very slow.\n\nThe question I have is why didn't the system realize that there would be\nlots of matches on pid? The \"dispersion\" statistics it keeps ought to\nhave given it a clue that this approach wouldn't be very selective.\n\nThe second example is fast because the scan over postal looking for city\nname matches finds only one match, so prefecture is scanned only once.\nHowever the cost estimates there also look bogus --- the system is again\nmis-guessing how many entries will be selected. It seems to think that\nall 47 prefecture entries will be matched by a scan for a specific pid.\nSo, bogus dispersion values again (or bad use of them).\n\nSomething is fishy here. Have you done a \"vacuum analyze\" since loading\nthe data in these tables?\n\n\t\t\tregards, tom lane\n", "msg_date": "Fri, 05 Mar 1999 18:24:41 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Why is that so slow? " }, { "msg_contents": "> Tatsuo Ishii <[email protected]> writes:\n> > postal=> explain select * from postal,prefecture where city ~ '^aaa' and postal.pid = prefecture.pid;\n> > NOTICE: QUERY PLAN:\n> > \n> > Nested Loop (cost=98.90 size=1 width=100)\n> > -> Seq Scan on prefecture (cost=2.55 size=47 width=26)\n> > -> Index Scan using pidindex on postal (cost=2.05 size=1 width=74)\n> > \n> > This is so slooow. Can anybody explain this? Am I missing something?\n> \n> and later:\n> > I had defined a btree index on pid and it has 2000\n> > duplicate entries in average! After I removed the index, the query\n> > runs unbelievably fast! Now explain shows:\n> \n> > Nested Loop (cost=933.82 size=1 width=100)\n> > -> Index Scan using cityindex on postal (cost=931.77 size=1 width=74)\n> > -> Index Scan using prefpidindex on prefecture (cost=2.05 size=47 width=26)\n> \n> Hmm. Removal of the index is just a hack --- the system should have\n> been smart enough not to use it. It looks like the system chose the\n> first plan shown above because it thought that selecting postal entries\n> matching a particular pid value would on average match only one postal\n> tuple (note the \"size\" fields, which are estimates of the numbers of\n> resulting tuples). But in reality, each scan produced 2000 matching\n> entries on average, according to your second message --- and each of\n> those entries had to be tested to see if it had the right city name.\n> So, very slow.\n> \n> The question I have is why didn't the system realize that there would be\n> lots of matches on pid? The \"dispersion\" statistics it keeps ought to\n> have given it a clue that this approach wouldn't be very selective.\n> \n> The second example is fast because the scan over postal looking for city\n> name matches finds only one match, so prefecture is scanned only once.\n\nActulally not only one since I use ~ operator. Anyway matching rows\nwould be reasonably small.\n\n> However the cost estimates there also look bogus --- the system is again\n> mis-guessing how many entries will be selected. It seems to think that\n> all 47 prefecture entries will be matched by a scan for a specific pid.\n> So, bogus dispersion values again (or bad use of them).\n> \n> Something is fishy here. Have you done a \"vacuum analyze\" since loading\n> the data in these tables?\n\nOh, I never thought about that. After re-made the index I removed in\nthe next letter and did vacuum analyze, I got:\n\nHash Join (cost=951.50 size=19 width=100)\n -> Index Scan using cityindex on postal (cost=944.77 size=19 width=74)\n -> Hash (cost=0.00 size=0 width=0)\n -> Seq Scan on prefecture (cost=2.55 size=47 width=26)\n\nThis plan looks good(and actually as fast as the previous\none). However, the cost estimate for prefecture is again 47?\n--\nTatsuo Ishii\n", "msg_date": "Sat, 06 Mar 1999 23:08:09 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Why is that so slow? " }, { "msg_contents": "Tatsuo Ishii <[email protected]> writes:\n>> Something is fishy here. Have you done a \"vacuum analyze\" since loading\n>> the data in these tables?\n\n> Oh, I never thought about that.\n\nAh. OK, that explains the system's poor choice of plan --- it was\neffectively operating on the assumption that these tables were small.\n\n(Note to hackers: maybe a freshly created table should be given dummy\nstatistics, say having 1000 rows instead of 0 rows? That would help\nto prevent the optimizer from making really foolish choices when no\nvacuum's been done yet for the table. But I dunno whether we could\ninvent plausible default values for all the stats...)\n\n> After re-made the index I removed in\n> the next letter and did vacuum analyze, I got:\n\n> Hash Join (cost=951.50 size=19 width=100)\n> -> Index Scan using cityindex on postal (cost=944.77 size=19 width=74)\n> -> Hash (cost=0.00 size=0 width=0)\n> -> Seq Scan on prefecture (cost=2.55 size=47 width=26)\n\n> This plan looks good(and actually as fast as the previous\n> one). However, the cost estimate for prefecture is again 47?\n\nNo, that looks OK in this context: it's proposing to load the whole\nprefecture table into an internal hashtable, so it will have to scan\nall 47 prefecture rows to do it. The only guesstimating in this plan\nis the \"size=19\" for the index scan, ie, an estimated 19 hits from the\nmatch on city name. That seems fairly reasonable, although of course\nit could be badly off depending on your match pattern.\n\n\t\t\tregards, tom lane\n", "msg_date": "Sat, 06 Mar 1999 11:42:05 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Why is that so slow? " }, { "msg_contents": "> Tatsuo Ishii <[email protected]> writes:\n> >> Something is fishy here. Have you done a \"vacuum analyze\" since loading\n> >> the data in these tables?\n> \n> > Oh, I never thought about that.\n> \n> Ah. OK, that explains the system's poor choice of plan --- it was\n> effectively operating on the assumption that these tables were small.\n> \n> (Note to hackers: maybe a freshly created table should be given dummy\n> statistics, say having 1000 rows instead of 0 rows? That would help\n> to prevent the optimizer from making really foolish choices when no\n> vacuum's been done yet for the table. But I dunno whether we could\n> invent plausible default values for all the stats...)\n\nNo way to really make a default. Zero is the correct number when the\ntable is created, right? Improved optimize may be even worse or better\nfor un-analyzed tables.\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 7 Mar 1999 06:27:54 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Why is that so slow?" }, { "msg_contents": "Bruce Momjian <[email protected]> writes:\n>> (Note to hackers: maybe a freshly created table should be given dummy\n>> statistics, say having 1000 rows instead of 0 rows? That would help\n>> to prevent the optimizer from making really foolish choices when no\n>> vacuum's been done yet for the table. But I dunno whether we could\n>> invent plausible default values for all the stats...)\n\n> No way to really make a default. Zero is the correct number when the\n> table is created, right?\n\nWell, it's right at the instant of creation, but I think that's much too\nsimplistic a way of looking at it. Tables are generally created with\nthe intention of putting data into them. It's a reasonable assumption\nthat the table will shortly have some rows in it.\n\nNow, any particular estimate like 1000 is obviously going to be wrong.\nThe point I'm trying to make is that the optimizer is more likely to\ngenerate a sane plan if it assumes that the table contains a moderate\nnumber of rows. We have seen gripes time and time again from people\nwho made a table, didn't bother to do a vacuum, and got horribly slow\nnested-loop plans from the optimizer because it assumed their table\nwas empty. With a nonzero initial estimate, the optimizer will choose\na plan that might be somewhat inefficient if the table really is small;\nbut it won't be seriously unusable if the table is large.\n\nOnce you've done a vacuum, of course, the whole question is moot.\nBut I think the system's behavior would be more robust if it assumed\nthat a never-yet-vacuumed table contained some rows, not no rows.\n\n\t\t\tregards, tom lane\n", "msg_date": "Sun, 07 Mar 1999 12:00:04 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Why is that so slow? " }, { "msg_contents": "> Well, it's right at the instant of creation, but I think that's much too\n> simplistic a way of looking at it. Tables are generally created with\n> the intention of putting data into them. It's a reasonable assumption\n> that the table will shortly have some rows in it.\n> \n> Now, any particular estimate like 1000 is obviously going to be wrong.\n> The point I'm trying to make is that the optimizer is more likely to\n> generate a sane plan if it assumes that the table contains a moderate\n> number of rows. We have seen gripes time and time again from people\n> who made a table, didn't bother to do a vacuum, and got horribly slow\n> nested-loop plans from the optimizer because it assumed their table\n> was empty. With a nonzero initial estimate, the optimizer will choose\n> a plan that might be somewhat inefficient if the table really is small;\n> but it won't be seriously unusable if the table is large.\n> \n> Once you've done a vacuum, of course, the whole question is moot.\n> But I think the system's behavior would be more robust if it assumed\n> that a never-yet-vacuumed table contained some rows, not no rows.\n\nTrue, but the new optimizer code favors ordered/mergejoin over nested\nloop because it favors ordered results over unordered ones like nested\nloop. Should fix the problem. \n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 7 Mar 1999 12:28:33 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Why is that so slow?" } ]
[ { "msg_contents": "> > > | NULL_P '=' a_expr\n> > > { $$ = makeA_Expr(ISNULL, NULL, $3, NULL); }\n> > This leads to a shift/reduce conflict in yacc.\n> What's wrong with shift/reduce conflicts?\n\nyacc looks ahead only one token to determine the parsing possibilities,\nand will maintain multiple, parallel possibilities until it is able to\nresolve them (keeping in mind this one-token constraint). With a\n\"shift/reduce\" conflict, at least one path will *never* be possible,\neven though you thought it should be from the grammar. So you will end\nup with some language feature permanently unavailable. And worse, when a\nnew conflict is introduced it might be the older feature which is\ntrashed :(\n\n - Tom\n", "msg_date": "Fri, 05 Mar 1999 06:57:24 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [INTERFACES] Foreign Keys" } ]
[ { "msg_contents": "> aasc24>gmake all\n> xlc -I../../include -I../../backend -I/usr/local/include -O2\n> -qchars=signed -qmaxmem=8192 -qhalt=w -qsrcmsg -qcheck=divzero -qlonglong\n> -DFIX_AIX43_TYPES -I../../interfaces/libpq -I../../include \n> -c regress.c -o regress.o\n> ../../backend/port/aix/mkldexport.sh regress.o `pwd` > regress.exp\n> Making share library regress.so from regress.o, regress.exp, and installed\n> postgres.imp\n> ld -H512 -bM:SRE -bI:/usr/local/pgsql/lib/postgres.imp -bE:regress.exp -o\n> regress.so regress.o -L/usr/local/lib -L/lib -lPW -lcrypt -lld -lnsl -ldl\n> -lm -lreadline -lhistory -lcurses \n> ld: 0711-224 WARNING: Duplicate symbol: fmgr_pl_finfo\n> ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more\n> information.\n> ld: 0711-317 ERROR: Undefined symbol: _iob\n> gmake: *** [regress.so] Error 8\n rm regress.exp regress.o\n\nYou only need to link -lc to the shared object. Somehow there is a different\nMakefile rule\nin the regression test, than for the rest of the shared libraries. \nAlso I think the template for aix is wrong in doing -qchars=signed. There\nwas a discussion saying\nthat chars should be unsigned.\n\nAndreas\n\n", "msg_date": "Fri, 5 Mar 1999 09:26:34 +0100 ", "msg_from": "Zeugswetter Andreas IZ5 <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] PostgreSQL 6.4.2 on AIX 4.3.2: regression test comp\n\tile error" } ]
[ { "msg_contents": "Hi!\n\n I am continuing playing with the query (I reduced it to):\nSELECT sh.distr_id , d.distr_id \n FROM central cn, shops sh, districts d \n WHERE cn.shop_id = sh.shop_id \n\nand got 27963 rows. I filtered it out (with awk:) where $1 == $2 (1st\ncolumn is equal to second) and got 3104 rows.\n\n But the query\nSELECT d.* \n FROM central cn, shops sh, districts d \n WHERE cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id \n\nreturned 0 rows.\n\n Where is the bug?\n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Fri, 5 Mar 1999 16:13:12 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Bug on complex join" }, { "msg_contents": "> SELECT sh.distr_id , d.distr_id\n> FROM central cn, shops sh, districts d\n> WHERE cn.shop_id = sh.shop_id\n> and got 27963 rows.\n> But the query\n> SELECT d.*\n> FROM central cn, shops sh, districts d\n> WHERE cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n\nWhy did you change both the target columns *and* the query qualification\nbetween these two examples? Is the \"SELECT d.*\" required to get the\nquery to fail?? If not, then...\n\nFor some reason sh.distr_id is not equal to d.distr_id. Are they\ndifferent data types? Do they have some embedded blanks?? Probably not a\nPostgres bug, since the query itself looks pretty simple...\n\n - Tom\n", "msg_date": "Fri, 05 Mar 1999 15:07:08 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Bug on complex join" }, { "msg_contents": "Hi!\n\nOn Fri, 5 Mar 1999, Thomas G. Lockhart wrote:\n> > SELECT sh.distr_id , d.distr_id\n> > FROM central cn, shops sh, districts d\n> > WHERE cn.shop_id = sh.shop_id\n> > and got 27963 rows.\n> > But the query\n> > SELECT d.*\n> > FROM central cn, shops sh, districts d\n> > WHERE cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n> \n> Why did you change both the target columns *and* the query qualification\n> between these two examples? Is the \"SELECT d.*\" required to get the\n> query to fail?? If not, then...\n\n Cause I need something in the target list. Wrong way to test it? What is\na better way?\n\n> For some reason sh.distr_id is not equal to d.distr_id. Are they\n> different data types? Do they have some embedded blanks?? Probably not a\n\n Both are int2. No blanks (at least I cannot imagine blanks in int2 :).\n\n> Postgres bug, since the query itself looks pretty simple...\n\n That's why I am very confused. :(\n\n> \n> - Tom\n> \n\nOleg.\n---- \n Oleg Broytmann National Research Surgery Centre http://sun.med.ru/~phd/\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Fri, 5 Mar 1999 18:13:50 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Bug on complex join" }, { "msg_contents": "Oleg Broytmann wrote:\n> \n> Hi!\n> \n> I am continuing playing with the query (I reduced it to):\n> SELECT sh.distr_id , d.distr_id\n> FROM central cn, shops sh, districts d\n> WHERE cn.shop_id = sh.shop_id\n> \n> and got 27963 rows. I filtered it out (with awk:) where $1 == $2 (1st\n> column is equal to second) and got 3104 rows.\n> \n> But the query\n> SELECT d.*\n> FROM central cn, shops sh, districts d\n> WHERE cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n> \n> returned 0 rows.\n> \n> Where is the bug?\n\nPlease post me EXPLAIN VERBOSE for second query.\n\nVadim\n", "msg_date": "Sat, 06 Mar 1999 17:06:48 +0700", "msg_from": "Vadim Mikheev <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Bug on complex join" }, { "msg_contents": "On Sat, 6 Mar 1999, Vadim Mikheev wrote:\n\n> Oleg Broytmann wrote:\n> > \n> > Hi!\n> > \n> > I am continuing playing with the query (I reduced it to):\n> > SELECT sh.distr_id , d.distr_id\n> > FROM central cn, shops sh, districts d\n> > WHERE cn.shop_id = sh.shop_id\n> > \n> > and got 27963 rows. I filtered it out (with awk:) where $1 == $2 (1st\n> > column is equal to second) and got 3104 rows.\n> > \n> > But the query\n> > SELECT d.*\n> > FROM central cn, shops sh, districts d\n> > WHERE cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n> > \n> > returned 0 rows.\n> > \n> > Where is the bug?\n> \n> Please post me EXPLAIN VERBOSE for second query.\n\n\nEXPLAIN VERBOSE SELECT d.*\n FROM central cn, shops sh, districts d\n WHERE cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n;\npqReadData() -- backend closed the channel unexpectedly.\n\tThis probably means the backend terminated abnormally before or while processing the request.\nWe have lost the connection to the backend, so further processing is impossible. Terminating.\n\n :(((\n\n> Vadim\n> \n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Tue, 9 Mar 1999 12:33:08 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Bug on complex join" }, { "msg_contents": "Hello!\n\n Another symptom. The query\n\nSELECT cn.date_i, cn.pos_id \n FROM central cn \n WHERE cn.date_i >= current_date - '300 days'::timespan \n\nreturns 3156 rows. But this:\n\nSELECT cn.date_i, p.subsec_id, cn.pos_id, p.pos_id\n FROM central cn, shops sh, districts d, positions p\n WHERE cn.date_i >= current_date - '300 days'::timespan\n\nfailed:\n\npqReadData() -- backend closed the channel unexpectedly.\n\tThis probably means the backend terminated abnormally before or while processing the request.\nWe have lost the connection to the backend, so further processing is impossible. Terminating.\n\n Tables attached (ZIP file with script to recreate tables and SQL\ncommands).\n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.", "msg_date": "Tue, 9 Mar 1999 12:40:09 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Bug on complex join" }, { "msg_contents": "Oleg Broytmann wrote:\n> \n> Hello!\n> \n> Another symptom. The query\n> \n> SELECT cn.date_i, cn.pos_id\n> FROM central cn\n> WHERE cn.date_i >= current_date - '300 days'::timespan\n> \n> returns 3156 rows. But this:\n> \n> SELECT cn.date_i, p.subsec_id, cn.pos_id, p.pos_id\n> FROM central cn, shops sh, districts d, positions p\n> WHERE cn.date_i >= current_date - '300 days'::timespan\n\nthis should return\n 3156 * count(shops) * count(districts) * count(positions)\n\nwhich is probably too much for the backend ;(\n\n-----------------------\nHannu\n", "msg_date": "Tue, 09 Mar 1999 14:56:30 +0200", "msg_from": "Hannu Krosing <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Bug on complex join" }, { "msg_contents": "Hi!\n\nOn Tue, 9 Mar 1999, Hannu Krosing wrote:\n> > SELECT cn.date_i, cn.pos_id\n> > FROM central cn\n> > WHERE cn.date_i >= current_date - '300 days'::timespan\n> > \n> > returns 3156 rows. But this:\n> > \n> > SELECT cn.date_i, p.subsec_id, cn.pos_id, p.pos_id\n> > FROM central cn, shops sh, districts d, positions p\n> > WHERE cn.date_i >= current_date - '300 days'::timespan\n> \n> this should return\n> 3156 * count(shops) * count(districts) * count(positions)\n> \n> which is probably too much for the backend ;(\n\n Bad news. Thanks for pointing this.\n\n> -----------------------\n> Hannu\n> \n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Tue, 9 Mar 1999 18:47:24 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Bug on complex join" }, { "msg_contents": "Oleg Broytmann wrote:\n> \n> Hi!\n> \n> On Tue, 9 Mar 1999, Hannu Krosing wrote:\n> > > SELECT cn.date_i, cn.pos_id\n> > > FROM central cn\n> > > WHERE cn.date_i >= current_date - '300 days'::timespan\n> > >\n> > > returns 3156 rows. But this:\n> > >\n> > > SELECT cn.date_i, p.subsec_id, cn.pos_id, p.pos_id\n> > > FROM central cn, shops sh, districts d, positions p\n> > > WHERE cn.date_i >= current_date - '300 days'::timespan\n> >\n> > this should return\n> > 3156 * count(shops) * count(districts) * count(positions)\n> >\n> > which is probably too much for the backend ;(\n ^^^^^^^^^^^^^^^\nFor the client-side, not for the backend - backend doesn't\nkeep all result tuples in memory.\n\n> \n> Bad news. Thanks for pointing this.\n\nVadim\n", "msg_date": "Wed, 10 Mar 1999 09:20:13 +0700", "msg_from": "Vadim Mikheev <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Bug on complex join" } ]
[ { "msg_contents": "Thanks. Declaring row as a record did the trick. I have a lot of tables\nand I am looking to more deeply understand the PostgreSQL security module. \n\n\t-----Original Message-----\n\tFrom:\[email protected] [SMTP:[email protected]]\n\tSent:\tFriday, March 05, 1999 3:00 AM\n\tTo:\tMichael Davis\n\tCc:\[email protected]\n\tSubject:\tRe: [HACKERS] for row in select loop question\n\n\t>\n\t> The following function does not work:\n\t>\n\t> CREATE FUNCTION InstallPermissions(varchar, varchar) RETURNS int2\nAS '\n\t> DECLARE\n\t> options ALIAS FOR $1;\n\t> username ALIAS FOR $2;\n\t> BEGIN\n\t> FOR row IN select * from pg_tables LOOP\n\t> REVOKE ALL ON row.tablename FROM username;\n\t> GRANT options on row.tablename TO username;\n\t> END LOOP;\n\t> return 0;\n\t> END; ' LANGUAGE 'plpgsql';\n\t>\n\t> \"select InstallPermissions('test1', 'test2);\" returns an error at\n\"select\".\n\t> Any suggestions on how to eliminate the error?\n\n\t First you forgot to declare 'row'. Could be type record.\n\n\t Second REVOKE and GRANT are utility statements not supported\n\t for prepared SPI plans and thus PL/pgSQL currently cannot\n\t execute them.\n\n\n\tJan\n\n\t--\n\n\t\n#======================================================================#\n\t# It's easier to get forgiveness for being wrong than for being\nright. #\n\t# Let's break this rule - forgive me.\n#\n\t#======================================== [email protected] (Jan\nWieck) #\n\t\n", "msg_date": "Fri, 5 Mar 1999 09:34:11 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "RE: [HACKERS] for row in select loop question" } ]
[ { "msg_contents": "Hi all\n\nI'm building up a new server based on a min. install of Red Hat 5.2 +\nupdates. Compiling all the server software myself, including Postgresql\n6.4.2. (Never installed the Red Hat server stuff to start with.)\n(server stuff == PHP3, mod_ssl, Apache, PostgreSQL, ...)\n\nAny way, int2, int4 (both on line 10), and geometry all say they fail.\nA diff of expected/results for int2 and int4 show that they are the same\nexcept for the phrasing of the error message (over flow test?)\n\nexpected: \nMath result not representable\n\nresults:\nNumerical result out of range\n\nThe differance for the geometry type is harder to tell as each fill a\nlittle more then a screen full and at a glance look about the same --\ngiberish. But as I have never use the geometry type, and am not even sure\nwhat it would be used for, I'm not worried about it. Every thing else\npassed.\n\nIf you like I can supply the \"result\" files.\n\nHave a great day\nTerry Mackintosh <[email protected]> http://www.terrym.com\nsysadmin/owner I'm excited about life! How about YOU!?\nProfessional Web Hosting and site design to include programming\nProudly powered by R H Linux 4.2, Apache 1.3.x, PHP 3.x, PostgreSQL 6.x\n-----------------------------------------------------------------------\nOnly if you know where you're going can you get there.\n\n", "msg_date": "Fri, 5 Mar 1999 10:36:25 -0500 (EST)", "msg_from": "Terry Mackintosh <[email protected]>", "msg_from_op": true, "msg_subject": "Regression differences on Red Hat 5.2" } ]
[ { "msg_contents": "Hello!\n\n Is anybody here who know Greek, or at least have greek locale (iso8859-7)?\nI got test data from Angelos Karageorgiou <[email protected]>,\nhe reported the test passed right on his BSD/OS. But when I ran locale test\non my linux box (Debian 2.0) I got a lot of diffs. Anyone to test the test\ndata?\n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Fri, 5 Mar 1999 18:42:12 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Greek locale" } ]
[ { "msg_contents": "It's currently not installed by default. \nJust wondering if this was intentional.\nCVS source (3/5/1999 9:00am CST)\n\t-DEJ\n", "msg_date": "Fri, 5 Mar 1999 11:49:21 -0600", "msg_from": "\"Jackson, DeJuan\" <[email protected]>", "msg_from_op": true, "msg_subject": "PL/PGSQL" } ]
[ { "msg_contents": "More info...\n\nmklang_pl.sql isn't getting copied to the lib directory on install, so\ninitdb isn't placing PL/PGSQL in template1. \n\n> -----Original Message-----\n> It's currently not installed by default. \n> Just wondering if this was intentional.\n> CVS source (3/5/1999 9:00am CST)\n> \t-DEJ\n> \n", "msg_date": "Fri, 5 Mar 1999 12:20:04 -0600", "msg_from": "\"Jackson, DeJuan\" <[email protected]>", "msg_from_op": true, "msg_subject": "RE: [HACKERS] PL/PGSQL" } ]
[ { "msg_contents": "All things considered, I would be content with logging insert, update, and\ndelete instructions to an external file. It would be extremely valuable to\nallow this log file to exist on a mounted file system so it could be stored\non a different machine. It would also be valuable to have the user name\noriginating the transaction, a date time stamp, and possibly a transaction\nnumber as optional fields appended to the logs. Maybe these optional fields\ncould be added to the sql statement as comments? Are comments allowed in a\nsql statement? This would allow for a visual of what has been updated,\nwhen it was updated, and who performed the update. At a later date this\ncould be extended or an alternate means of logging can be added. My vote\nwill always go for additional options.\n\nThe sooner this is implemented the better for me. I am willing to be\ncoached on how to implement this. I will be busy for the next 2-4 weeks as\nI migrate my Access97 database to PostgreSQL. Once this migration is\ncompleted and released to my users, this will become a hot topic for me.\n\nThanks, Michael\n\n\t-----Original Message-----\n\tFrom:\[email protected] [SMTP:[email protected]]\n\tSent:\tFriday, March 05, 1999 11:24 AM\n\tTo:\tMichael Davis\n\tCc:\[email protected]; [email protected]\n\tSubject:\tRe: [HACKERS] RE: [GENERAL] Transaction logging\n\n\tMichael Davis wrote:\n\n\t>\n\t> What would it take to have transaction logging added to Postgres.\nI am a\n\t> c/c++ programmer and will consider contributing to the Postgres\ndevelopment\n\t> effort. I really like everything I see and read about Postgres.\nAs a\n\n\t I spent some time on transaction logging since it's a feature\n\t I'm missing too. There are mainly two different transaction\n\t log mechanisms out.\n\n\t 1. Log queries sent to the backend.\n\n\t 2. Log images of inserted/updated rows and row ID's of\n\t deleted ones.\n\n\t The query level logging will write less information if\n\t queries usually affect a large number of rows. Unfortunately\n\t the extensibility of Postgres work's against this approach.\n\t There could be any number of user written functions who's\n\t results aren't reproduceable during recovery. And core\n\t features of Postgres itself would introduce the same problem.\n\t Have a sequence which is used to create default values for\n\t multiple tables, so that one ID is unique across them. Now\n\t two backends insert (with INSERT ... SELECT) concurrently\n\t into different tables using the same sequence. It's a\n\t classic race condition and it depends on context switching\n\t and page faults which backend will get which sequence\n\t numbers. You cannot foresee and you cannot reproduce, except\n\t you hook into the sequence generator and log this too. Later\n\t when recovering, another hook into the sequence generator\n\t must reproduce the logged results on the per\n\t backend/transaction/command base, and the same must be done\n\t for each function that usually returns unreproduceable\n\t results (anything dealing with time, pid's, etc.).\n\n\t As said, this must also cover user functions. So at least\n\t there must be a general log API that provides such a\n\t functionality for user written functions.\n\n\t The image logging approach also has problems. First, the only\n\t thing given to the heap access methods to outdate a tuple on\n\t update/delete is the current tuple ID (information that tells\n\t which tuple in which block is meant). So you need to save\n\t the database files in binary format, because during the\n\t actually existing dump/restore this could change and the\n\t logged CTID's would hit the wrong tuples.\n\n\t Second, you must remember in the log which transaction ID\n\t these informations came from and later if the transaction\n\t committed or not, so the recovery can set this commit/abort\n\t information in pg_log too. pg_log is a shared system file and\n\t the transaction ID's are unique only for one server. Using\n\t this information for online replication of a single database\n\t to another Postgres installation will not work.\n\n\t Third, there are still some shared system catalogs across all\n\t databases (pg_database, pg_group, pg_log!!!, pg_shadow and\n\t pg_variable). Due to that it would be impossible (or at least\n\t very, very tricky) to restore/recover (maybe point in time)\n\t one single database. If you destroy one database and restore\n\t it from the binary backup, these shared catalogs cannot be\n\t restored too, so they're out of sync with the backup time.\n\t How should the recovery now hit the right things (which\n\t probably must not be there at all)?.\n\n\t All this is really a mess. I think the architecture of\n\t Postgres will only allow something on query level with some\n\t general API for things that must reproduce the same result\n\t during recovery. For example time(). Inside the backend,\n\t time() should never be called directly. Instead another\n\t function is to be called that log's during normal operation\n\t which time get's returned by this particular function call\n\t and if the backend is in recovery mode, returns the value\n\t from the log.\n\n\t And again, this all means trouble. Usually, most queries sent\n\t to the database don't change any data because they are\n\t SELECT's. It would dramatically blow up the log amount if you\n\t log ALL queries instead of only those that modify things. But\n\t when the query begins, you don't know this, because a SELECT\n\t might call a function that uses SPI to UPDATE something else.\n\t So the decision if the query must be logged or not can only\n\t be made when the query is done (by having some global\n\t variable where the heap access methods set a flag that\n\t something got written). Now you have to log function call's\n\t like time() even if the query will not modify any single row\n\t in the database because the query is a\n\n\t SELECT 'now'::datetime - updtime FROM ...\n\n\t Doing this on a table with thousands of rows will definitely\n\t waste much logging space and slowdown the whole thing by\n\t unnecessary logging.\n\n\t Maybe it's a compromise if at each query start the actual\n\t time and other such information is remembered by the backend,\n\t all time() calls return this remembered value instead of the\n\t real one (wouldn't be bad anyway IMHO), and this information\n\t is logged only if the query is to be logged.\n\n\t Finally I think I must have missed some more problems, but\n\t aren't these enough already to frustrate you :-?\n\n\n\tJan\n\n\t--\n\n\t\n#======================================================================#\n\t# It's easier to get forgiveness for being wrong than for being\nright. #\n\t# Let's break this rule - forgive me.\n#\n\t#======================================== [email protected] (Jan\nWieck) #\n\t\n", "msg_date": "Fri, 5 Mar 1999 12:44:10 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "RE: [HACKERS] RE: [GENERAL] Transaction logging" }, { "msg_contents": "On Fri, 5 Mar 1999, Michael Davis wrote:\n\n> All things considered, I would be content with logging insert, update, and\n> delete instructions to an external file. It would be extremely valuable to\n> allow this log file to exist on a mounted file system so it could be stored\n> on a different machine. It would also be valuable to have the user name\n> originating the transaction, a date time stamp, and possibly a transaction\n> number as optional fields appended to the logs. \n\nI agree that this would be an _excellent_ first cut at logging. Yes, if\nyou have \"insert time()\", then you will have problems, but for the majority\nof applications, what is suggested above will be good enough, and I'm a big\nanti-fan of letting the best be the enemy of the good.\n\n> Maybe these optional fields\n> could be added to the sql statement as comments? Are comments allowed in a\n> sql statement? This would allow for a visual of what has been updated,\n> when it was updated, and who performed the update. At a later date this\n> could be extended or an alternate means of logging can be added. My vote\n> will always go for additional options.\n\nJust define a document format for the logs which allow for this data.\nIf you do it in XML (GNOME and mozilla both have great free xml libraries\navailable), then all of the normal, nasty issues, like whitespace and\nparsing, are completely taken care of for you. Just a suggestion.\n\n--\nTodd Graham Lewis 32�49'N,83�36'W (800) 719-4664, x2804\n******Linux****** MindSpring Enterprises [email protected]\n\n\"A pint of sweat will save a gallon of blood.\" -- George S. Patton\n\n", "msg_date": "Sat, 6 Mar 1999 16:27:15 -0500 (EST)", "msg_from": "Todd Graham Lewis <[email protected]>", "msg_from_op": false, "msg_subject": "RE: [HACKERS] RE: [GENERAL] Transaction logging" } ]
[ { "msg_contents": "Wouldn't it be possible to have the commit take care of the logging. I\ndon't know enough about the backend, but all those functions have to get\nevaluated at some point before the commit in order to be written to the\ndb. You could actually have the commit log the needed sql-commands\nwithout any function interference. Also doesn't commit know if any rows\nhave been updated by any of the transaction's statements. This\ncould/would require filtering rules and triggers from the output, or\nturning them off on the restore.\nAs far as the shared-tables/catalogues are concerned. We could have\neach full-database dump take care of it's entries into the shared\ntables. Having the transaction logs be incremental from the point of\nthe dump allows the catalogues to be self updating (assuming we are\nlogging vacuums).\n\nAll this depends on getting a database commit to output it's information\neasily. It also ignores MVCC's effect on transactions, but I think it's\na workable model.\n\nYou can tell me I'm blowing smoke now.\n\t-DEJ\n\n> -----Original Message-----\n> From: [email protected] [mailto:[email protected]]\n> Sent: Friday, March 05, 1999 12:24 PM\n> To: [email protected]\n> Cc: [email protected]; [email protected]\n> Subject: Re: [HACKERS] RE: [GENERAL] Transaction logging\n> \n> \n> Michael Davis wrote:\n> \n> >\n> > What would it take to have transaction logging added to \n> Postgres. I am a\n> > c/c++ programmer and will consider contributing to the \n> Postgres development\n> > effort. I really like everything I see and read about \n> Postgres. As a\n> \n> I spent some time on transaction logging since it's a feature\n> I'm missing too. There are mainly two different transaction\n> log mechanisms out.\n> \n> 1. Log queries sent to the backend.\n> \n> 2. Log images of inserted/updated rows and row ID's of\n> deleted ones.\n> \n> The query level logging will write less information if\n> queries usually affect a large number of rows. Unfortunately\n> the extensibility of Postgres work's against this approach.\n> There could be any number of user written functions who's\n> results aren't reproduceable during recovery. And core\n> features of Postgres itself would introduce the same problem.\n> Have a sequence which is used to create default values for\n> multiple tables, so that one ID is unique across them. Now\n> two backends insert (with INSERT ... SELECT) concurrently\n> into different tables using the same sequence. It's a\n> classic race condition and it depends on context switching\n> and page faults which backend will get which sequence\n> numbers. You cannot foresee and you cannot reproduce, except\n> you hook into the sequence generator and log this too. Later\n> when recovering, another hook into the sequence generator\n> must reproduce the logged results on the per\n> backend/transaction/command base, and the same must be done\n> for each function that usually returns unreproduceable\n> results (anything dealing with time, pid's, etc.).\n> \n> As said, this must also cover user functions. So at least\n> there must be a general log API that provides such a\n> functionality for user written functions.\n> \n> The image logging approach also has problems. First, the only\n> thing given to the heap access methods to outdate a tuple on\n> update/delete is the current tuple ID (information that tells\n> which tuple in which block is meant). So you need to save\n> the database files in binary format, because during the\n> actually existing dump/restore this could change and the\n> logged CTID's would hit the wrong tuples.\n> \n> Second, you must remember in the log which transaction ID\n> these informations came from and later if the transaction\n> committed or not, so the recovery can set this commit/abort\n> information in pg_log too. pg_log is a shared system file and\n> the transaction ID's are unique only for one server. Using\n> this information for online replication of a single database\n> to another Postgres installation will not work.\n> \n> Third, there are still some shared system catalogs across all\n> databases (pg_database, pg_group, pg_log!!!, pg_shadow and\n> pg_variable). Due to that it would be impossible (or at least\n> very, very tricky) to restore/recover (maybe point in time)\n> one single database. If you destroy one database and restore\n> it from the binary backup, these shared catalogs cannot be\n> restored too, so they're out of sync with the backup time.\n> How should the recovery now hit the right things (which\n> probably must not be there at all)?.\n> \n> All this is really a mess. I think the architecture of\n> Postgres will only allow something on query level with some\n> general API for things that must reproduce the same result\n> during recovery. For example time(). Inside the backend,\n> time() should never be called directly. Instead another\n> function is to be called that log's during normal operation\n> which time get's returned by this particular function call\n> and if the backend is in recovery mode, returns the value\n> from the log.\n> \n> And again, this all means trouble. Usually, most queries sent\n> to the database don't change any data because they are\n> SELECT's. It would dramatically blow up the log amount if you\n> log ALL queries instead of only those that modify things. But\n> when the query begins, you don't know this, because a SELECT\n> might call a function that uses SPI to UPDATE something else.\n> So the decision if the query must be logged or not can only\n> be made when the query is done (by having some global\n> variable where the heap access methods set a flag that\n> something got written). Now you have to log function call's\n> like time() even if the query will not modify any single row\n> in the database because the query is a\n> \n> SELECT 'now'::datetime - updtime FROM ...\n> \n> Doing this on a table with thousands of rows will definitely\n> waste much logging space and slowdown the whole thing by\n> unnecessary logging.\n> \n> Maybe it's a compromise if at each query start the actual\n> time and other such information is remembered by the backend,\n> all time() calls return this remembered value instead of the\n> real one (wouldn't be bad anyway IMHO), and this information\n> is logged only if the query is to be logged.\n> \n> Finally I think I must have missed some more problems, but\n> aren't these enough already to frustrate you :-?\n> \n> \n> Jan\n> \n> --\n> \n> #=============================================================\n> =========#\n> # It's easier to get forgiveness for being wrong than for \n> being right. #\n> # Let's break this rule - forgive me. \n> #\n> #======================================== [email protected] \n> (Jan Wieck) #\n> \n> \n> \n", "msg_date": "Fri, 5 Mar 1999 13:02:11 -0600", "msg_from": "\"Jackson, DeJuan\" <[email protected]>", "msg_from_op": true, "msg_subject": "RE: [HACKERS] RE: [GENERAL] Transaction logging" } ]
[ { "msg_contents": "I like it. \n\n\t-----Original Message-----\n\tFrom:\tJackson, DeJuan [SMTP:[email protected]]\n\tSent:\tFriday, March 05, 1999 12:02 PM\n\tTo:\[email protected]; Michael Davis\n\tCc:\[email protected]; [email protected]\n\tSubject:\tRE: [HACKERS] RE: [GENERAL] Transaction logging\n\n\tWouldn't it be possible to have the commit take care of the logging.\nI\n\tdon't know enough about the backend, but all those functions have to\nget\n\tevaluated at some point before the commit in order to be written to\nthe\n\tdb. You could actually have the commit log the needed sql-commands\n\twithout any function interference. Also doesn't commit know if any\nrows\n\thave been updated by any of the transaction's statements. This\n\tcould/would require filtering rules and triggers from the output, or\n\tturning them off on the restore.\n\tAs far as the shared-tables/catalogues are concerned. We could have\n\teach full-database dump take care of it's entries into the shared\n\ttables. Having the transaction logs be incremental from the point\nof\n\tthe dump allows the catalogues to be self updating (assuming we are\n\tlogging vacuums).\n\n\tAll this depends on getting a database commit to output it's\ninformation\n\teasily. It also ignores MVCC's effect on transactions, but I think\nit's\n\ta workable model.\n\n\tYou can tell me I'm blowing smoke now.\n\t\t-DEJ\n\n\t> -----Original Message-----\n\t> From: [email protected] [mailto:[email protected]]\n\t> Sent: Friday, March 05, 1999 12:24 PM\n\t> To: [email protected]\n\t> Cc: [email protected]; [email protected]\n\t> Subject: Re: [HACKERS] RE: [GENERAL] Transaction logging\n\t> \n\t> \n\t> Michael Davis wrote:\n\t> \n\t> >\n\t> > What would it take to have transaction logging added to \n\t> Postgres. I am a\n\t> > c/c++ programmer and will consider contributing to the \n\t> Postgres development\n\t> > effort. I really like everything I see and read about \n\t> Postgres. As a\n\t> \n\t> I spent some time on transaction logging since it's a feature\n\t> I'm missing too. There are mainly two different transaction\n\t> log mechanisms out.\n\t> \n\t> 1. Log queries sent to the backend.\n\t> \n\t> 2. Log images of inserted/updated rows and row ID's of\n\t> deleted ones.\n\t> \n\t> The query level logging will write less information if\n\t> queries usually affect a large number of rows. Unfortunately\n\t> the extensibility of Postgres work's against this approach.\n\t> There could be any number of user written functions who's\n\t> results aren't reproduceable during recovery. And core\n\t> features of Postgres itself would introduce the same problem.\n\t> Have a sequence which is used to create default values for\n\t> multiple tables, so that one ID is unique across them. Now\n\t> two backends insert (with INSERT ... SELECT) concurrently\n\t> into different tables using the same sequence. It's a\n\t> classic race condition and it depends on context switching\n\t> and page faults which backend will get which sequence\n\t> numbers. You cannot foresee and you cannot reproduce, except\n\t> you hook into the sequence generator and log this too. Later\n\t> when recovering, another hook into the sequence generator\n\t> must reproduce the logged results on the per\n\t> backend/transaction/command base, and the same must be done\n\t> for each function that usually returns unreproduceable\n\t> results (anything dealing with time, pid's, etc.).\n\t> \n\t> As said, this must also cover user functions. So at least\n\t> there must be a general log API that provides such a\n\t> functionality for user written functions.\n\t> \n\t> The image logging approach also has problems. First, the only\n\t> thing given to the heap access methods to outdate a tuple on\n\t> update/delete is the current tuple ID (information that tells\n\t> which tuple in which block is meant). So you need to save\n\t> the database files in binary format, because during the\n\t> actually existing dump/restore this could change and the\n\t> logged CTID's would hit the wrong tuples.\n\t> \n\t> Second, you must remember in the log which transaction ID\n\t> these informations came from and later if the transaction\n\t> committed or not, so the recovery can set this commit/abort\n\t> information in pg_log too. pg_log is a shared system file and\n\t> the transaction ID's are unique only for one server. Using\n\t> this information for online replication of a single database\n\t> to another Postgres installation will not work.\n\t> \n\t> Third, there are still some shared system catalogs across all\n\t> databases (pg_database, pg_group, pg_log!!!, pg_shadow and\n\t> pg_variable). Due to that it would be impossible (or at least\n\t> very, very tricky) to restore/recover (maybe point in time)\n\t> one single database. If you destroy one database and restore\n\t> it from the binary backup, these shared catalogs cannot be\n\t> restored too, so they're out of sync with the backup time.\n\t> How should the recovery now hit the right things (which\n\t> probably must not be there at all)?.\n\t> \n\t> All this is really a mess. I think the architecture of\n\t> Postgres will only allow something on query level with some\n\t> general API for things that must reproduce the same result\n\t> during recovery. For example time(). Inside the backend,\n\t> time() should never be called directly. Instead another\n\t> function is to be called that log's during normal operation\n\t> which time get's returned by this particular function call\n\t> and if the backend is in recovery mode, returns the value\n\t> from the log.\n\t> \n\t> And again, this all means trouble. Usually, most queries sent\n\t> to the database don't change any data because they are\n\t> SELECT's. It would dramatically blow up the log amount if you\n\t> log ALL queries instead of only those that modify things. But\n\t> when the query begins, you don't know this, because a SELECT\n\t> might call a function that uses SPI to UPDATE something else.\n\t> So the decision if the query must be logged or not can only\n\t> be made when the query is done (by having some global\n\t> variable where the heap access methods set a flag that\n\t> something got written). Now you have to log function call's\n\t> like time() even if the query will not modify any single row\n\t> in the database because the query is a\n\t> \n\t> SELECT 'now'::datetime - updtime FROM ...\n\t> \n\t> Doing this on a table with thousands of rows will definitely\n\t> waste much logging space and slowdown the whole thing by\n\t> unnecessary logging.\n\t> \n\t> Maybe it's a compromise if at each query start the actual\n\t> time and other such information is remembered by the backend,\n\t> all time() calls return this remembered value instead of the\n\t> real one (wouldn't be bad anyway IMHO), and this information\n\t> is logged only if the query is to be logged.\n\t> \n\t> Finally I think I must have missed some more problems, but\n\t> aren't these enough already to frustrate you :-?\n\t> \n\t> \n\t> Jan\n\t> \n\t> --\n\t> \n\t> #=============================================================\n\t> =========#\n\t> # It's easier to get forgiveness for being wrong than for \n\t> being right. #\n\t> # Let's break this rule - forgive me. \n\t> #\n\t> #======================================== [email protected] \n\t> (Jan Wieck) #\n\t> \n\t> \n\t> \n", "msg_date": "Fri, 5 Mar 1999 13:03:21 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "RE: [HACKERS] RE: [GENERAL] Transaction logging" } ]
[ { "msg_contents": "\nI've been using the cvs,which I assumed was pre 6.5, copy of postgresql in\na project I'm working on. With the recent talk of a 6.4.3 being released I\nwas wondering which version cvs actually reflects. I'm green with CVS so\nif I'm missing something obvious feel free to whack me on the head.\nThanks!\n\nI log into cvs and download like this\n\ncvs -d :pserver:[email protected]:/usr/local/cvsroot login\ncvs -d :pserver:[email protected]:/usr/local/cvsroot co pgsql\n\n->->->->->->->->->->->->->->->->->->---<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<\nJames Thompson 138 Cardwell Hall Manhattan, Ks 66506 785-532-0561 \nKansas State University Department of Mathematics\n->->->->->->->->->->->->->->->->->->---<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<\n\n\n", "msg_date": "Fri, 5 Mar 1999 13:35:53 -0600 (CST)", "msg_from": "James Thompson <[email protected]>", "msg_from_op": true, "msg_subject": "Stupid CVS question" }, { "msg_contents": "On Fri, 5 Mar 1999, James Thompson wrote:\n\n> \n> I've been using the cvs,which I assumed was pre 6.5, copy of postgresql in\n> a project I'm working on. With the recent talk of a 6.4.3 being released I\n> was wondering which version cvs actually reflects. I'm green with CVS so\n> if I'm missing something obvious feel free to whack me on the head.\n> Thanks!\n> \n> I log into cvs and download like this\n> \n> cvs -d :pserver:[email protected]:/usr/local/cvsroot login\n> cvs -d :pserver:[email protected]:/usr/local/cvsroot co pgsql\n\nwith the above, youare getting the development branch...\n\nIf you do:\n\ncvs -d :pserver:[email protected]:/usr/local/cvsroot co -rREL6_4\npgsql\n\nyou'll get the \"stable\" branch\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Fri, 5 Mar 1999 16:15:25 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Stupid CVS question" } ]
[ { "msg_contents": "I create a group:\n\n\tINSERT INTO pg_group VALUES ('dba');\n\nI create the following user:\n\nCREATE USER admin IN GROUP dba;\n\nNext: \n\ngrant all on table_name to GROUP dba;\nselect * from pg_class where relname = 'table_name';\n\nPsql aborts with the following error:\n\n\tERROR: parser: parse error at or near \";\"\n\tmp=> select * From pg_class where relname = 'status';;\n\tpqReadData() -- backend closed the channel unexpectedly.\n\t This probably means the backend terminated abnormally before\nor while processing the request.\n\tWe have lost the connection to the backend, so further processing is\nimpossible. Terminating.\n\nAny ideas on what could be causing this and/or how to work around it? I am\nrunning 6.4.2 on linux Red Hat 5.1.\n\nThanks, Michael\n", "msg_date": "Fri, 5 Mar 1999 17:14:35 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "permissions problem" } ]
[ { "msg_contents": "The following appears to have fixed my problems:\n\n\t\tINSERT INTO pg_group VALUES ('dba', 1);\n\nThe 1 seems to be the trick.\n\n\t-----Original Message-----\n\tFrom:\tMichael Davis \n\tSent:\tFriday, March 05, 1999 4:15 PM\n\tTo:\t'[email protected]'; [email protected]\n\tSubject:\t[GENERAL] permissions problem\n\n\tI create a group:\n\n\t\tINSERT INTO pg_group VALUES ('dba');\n\n\tI create the following user:\n\n\tCREATE USER admin IN GROUP dba;\n\n\tNext: \n\n\tgrant all on table_name to GROUP dba;\n\tselect * from pg_class where relname = 'table_name';\n\n\tPsql aborts with the following error:\n\n\t\tERROR: parser: parse error at or near \";\"\n\t\tmp=> select * From pg_class where relname = 'status';;\n\t\tpqReadData() -- backend closed the channel unexpectedly.\n\t\t This probably means the backend terminated\nabnormally before\n\tor while processing the request.\n\t\tWe have lost the connection to the backend, so further\nprocessing is\n\timpossible. Terminating.\n\n\tAny ideas on what could be causing this and/or how to work around\nit? I am\n\trunning 6.4.2 on linux Red Hat 5.1.\n\n\tThanks, Michael\n", "msg_date": "Fri, 5 Mar 1999 19:05:33 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "RE: [GENERAL] permissions problem" } ]
[ { "msg_contents": "The following select statement produces the provided error below:\n\nselect * from pg_tables;\nERROR: cache lookup of attribute 0 in relation 6187373 failed\n\nThis happens after creating view Membership_Addresses (provided below). I\nalso included the function used in Membership_Addresss. Selecting from the\nview Membership_Addresses works great (Select * from Membership_Addresses;).\nWhat is causing the error in the pg_tables view? If I drop view\nMembership_Addresses the select * from pg_tables works great.\n\nCREATE FUNCTION FormatAddress(ADDRESSES, int2) RETURNS varchar AS '\n\tDECLARE \n\t\tadr ALIAS FOR $1;\n\t\tincludecountry ALIAS FOR $2;\nBEGIN\n\t\tif (adr.address is null) then\n\t\t\tstr1 := '''';\n\t\telse\n\t\t\tstr1 := adr.address || ''\\n''; \n\t\tend if;\n\t\t... lots of other stuff ...\n\t\treturn str1;\n END; ' LANGUAGE 'plpgsql';\n\nCREATE VIEW Membership_Addresses as\n\tSELECT addressid, memberid, address, city, state, zipcode, country,\nphone, altphone, fax,\n\t\tFormatAddress(addresses, 0) AS MemberAddress\n\tFROM addresses\n\tWHERE active = -1;\n\n\n\n\t-----Original Message-----\n\tFrom:\[email protected] [SMTP:[email protected]]\n\tSent:\tFriday, March 05, 1999 11:24 AM\n\tTo:\tMichael Davis\n\tCc:\[email protected]; [email protected]\n\tSubject:\tRe: [HACKERS] RE: [GENERAL] Transaction logging\n\n\tMichael Davis wrote:\n\n\t>\n\t> What would it take to have transaction logging added to Postgres.\nI am a\n\t> c/c++ programmer and will consider contributing to the Postgres\ndevelopment\n\t> effort. I really like everything I see and read about Postgres.\nAs a\n\n\t I spent some time on transaction logging since it's a feature\n\t I'm missing too. There are mainly two different transaction\n\t log mechanisms out.\n\n\t 1. Log queries sent to the backend.\n\n\t 2. Log images of inserted/updated rows and row ID's of\n\t deleted ones.\n\n\t The query level logging will write less information if\n\t queries usually affect a large number of rows. Unfortunately\n\t the extensibility of Postgres work's against this approach.\n\t There could be any number of user written functions who's\n\t results aren't reproduceable during recovery. And core\n\t features of Postgres itself would introduce the same problem.\n\t Have a sequence which is used to create default values for\n\t multiple tables, so that one ID is unique across them. Now\n\t two backends insert (with INSERT ... SELECT) concurrently\n\t into different tables using the same sequence. It's a\n\t classic race condition and it depends on context switching\n\t and page faults which backend will get which sequence\n\t numbers. You cannot foresee and you cannot reproduce, except\n\t you hook into the sequence generator and log this too. Later\n\t when recovering, another hook into the sequence generator\n\t must reproduce the logged results on the per\n\t backend/transaction/command base, and the same must be done\n\t for each function that usually returns unreproduceable\n\t results (anything dealing with time, pid's, etc.).\n\n\t As said, this must also cover user functions. So at least\n\t there must be a general log API that provides such a\n\t functionality for user written functions.\n\n\t The image logging approach also has problems. First, the only\n\t thing given to the heap access methods to outdate a tuple on\n\t update/delete is the current tuple ID (information that tells\n\t which tuple in which block is meant). So you need to save\n\t the database files in binary format, because during the\n\t actually existing dump/restore this could change and the\n\t logged CTID's would hit the wrong tuples.\n\n\t Second, you must remember in the log which transaction ID\n\t these informations came from and later if the transaction\n\t committed or not, so the recovery can set this commit/abort\n\t information in pg_log too. pg_log is a shared system file and\n\t the transaction ID's are unique only for one server. Using\n\t this information for online replication of a single database\n\t to another Postgres installation will not work.\n\n\t Third, there are still some shared system catalogs across all\n\t databases (pg_database, pg_group, pg_log!!!, pg_shadow and\n\t pg_variable). Due to that it would be impossible (or at least\n\t very, very tricky) to restore/recover (maybe point in time)\n\t one single database. If you destroy one database and restore\n\t it from the binary backup, these shared catalogs cannot be\n\t restored too, so they're out of sync with the backup time.\n\t How should the recovery now hit the right things (which\n\t probably must not be there at all)?.\n\n\t All this is really a mess. I think the architecture of\n\t Postgres will only allow something on query level with some\n\t general API for things that must reproduce the same result\n\t during recovery. For example time(). Inside the backend,\n\t time() should never be called directly. Instead another\n\t function is to be called that log's during normal operation\n\t which time get's returned by this particular function call\n\t and if the backend is in recovery mode, returns the value\n\t from the log.\n\n\t And again, this all means trouble. Usually, most queries sent\n\t to the database don't change any data because they are\n\t SELECT's. It would dramatically blow up the log amount if you\n\t log ALL queries instead of only those that modify things. But\n\t when the query begins, you don't know this, because a SELECT\n\t might call a function that uses SPI to UPDATE something else.\n\t So the decision if the query must be logged or not can only\n\t be made when the query is done (by having some global\n\t variable where the heap access methods set a flag that\n\t something got written). Now you have to log function call's\n\t like time() even if the query will not modify any single row\n\t in the database because the query is a\n\n\t SELECT 'now'::datetime - updtime FROM ...\n\n\t Doing this on a table with thousands of rows will definitely\n\t waste much logging space and slowdown the whole thing by\n\t unnecessary logging.\n\n\t Maybe it's a compromise if at each query start the actual\n\t time and other such information is remembered by the backend,\n\t all time() calls return this remembered value instead of the\n\t real one (wouldn't be bad anyway IMHO), and this information\n\t is logged only if the query is to be logged.\n\n\t Finally I think I must have missed some more problems, but\n\t aren't these enough already to frustrate you :-?\n\n\n\tJan\n\n\t--\n\n\t\n#======================================================================#\n\t# It's easier to get forgiveness for being wrong than for being\nright. #\n\t# Let's break this rule - forgive me.\n#\n\t#======================================== [email protected] (Jan\nWieck) #\n\n\t\n", "msg_date": "Fri, 5 Mar 1999 19:54:13 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "View causes error in pg_tables" } ]
[ { "msg_contents": "I know, it sounds like a personal problem. Well it is but I am hoping\nsomeone cant point in the right direction. I am trying to get a handle on\nthe whole security thing.\n\nselect * from status;\nNOTICE: in_group: group 4 not found\nERROR: status: Permission denied.\n\nDoes anyone know what this means? Here is how to duplicate the problem:\n\nAdd group 4:\n\nINSERT INTO pg_group VALUES ('dba', 4);\n\nCreate a user in that group\n\nCREATE USER \"Admin\" IN GROUP dba;\n\nGrant permissions on group dba: \n\nGRANT ALL ON status to GROUP dba; \n\nLog into psql as Admin and select * from status. You should get the error\nabove. Any ideas what I am doing wrong. I am running PostgreSQL 6.4.2 on\nRed Hat 5.1. Access97 does not want to all me write access to any of my\ntables. Once I install any kind of permissions (using grant), it would let\nme even open the tables.\n", "msg_date": "Fri, 5 Mar 1999 20:28:16 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "Permissions problems" } ]
[ { "msg_contents": "This is another example of why not allowing utility functions in SPI would\nbe a Bad Thing.\n\nFor what it's worth, I found another case in libpq where you can get a T\nmessage without a D that my utility patch needs to handle. I have attached\nthe updated patch against the 6.4.2 version of\nsrc/interfaces/libpq/fe-exec.c\n\nJerry\n\n\n> -----Original Message-----\n> From: [email protected] [SMTP:[email protected]]\n> Sent: Friday, March 05, 1999 3:00 AM\n> To: Michael Davis\n> Cc: [email protected]\n> Subject: Re: [HACKERS] for row in select loop question\n>\n> >\n>\n> Second REVOKE and GRANT are utility statements not supported\n> for prepared SPI plans and thus PL/pgSQL currently cannot\n> execute them.\n>\n>\n> Jan\n>\n> --\n>\n>\n>#======================================================================#\n> # It's easier to get forgiveness for being wrong than for being\n>right. #\n> # Let's break this rule - forgive me.\n>#\n> #======================================== [email protected] (Jan\n>Wieck) #\n>", "msg_date": "Sat, 6 Mar 1999 12:44:05 +0900", "msg_from": "\"Gerald L. Gay\" <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] for row in select loop question" }, { "msg_contents": "Older patch reversed out, and new patch applied.\n\nThanks.\n\n[Charset iso-8859-1 unsupported, filtering to ASCII...]\n> This is another example of why not allowing utility functions in SPI would\n> be a Bad Thing.\n> \n> For what it's worth, I found another case in libpq where you can get a T\n> message without a D that my utility patch needs to handle. I have attached\n> the updated patch against the 6.4.2 version of\n> src/interfaces/libpq/fe-exec.c\n> \n> Jerry\n> \n> \n> > -----Original Message-----\n> > From: [email protected] [SMTP:[email protected]]\n> > Sent: Friday, March 05, 1999 3:00 AM\n> > To: Michael Davis\n> > Cc: [email protected]\n> > Subject: Re: [HACKERS] for row in select loop question\n> >\n> > >\n> >\n> > Second REVOKE and GRANT are utility statements not supported\n> > for prepared SPI plans and thus PL/pgSQL currently cannot\n> > execute them.\n> >\n> >\n> > Jan\n> >\n> > --\n> >\n> >\n> >#======================================================================#\n> > # It's easier to get forgiveness for being wrong than for being\n> >right. #\n> > # Let's break this rule - forgive me.\n> >#\n> > #======================================== [email protected] (Jan\n> >Wieck) #\n> >\n\n[Attachment, skipping...]\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 14 Mar 1999 11:46:44 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] for row in select loop question" } ]
[ { "msg_contents": "Hi,\n\ne.g.\ncreate table super (name varchar(50));\ncreate table sub (val int4) inherits (super);\ncreate table second (name varchar(50), supernr int4 references\nsuper(oid));\n\nis there any way to get a chart (er, omt), etc. out of a database? I\nknow that this would be a great programming work but are there any\nalready fixed solutions? The output does not have to be \"graphically\" a\nkind of textinfo like: \"super is a superclass of sub and offers name to\nsub\" or \"table second has a field supernr which is a foreign key of\nsuper\" would be okay too.\n\nMArtin\n", "msg_date": "Sat, 06 Mar 1999 09:41:50 +0100", "msg_from": "Martin =?iso-8859-1?Q?M=F6derndorfer?=\n\t<[email protected]>", "msg_from_op": true, "msg_subject": "ER, OMT chart of a database" } ]
[ { "msg_contents": "\"Stan Ng\" <[email protected]> writes:\n> I'm running into a limit problem with CREATE FUNCTION. Whenever I exceed\n> the limit on parameters, somewhere around 12 for me, psql bombs on me\n> with the \"pqReadData() -- backend closed the channel unexpectedly\" error\n> message. Is there any way to increase the parameter limit for\n> functions?\n\nIt shouldn't be letting you get that high; the system limit is 8\nparameters for a function. (It looks to me like parse_func.c should\nbe, but is failing to, verify that the length of the parameter lists\nit is passed are within bounds --- or else gram.y itself should complain\nif the list is overlength.)\n\nIncreasing the limit would be rather a major project I suspect.\nA quick look around the sources found a lot of dependencies, some\nbeing symbolic constants (not all the same one :-() and some being\nhard coded \"8\"s. Ick. Also, you would certainly be facing an initdb\nand database reload, because the number of argument-type columns in\npg_proc would have to increase.\n\nIt'd be good to clean this up and replace pg_proc's fixed-size\nproargtypes column with a variable-size \"array of oid\" column.\nAny volunteers?\n\n\t\t\tregards, tom lane\n", "msg_date": "Sat, 06 Mar 1999 12:20:49 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": true, "msg_subject": "Re: Number of parameters in a sql function" }, { "msg_contents": "Looks like this was fixed, because we now generate an error:\n\ntest=> create function klasd (int, int, int, int, int, int, int, int,\nint) returns int as 'asdf' language 'c';\nERROR: Procedures cannot take more than 8 arguments\n\n\n> \"Stan Ng\" <[email protected]> writes:\n> > I'm running into a limit problem with CREATE FUNCTION. Whenever I exceed\n> > the limit on parameters, somewhere around 12 for me, psql bombs on me\n> > with the \"pqReadData() -- backend closed the channel unexpectedly\" error\n> > message. Is there any way to increase the parameter limit for\n> > functions?\n> \n> It shouldn't be letting you get that high; the system limit is 8\n> parameters for a function. (It looks to me like parse_func.c should\n> be, but is failing to, verify that the length of the parameter lists\n> it is passed are within bounds --- or else gram.y itself should complain\n> if the list is overlength.)\n> \n> Increasing the limit would be rather a major project I suspect.\n> A quick look around the sources found a lot of dependencies, some\n> being symbolic constants (not all the same one :-() and some being\n> hard coded \"8\"s. Ick. Also, you would certainly be facing an initdb\n> and database reload, because the number of argument-type columns in\n> pg_proc would have to increase.\n> \n> It'd be good to clean this up and replace pg_proc's fixed-size\n> proargtypes column with a variable-size \"array of oid\" column.\n> Any volunteers?\n> \n> \t\t\tregards, tom lane\n> \n> \n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 9 May 1999 12:02:40 -0400 (EDT)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Number of parameters in a sql function" }, { "msg_contents": "Bruce Momjian <[email protected]> writes:\n> Looks like this was fixed, because we now generate an error:\n> test=> create function klasd (int, int, int, int, int, int, int, int,\n> int) returns int as 'asdf' language 'c';\n> ERROR: Procedures cannot take more than 8 arguments\n\nI'm not sure if this is a sufficient defense or not --- IIRC, there were\nplaces in the parser that stuffed info into size-8 arrays without any\nnearby check that no more than 8 items would be forthcoming. Offhand\nthe obvious try doesn't crash it:\n\nregression=> select foo(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18);\nERROR: No such function 'foo' with the specified attributes\n\nbut I am hoping to go back and look more carefully.\n\n\t\t\tregards, tom lane\n", "msg_date": "Sun, 09 May 1999 13:40:05 -0400", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Number of parameters in a sql function " } ]
[ { "msg_contents": "\nJust trying to clean up alot of old code, and came across this:\n\nacctng=> select max(start_time) from radhist where userid='billchuck';\nmax\n---\n \n(1 row)\n\nacctng=> select * from radhist where userid='billchuck';\nuniq_id|term_server|userid|ip_addr|port|start_time|stop_time|slipl\n-------+-----------+------+-------+----+----------+---------+-----\n(0 rows)\n\nacctng=> \n\nIMHO, the first select should return 0 rows, not 1, if no values are\nfound...no?\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Sat, 6 Mar 1999 16:28:28 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": true, "msg_subject": "a \"bug\" in v6.4.3 ..." }, { "msg_contents": "> acctng=> select max(start_time) from radhist where userid='billchuck';\n> max\n> ---\n> \n> (1 row)\n> IMHO, (this) should return 0 rows, not 1, if no values are\n> found...no?\n\nWe've had discussions on this. I vaguely recall that others disagreed\nwith my opinion that this behavior is consistant. Can't recall if others\nhad standards on their side ;)\n\nGiven that NULL means \"don't know\", I'm happy with the current result...\n\n - Tom\n", "msg_date": "Tue, 09 Mar 1999 02:28:31 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] a \"bug\" in v6.4.3 ..." } ]
[ { "msg_contents": "Hello.\n\nThe web page does not, so far as I can tell, detail how to get access to\nthe PostgreSQL CVS tree. I just want read access so that I can take a look\nat the latest source and do some poking around.\n\n1) Can someone tell me how to get to it, and\n\n2) Can someone please put it in the download section of the web page?\n\nThanks!\n\n--\nTodd Graham Lewis 32�49'N,83�36'W (800) 719-4664, x2804\n******Linux****** MindSpring Enterprises [email protected]\n\n\"A pint of sweat will save a gallon of blood.\" -- George S. Patton\n\n", "msg_date": "Sat, 6 Mar 1999 16:29:43 -0500 (EST)", "msg_from": "Todd Graham Lewis <[email protected]>", "msg_from_op": true, "msg_subject": "CVS access?" }, { "msg_contents": "On Sat, 6 Mar 1999, Todd Graham Lewis wrote:\n\n> Hello.\n> \n> The web page does not, so far as I can tell, detail how to get access to\n> the PostgreSQL CVS tree. I just want read access so that I can take a look\n> at the latest source and do some poking around.\n> \n> 1) Can someone tell me how to get to it, and\n> \n\nLogin\ncvs -d :pserver:[email protected]:/usr/local/cvsroot login\n password is 'postgresql'\n\nDevelopment copy\ncvs -d :pserver:[email protected]:/usr/local/cvsroot co pgsql\n\nStable copy (i think)\ncvs -d :pserver:[email protected]:/usr/local/cvsroot co -rREL6_4 pgsql\n\n->->->->->->->->->->->->->->->->->->---<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<\nJames Thompson 138 Cardwell Hall Manhattan, Ks 66506 785-532-0561 \nKansas State University Department of Mathematics\n->->->->->->->->->->->->->->->->->->---<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<-<\n\n\n\n", "msg_date": "Sat, 6 Mar 1999 15:51:56 -0600 (CST)", "msg_from": "James Thompson <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] CVS access?" }, { "msg_contents": "Todd Graham Lewis <[email protected]> writes:\n> The web page does not, so far as I can tell, detail how to get access to\n> the PostgreSQL CVS tree. I just want read access so that I can take a look\n> at the latest source and do some poking around.\n> 1) Can someone tell me how to get to it, and\n> 2) Can someone please put it in the download section of the web page?\n\nThere is a CVS FAQ, which you'll find off the Getting Help/Documentation\npage of the website. I agree it oughta be linked from the Download page\nas well...\n\nIt's a little bit out of date --- it recommends cvs 1.9 but the latest\nis 1.10. That nasty permissions bug it mentions *is* fixed in 1.10 BTW.\n\n\t\t\tregards, tom lane\n", "msg_date": "Sat, 06 Mar 1999 18:43:59 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] CVS access? " }, { "msg_contents": "> > The web page does not, so far as I can tell, detail how to get \n> > access to the PostgreSQL CVS tree.\n> There is a CVS FAQ, which you'll find off the Getting \n> Help/Documentation page of the website.\n\nI've started a chapter for the sgml docs (cvs.sgml) which includes the\nCVS_FAQ and various e-mail communications from scrappy about the CVS\ntree structure. Haven't finished marking it up though...\n\n - Tom\n", "msg_date": "Tue, 09 Mar 1999 02:34:09 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] CVS access?" } ]
[ { "msg_contents": "When will updateable views be supported?\n\n\nThanks, Michael\n", "msg_date": "Sat, 6 Mar 1999 19:15:40 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "Updatable views?" }, { "msg_contents": ">\n> When will updateable views be supported?\n\n Since v6.4 :-)\n\n Look at the rule system documentation in the programmers\n manual for details.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Mon, 8 Mar 1999 09:38:16 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Updatable views?" } ]
[ { "msg_contents": "Holm Tiffe wrote:\n\n> Hi all,\n\nHi, Holm,\n\n> While compiling of 6.4.2 I've realized that int8's are not supported\n> on FreeBSD by default. The Configure script looks if %ld or %lld\n> will do the job but FreeBSD is using %qd as the long long int format.\n> After a little fiddeling around with the configure script and\n> include/utils/int8.h, I've got the int8 type working.\n> Since I've read somtimes in this Mailinglists, that FreeBSD is one\n> of the development platforms for PostgreSQL, I wonder why PGSQL\n> doesn't support FreeBSD's own format ...\n\nHere are the patches I've used to enable int8 support.\nSince my patches modify configure.in, you'll need to install\nautoconf and regenerate the configure script.\n\nHope this helps,\nKevin.\n\n==============================================================\n\n--- ./src/backend/port/snprintf.c 1998/12/25 02:20:41 64.2\n+++ ./src/backend/port/snprintf.c 1999/01/19 00:37:40 64.2.1.1\n\n@@ -49,7 +49,7 @@\n #include <sys/param.h>\n\n /* IRIX doesn't do 'long long' in va_arg(), so use a typedef */\n-#ifdef HAVE_LONG_LONG_INT_64\n+#if defined(HAVE_LONG_INT_64) || defined(HAVE_LONG_LONG_INT) ||\ndefined(HAVE_QUAD_INT_64)\n typedef long long long_long;\n typedef unsigned long long ulong_long;\n #endif\n--- ./src/include/utils/int8.h 1998/09/11 17:16:11 64.2\n+++ ./src/include/utils/int8.h 1999/01/19 00:37:41 64.2.1.1\n@@ -35,11 +35,18 @@\n\n #define INT64_FORMAT \"%lld\"\n #else\n+#ifdef HAVE_QUAD_INT_64\n+/* We have working support for \"long long\", use that */\n+typedef long long int64;\n+\n+#define INT64_FORMAT \"%qd\"\n+#else\n /* Won't actually work, but fall back to long int so that int8.c\ncompiles\n*/\n typedef long int int64;\n\n #define INT64_FORMAT \"%ld\"\n #define INT64_IS_BUSTED\n+#endif\n #endif\n #endif\n\n--- ./src/include/config.h.in 1998/12/13 20:08:24 64.2\n+++ ./src/include/config.h.in 1999/01/19 00:37:41 64.2.1.1\n@@ -267,6 +267,9 @@\n /* Set to 1 if type \"long long int\" works and is 64 bits */\n #undef HAVE_LONG_LONG_INT_64\n\n+/* Set to 1 if type \"long long\" works and is 64 bits */\n+#undef HAVE_QUAD_INT_64\n+\n /* Define as the base type of the last arg to accept */\n #undef SOCKET_SIZE_TYPE\n\n--- ./src/configure.in 1998/12/13 20:08:20 64.2\n+++ ./src/configure.in 1999/01/19 00:37:39 64.2.1.1\n@@ -688,6 +688,43 @@\n AC_MSG_RESULT(no),\n AC_MSG_RESULT(assuming not on target machine))\n\n+AC_MSG_CHECKING(whether 'long long' is 64 bits)\n+AC_TRY_RUN([#include <stdio.h>\n+typedef long long int64;\n+#define INT64_FORMAT \"%qd\"\n+\n+int64 a = 20000001;\n+int64 b = 40000005;\n+\n+int does_int64_work()\n+{\n+ int64 c,d,e;\n+ char buf[100];\n+\n+ if (sizeof(int64) != 8)\n+ return 0; /* doesn't look like the right size */\n+\n+ /* we do perfunctory checks on multiply, divide, sprintf, sscanf */\n+ c = a * b;\n+ sprintf(buf, INT64_FORMAT, c);\n+ if (strcmp(buf, \"800000140000005\") != 0)\n+ return 0; /* either multiply or sprintf is busted\n*/\n+ if (sscanf(buf, INT64_FORMAT, &d) != 1)\n+ return 0;\n+ if (d != c)\n+ return 0;\n+ e = d / b;\n+ if (e != a)\n+ return 0;\n+ return 1;\n+}\n+main() {\n+ exit(! does_int64_work());\n+}],\n+ [AC_DEFINE(HAVE_QUAD_INT_64) AC_MSG_RESULT(yes)],\n+ AC_MSG_RESULT(no),\n+ AC_MSG_RESULT(assuming not on target machine))\n+\n dnl Checks for library functions.\n AC_FUNC_MEMCMP\n AC_TYPE_SIGNAL\n\n\n\n", "msg_date": "Sun, 07 Mar 1999 12:53:20 +0800", "msg_from": "Kevin Lo <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "Kevin Lo <[email protected]> writes:\n>> While compiling of 6.4.2 I've realized that int8's are not supported\n>> on FreeBSD by default. The Configure script looks if %ld or %lld\n>> will do the job but FreeBSD is using %qd as the long long int format.\n\n> Here are the patches I've used to enable int8 support.\n\nActually, given the way that the 6.5 sources currently stand, the right\nlong-term solution is to make two separate tests: (1) does the compiler\ndo 64-bit arithmetic correctly, and if so then (2) does snprintf have a\nworking format for the chosen type. If (1) passes but (2) fails we have\na fallback: enable use of our own snprintf code. When I wrote the 6.4\nconfigure test for int8, we hadn't yet developed that fallback, so there\nwasn't much point in distinguishing compiler support from library\nsupport. But now we should.\n\nIf we do it this way then it doesn't matter a whole lot whether \"%qd\" is\none of the tested alternatives or not ;-).\n\nI will pursue fixing things that way in the 6.5 sources. In the\nmeantime, do we want to check Kevin's fixes into REL6_4, or is adding\nint8 support for more machines too low-priority to justify taking any\nchance of breaking 6.4.3?\n\n\t\t\tregards, tom lane\n", "msg_date": "Sun, 07 Mar 1999 11:36:48 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD " }, { "msg_contents": "On Sun, 7 Mar 1999, Tom Lane wrote:\n\n> I will pursue fixing things that way in the 6.5 sources. In the\n> meantime, do we want to check Kevin's fixes into REL6_4, or is adding\n> int8 support for more machines too low-priority to justify taking any\n> chance of breaking 6.4.3?\n\nNote that I commited int8 fixes to the v6.4.3 tree, and they are in the\ncurrent beta3 tar file...did this mid-last week...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Sun, 7 Mar 1999 19:05:39 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD " }, { "msg_contents": "The Hermit Hacker <[email protected]> writes:\n\n> On Sun, 7 Mar 1999, Tom Lane wrote:\n> \n> > I will pursue fixing things that way in the 6.5 sources. In the\n> > meantime, do we want to check Kevin's fixes into REL6_4, or is adding\n> > int8 support for more machines too low-priority to justify taking any\n> > chance of breaking 6.4.3?\n> \n> Note that I commited int8 fixes to the v6.4.3 tree, and they are in the\n> current beta3 tar file...did this mid-last week...\n\nDid you also change the #ifdef line to include __NetBSD__ for the %qd\nformat? At least NetBSD/i386 uses that. Alpha might not, I'm not\nfully certain, but I think it will work with %qd.\n\n--Michael\n", "msg_date": "07 Mar 1999 18:05:30 -0800", "msg_from": "Michael Graff <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "\nTry what I just commited now...its only in v6.5 sources though...\n\nI've modified it so that it checks for use of %lld, and if that fails,\ncheck for use of %qd, before it falls back to our homebrew snprintf...\n\n\nOn 7 Mar 1999, Michael Graff wrote:\n\n> The Hermit Hacker <[email protected]> writes:\n> \n> > On Sun, 7 Mar 1999, Tom Lane wrote:\n> > \n> > > I will pursue fixing things that way in the 6.5 sources. In the\n> > > meantime, do we want to check Kevin's fixes into REL6_4, or is adding\n> > > int8 support for more machines too low-priority to justify taking any\n> > > chance of breaking 6.4.3?\n> > \n> > Note that I commited int8 fixes to the v6.4.3 tree, and they are in the\n> > current beta3 tar file...did this mid-last week...\n> \n> Did you also change the #ifdef line to include __NetBSD__ for the %qd\n> format? At least NetBSD/i386 uses that. Alpha might not, I'm not\n> fully certain, but I think it will work with %qd.\n> \n> --Michael\n> \n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 8 Mar 1999 00:18:20 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD" }, { "msg_contents": "The Hermit Hacker <[email protected]> writes:\n> Try what I just commited now...its only in v6.5 sources though...\n> I've modified it so that it checks for use of %lld, and if that fails,\n> check for use of %qd, before it falls back to our homebrew snprintf...\n\nWhat I committed this afternoon will work on a %qd box, it'll just\nchoose to insert our homebrew snprintf instead of using the local\nversion.\n\nIs it worth adding another test cycle to configure in order to use the\nnative snprintf on these boxes? There are not that many boxes that\nuse %qd; and it's not real clear that using our own snprintf is a loss\nanyway.\n\nJust my $0.02 ...\n\n\t\t\tregards, tom lane\n", "msg_date": "Mon, 08 Mar 1999 02:12:43 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD " }, { "msg_contents": "On Mon, 8 Mar 1999, Tom Lane wrote:\n\n> The Hermit Hacker <[email protected]> writes:\n> > Try what I just commited now...its only in v6.5 sources though...\n> > I've modified it so that it checks for use of %lld, and if that fails,\n> > check for use of %qd, before it falls back to our homebrew snprintf...\n> \n> What I committed this afternoon will work on a %qd box, it'll just\n> choose to insert our homebrew snprintf instead of using the local\n> version.\n> \n> Is it worth adding another test cycle to configure in order to use the\n> native snprintf on these boxes? There are not that many boxes that\n> use %qd; and it's not real clear that using our own snprintf is a loss\n> anyway.\n\nIn all cases, using a system function is preferable to a \"home brew\nsolution\"...as its then guaranteed to work, and work optimally, for that\nsystem...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 8 Mar 1999 08:50:19 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] int 8 on FreeBSD " } ]
[ { "msg_contents": "> You asked me to re-do the tests I had run to check the GEQO\n> threshold. With today's sources I get these runtimes:\n> \n> # Tables:\t\t7\t8\t9\t10\t11\n> \n> std, with indexes\t1.89\t2.26\t5.41\t24.21\t90.04\n> GEQO, with indexes\t3.73\t6.98\t16.98\t43.81*\t21.45\n> std, no indexes\t\t1.60\t1.79\t4.78\t15.75\t84.17\n> GEQO, no indexes\t3.30\t6.04\t15.52\t18.73\t22.35\n> \n> So, (a) the number of indexes is still not very relevant,\n> and (b) it looks like we oughta kick over to GEQO at 11 tables.\n> \n> WHat were we at before? 6 or 8 I think? Nice job :-)\n> \n> \t\t\tregards, tom lane\n> \n> * That number is obviously out of line ... probably it was\n> skewed by some other activity on my system at the time.\n> I'm too lazy to repeat the measurement however.\n> \n\nDone. Thanks for doing the checking.\n\nDoesn't seem like much of an increase, but it is N!, so I guess it is:\n\n\t> 6!\n\t 720\n\t> 8!\n\t 40320\n\t> 11!\n\t 39916800\n\t\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 7 Mar 1999 07:02:35 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: GEQO threshold results" } ]
[ { "msg_contents": "I just want to go on record stating that I don't think we need a 6.4.3\nrelease because I am concerned about making a more buggy release than\n6.4.2 because of inadequate testing and 6.5 retrofitting problems.\n\nI also don't like to see effort expended on it rather than 6.5. \nHowever, I realize many people disagree with me.\n\nI am ready to package up the 6.4.3 file changes needed when requested.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 7 Mar 1999 13:53:54 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "6.4.3 release" }, { "msg_contents": "On Sun, 7 Mar 1999, Bruce Momjian wrote:\n\n> I just want to go on record stating that I don't think we need a 6.4.3\n> release because I am concerned about making a more buggy release than\n> 6.4.2 because of inadequate testing and 6.5 retrofitting problems.\n\nI would like to roll out a v6.4.3 the day we make v6.5 beta, myself ...\nsince only bug fixes were supposed to go into the v6.4.3 source tree, and\neveryone would obviously have followed that simple rule to the letter,\nthere shouldn't be any more bugs in v6.4.3 then v6.4.2...right? :)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Sun, 7 Mar 1999 19:07:48 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] 6.4.3 release" }, { "msg_contents": "The Hermit Hacker <[email protected]> writes:\n> I would like to roll out a v6.4.3 the day we make v6.5 beta, myself ...\n> since only bug fixes were supposed to go into the v6.4.3 source tree, and\n> everyone would obviously have followed that simple rule to the letter,\n> there shouldn't be any more bugs in v6.4.3 then v6.4.2...right? :)\n\nOf course, the fact that the REL6_4 tree has been broken more than once\nsays that we've been sloppy :-(.\n\nI'd suggest trying to get 6.4.3 out the door a week or so before 6.5\ngoes beta --- that way, if anything is badly broken, we have time to\ndeal with a 6.4.4 before the 6.5 beta cycle starts consuming all the\nattention.\n\nWhat is the current plan for 6.5 beta anyway? Last I heard was\n1 Feb, which we definitely have not met...\n\n\t\t\tregards, tom lane\n", "msg_date": "Mon, 08 Mar 1999 02:18:34 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] 6.4.3 release " }, { "msg_contents": "On Mon, 8 Mar 1999, Tom Lane wrote:\n\n> The Hermit Hacker <[email protected]> writes:\n> > I would like to roll out a v6.4.3 the day we make v6.5 beta, myself ...\n> > since only bug fixes were supposed to go into the v6.4.3 source tree, and\n> > everyone would obviously have followed that simple rule to the letter,\n> > there shouldn't be any more bugs in v6.4.3 then v6.4.2...right? :)\n> \n> Of course, the fact that the REL6_4 tree has been broken more than once\n> says that we've been sloppy :-(.\n> \n> I'd suggest trying to get 6.4.3 out the door a week or so before 6.5\n> goes beta --- that way, if anything is badly broken, we have time to\n> deal with a 6.4.4 before the 6.5 beta cycle starts consuming all the\n> attention.\n> \n> What is the current plan for 6.5 beta anyway? Last I heard was\n> 1 Feb, which we definitely have not met...\n\nWe are awaiting stuff from Vadim, I believe concerning some code\ncompletions he wants to do on the MVCC stuff...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 8 Mar 1999 08:51:07 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] 6.4.3 release " }, { "msg_contents": "> The Hermit Hacker <[email protected]> writes:\n> > I would like to roll out a v6.4.3 the day we make v6.5 beta, myself ...\n> > since only bug fixes were supposed to go into the v6.4.3 source tree, and\n> > everyone would obviously have followed that simple rule to the letter,\n> > there shouldn't be any more bugs in v6.4.3 then v6.4.2...right? :)\n> \n> Of course, the fact that the REL6_4 tree has been broken more than once\n> says that we've been sloppy :-(.\n> \n> I'd suggest trying to get 6.4.3 out the door a week or so before 6.5\n> goes beta --- that way, if anything is badly broken, we have time to\n> deal with a 6.4.4 before the 6.5 beta cycle starts consuming all the\n> attention.\n> \n> What is the current plan for 6.5 beta anyway? Last I heard was\n> 1 Feb, which we definitely have not met...\n\nWe did not specify a year. :-)\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 08:10:20 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] 6.4.3 release" }, { "msg_contents": "On Mon, 8 Mar 1999, Bruce Momjian wrote:\n\n> > The Hermit Hacker <[email protected]> writes:\n> > > I would like to roll out a v6.4.3 the day we make v6.5 beta, myself ...\n> > > since only bug fixes were supposed to go into the v6.4.3 source tree, and\n> > > everyone would obviously have followed that simple rule to the letter,\n> > > there shouldn't be any more bugs in v6.4.3 then v6.4.2...right? :)\n> > \n> > Of course, the fact that the REL6_4 tree has been broken more than once\n> > says that we've been sloppy :-(.\n> > \n> > I'd suggest trying to get 6.4.3 out the door a week or so before 6.5\n> > goes beta --- that way, if anything is badly broken, we have time to\n> > deal with a 6.4.4 before the 6.5 beta cycle starts consuming all the\n> > attention.\n> > \n> > What is the current plan for 6.5 beta anyway? Last I heard was\n> > 1 Feb, which we definitely have not met...\n> \n> We did not specify a year. :-)\n\nDamn, the loopholes that I keep leaving you guys *slap forehead* *grin*\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 8 Mar 1999 09:45:47 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] 6.4.3 release" }, { "msg_contents": "> > > > I would like to roll out a v6.4.3 the day we make v6.5 beta, \n> > > > myself ... since only bug fixes were supposed to go into the \n> > > > v6.4.3 source tree, and everyone would obviously have followed \n> > > > that simple rule to the letter, there shouldn't be any more bugs \n> > > > in v6.4.3 then v6.4.2...right? :)\n\nNope. We had a *firm* declaration that the v6.4.x tree was dead. That\nusually means done, finished, kaput. I stopped committing changes to\nthat tree, but instead posted patches in /pub/patches/ as I would think\nthe unwritten procedure would call for.\n\nA long time later we then declared that there *would* be a v6.4.3, and I\nwas forced to try retrofitting the patches onto the dead source tree. I\ndon't know what other problems other people introduced, but it is a bit\nof work to raise the dead, and I'll agree with Bruce that it is a\ndistraction.\n\nHowever, we also have the feeling that v6.5 is a bit riskier than\nprevious recent versions, so want to continue the v6.4.x thread. I would\nsuggest that those who feel that way adopt the v6.4.x tree and ensure\nthat it is a solid release. Asking everyone to devote the same attention\nto both v6.4.3 and to v6.5 will lead to gaps in test coverage which we\nshould avoid.\n\nOr, perhaps we should declare a v6.4.3 beta period in which we do some\nfocused testing, then pick up v6.5 afterwards.\n\n> > > Of course, the fact that the REL6_4 tree has been broken more than \n> > > once says that we've been sloppy :-(.\n\nSloppy? See above...\n\n - Tom\n", "msg_date": "Tue, 09 Mar 1999 03:05:11 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] 6.4.3 release" } ]
[ { "msg_contents": "\nI'd like to re-ask on this thread of a few days ago because I suspect\nthere is a deeper problem here that changing compilers simply \"hides\" in\nsome way.\n\nI have a similar problem on Digital UNIX 4.0d with postgresql 6.4.2\ncompiled with vendor CC (V5.6 using \"cc -std1\" in compile).\n\n% mv ${PGDATA} ${PGDATA}.foo\n% mkdir $PGDATA\n% initdb\n% createdb test\n% psql test\ntest=> create table val (v_num decimal);\ntest=> \\d val\nERROR: typeidTypeRelid: Invalid type - oid = 0\n\nI have the same problem in an existing database that actually has data\nin it - queries (seem to) work ok, but using psql to examine tables and\nsuch is impossible.\n\nThe last discussion of this was on AIX 4.3.2 compiled with gcc, and the\nfix (if I read the resolution of that thread properly) was to use the\nvendor xlC on AIX instead.\n\nhttp://www.postgresql.org/mhonarc/pgsql-hackers/1999-03/msg00066.html\n\nThis doesn't help me on Digital UNIX.\n\nA previous mention of something similar (with no resolution I can find\nin the archives) was in:\n\nhttp://www.postgresql.org/mhonarc/pgsql-general/1999-01/msg00230.html\n\nwith pgsql 6.4.2 on LinuxPPC R4.\n\nAnd before that (again, no resolution I can find) in:\n\nhttp://www.postgresql.org/mhonarc/pgsql-hackers/1998-11/msg00215.html\n\non 6.4 where the problem is in IRIX with vendor cc and mentioned on\nDigital UNIX 4.0b with the same \"cc -std1\" that I am using.\n\nAny advice greatly appreciated.\n\nErik Riedel\nCarnegie Mellon University\[email protected]\n\n", "msg_date": "Sun, 7 Mar 1999 16:11:20 -0500 (EST)", "msg_from": "Erik Riedel <[email protected]>", "msg_from_op": true, "msg_subject": "PostgreSQL 6.4.2: typeidTypeRelid error" } ]
[ { "msg_contents": "\nI'd like to re-ask on this thread of a few days ago because I suspect\nthere is a deeper problem here that changing compilers simply \"hides\" in\nsome way.\n\nI have a similar problem on Digital UNIX 4.0d with postgresql 6.4.2\ncompiled with vendor CC (V5.6 using \"cc -std1\" in compile).\n\n% mv ${PGDATA} ${PGDATA}.foo\n% mkdir $PGDATA\n% initdb\n% createdb test\n% psql test\ntest=> create table val (v_num decimal);\ntest=> \\d val\nERROR: typeidTypeRelid: Invalid type - oid = 0\n\nI have the same problem in an existing database that actually has data\nin it - queries (seem to) work ok, but using psql to examine tables and\nsuch is impossible.\n\nThe last discussion of this was on AIX 4.3.2 compiled with gcc, and the\nfix (if I read the resolution of that thread properly) was to use the\nvendor xlC on AIX instead.\n\nhttp://www.postgresql.org/mhonarc/pgsql-hackers/1999-03/msg00066.html\n\nThis doesn't help me on Digital UNIX.\n\nA previous mention of something similar (with no resolution I can find\nin the archives) was in:\n\nhttp://www.postgresql.org/mhonarc/pgsql-general/1999-01/msg00230.html\n\nwith pgsql 6.4.2 on LinuxPPC R4.\n\nAnd before that (again, no resolution I can find) in:\n\nhttp://www.postgresql.org/mhonarc/pgsql-hackers/1998-11/msg00215.html\n\non 6.4 where the problem is in IRIX with vendor cc and mentioned on\nDigital UNIX 4.0b with the same \"cc -std1\" that I am using.\n\nAny advice greatly appreciated.\n\nErik Riedel\nCarnegie Mellon University\[email protected]\n\n\n", "msg_date": "Sun, 7 Mar 1999 16:20:39 -0500 (EST)", "msg_from": "Erik Riedel <[email protected]>", "msg_from_op": true, "msg_subject": "PostgreSQL 6.4.2: typeidTypeRelid error" }, { "msg_contents": "Erik Riedel wrote:\n> \n> I'd like to re-ask on this thread of a few days ago because I suspect\n> there is a deeper problem here that changing compilers simply \"hides\" in\n> some way.\n> \n> I have a similar problem on Digital UNIX 4.0d with postgresql 6.4.2\n> compiled with vendor CC (V5.6 using \"cc -std1\" in compile).\n> \n> % mv ${PGDATA} ${PGDATA}.foo\n> % mkdir $PGDATA\n> % initdb\n> % createdb test\n> % psql test\n> test=> create table val (v_num decimal);\n> test=> \\d val\n> ERROR: typeidTypeRelid: Invalid type - oid = 0\n> \n> I have the same problem in an existing database that actually has data\n> in it - queries (seem to) work ok, but using psql to examine tables and\n> such is impossible.\n> \n> The last discussion of this was on AIX 4.3.2 compiled with gcc, and the\n> fix (if I read the resolution of that thread properly) was to use the\n> vendor xlC on AIX instead.\n\nThis is correct. I changed two items.\n\n 1) switched from gcc to xlc\n\n 2) changed -qchars=unsigned (was -qchars=signed)\n\nMost of the regression tests failed with -qchars=signed and all passed without\nso much as a whimper when -qchars=unsigned. Don't know why this happened, or if\nit makes any difference on your platform, but that's what AIX needs to work.\n", "msg_date": "Mon, 08 Mar 1999 07:23:22 -0600", "msg_from": "\"David R. Favor\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL 6.4.2: typeidTypeRelid error" } ]
[ { "msg_contents": "Can someone point out where MVCC is implemented on the backend? I would\nlike to Use The Source to understand what's going on, but I can't seem\nto find it! `grep -i MVCC` turns up two comments, neither of which\nseems near the implementation. I would be more than happy to write up a\ntourist's guide to the implementation of MVCC if some kind hacker would\nbless me with some pointers as to how it all fits in.\n\nTIA!\n\n--\nTodd Graham Lewis 32�49'N,83�36'W (800) 719-4664, x2804\n******Linux****** MindSpring Enterprises [email protected]\n\n\"A pint of sweat will save a gallon of blood.\" -- George S. Patton\n\n", "msg_date": "Sun, 7 Mar 1999 18:14:26 -0500 (EST)", "msg_from": "Todd Graham Lewis <[email protected]>", "msg_from_op": true, "msg_subject": "MVCC? Where?" }, { "msg_contents": "Todd Graham Lewis wrote:\n> \n> Can someone point out where MVCC is implemented on the backend? I would\n> like to Use The Source to understand what's going on, but I can't seem\n> to find it! `grep -i MVCC` turns up two comments, neither of which\n> seems near the implementation. I would be more than happy to write up a\n> tourist's guide to the implementation of MVCC if some kind hacker would\n> bless me with some pointers as to how it all fits in.\n\ngrep XACT_SERIALIZABLE :)\n\nThere was discussion in june - july '98 about implementation.\n\nAlso, I'll write something after beta started...\n\nVadim\n", "msg_date": "Tue, 09 Mar 1999 09:30:25 +0700", "msg_from": "Vadim Mikheev <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] MVCC? Where?" } ]
[ { "msg_contents": "Hi all,\n\nI send a similar problem to the SQL list but had no replies. Perhaps you\ncan help.\n\nI have noticed a problem/bug/feature using nextval and rules. What I\nwould expect from the example below is for the nextval of the sequence\nto be evaluated once and this value to be used. What appears to be\nhappening is the nextval function is being referenced and then being\nre-evaluated a number of times.\n\nHere is the example of the problem:\n\nCREATE SEQUENCE seq1;\nCREATE\n\nCREATE TABLE tab1 (\n a INTEGER\n);\nCREATE\n\nCREATE TABLE tab2 (\n b1 INTEGER,\n b2 INTEGER,\n b3 INTEGER,\n b4 INTEGER\n);\nCREATE\n\nCREATE RULE rule1 AS ON INSERT TO tab1 DO INSERT INTO tab2 VALUES\n(NEW.a, NEW.a, NEW.a,\nNEW.a);\nCREATE\n\nINSERT INTO tab1 VALUES (NEXTVAL('seq1'));\nINSERT 29288 1\n\nSELECT * FROM tab1;\na\n-\n5\n(1 row)\n\n\nSELECT * FROM tab2;\nb1|b2|b3|b4\n--+--+--+--\n 1| 2| 3| 4\n(1 row)\n\n\n----\n\nThe only way I can see of getting round this is to use triggers and c\ncode.\n\nDoes anybody have any ideas on this?\n\nThanks,\nJustin.\n\n", "msg_date": "Sun, 07 Mar 1999 23:51:40 +0000", "msg_from": "Justin Subert <[email protected]>", "msg_from_op": true, "msg_subject": "[HACKERS] Problem with nextval and rules" } ]
[ { "msg_contents": "Hi,\nWhen I try on current:\nabr=> select sum(endt-begt-pausen)/count(*)::float8 from tage;\nERROR: Unable to identify an operator '/' for types 'timespan' and 'float8'\n You will have to retype this query using an explicit cast \nbut:\nabr=> \\df\n...\ntimespan |timespan_div |timespan float8 |divide\n...\n\nWhy? Directly used I get:\nabr=> select timespan_div(sum(endt-begt-pausen),count(*)::float8) from tage;\ntimespan_div\n-----------------------------\n@ 11 hours 24 mins 34.79 secs\n(1 row) \n\nIt seems that an entry for '/' is defined:\ngrep timespan_div catalog/*.bki*\n...\ncatalog/local1_template1.bki.source:insert OID = 1585 ( \"/\" PGUID 0 b t f 1186\n1186 1186 0 0 0 0 timespan_div - - ) \n...\n\nAny clues?\n\nBye!\n----\nMichael Reifenberger\nPlaut Software GmbH, R/3 Basis\n\n", "msg_date": "Mon, 8 Mar 1999 01:15:46 +0100 (CET)", "msg_from": "Michael Reifenberger <[email protected]>", "msg_from_op": true, "msg_subject": "Typeidentifying not working sometimes?" } ]
[ { "msg_contents": "[Cc'ed to the hackers list]\n\n>-- hi there,\n>\n>I saw \"Re: [HACKERS] large objects failing (hpux10.20 sparc/solaris 2.6,\n>gcc 2.8.1\"\n>and I wonder if you could help me ( a new postgres user).\n>\n>basically I think I am seeing the same problem and want to fix/patch it.\n>Is there a central place for such patches ? or are they incorporated\n>into\n>standard release(s).\n\nIncluded are patches should fix your problem. I tested on Solaris\n2.6/gcc 2.7.2. and it worked for me.\n\nbegin 644 lobj.patch.gz\nM'XL(\"*$KXS8``VQO8FHN<&%T8V@`U5CI4ALY$/[M/$63K6)G[/&)SQS40@*L\nM*\\10D&RV*IMRR3.R/<EXY&@T=DC\"`^];;+?F\\-C8\"PFPU/H'V%)+ZO[ZZT/*\nMY_,P%8$:21Y\\]HK-4KU4*P?2+@^8_8G[3CE00K(1+WM,CGA?##YR6Y5=?]9G\nM4[=DEX1T1[G7PH?73`*TH5IYLE-]4JU#M=/I/\"H6B[?9_>K&M4:T<7[Y0[^A\nMM6.U.J!_`D`^ESL4$H)P$/#/(?<5^$+QP(+WIWL?P`W@E`4V\\V#/=^3?\\(SA\nM/_[;S&4E;D]+0[FK]RCCWT=;\\(O#AZ[/H7MX=G!P?OK\"F)JYG'&*BA]Q=2@Y\nM/Y\\RF^,H%\"%POW(Q-'[G;/HFG'K\\)5,L,QXH&=H*9DQZW,].N+[:J9DF'IB>\nM]GKOS_WC5[E<KEWIU)8FNKUHHE&E<0TS6M^N`7XK/J3U?Z$\\ZI6#ER=O]X\\/\nM]HZ[1SU##(<!5ZN@6*H_<%5@;ERV`;'\"\"F+9Y?&,(\\*!QW\\2S'74:M0;5J/1\nM2,B%PHK,\\.$Y$%GGTN=S`]G;=WA@6S`(AQ;X@PM$'+7SY])5BOOF4[V4>P'7\nM7[Y%6KM#2)<6=R.L8!<60V-W-*:]S.CD6QR]./O*+L)SKMW%`KU(3P^YQ$VU\nM`6?<XRS@^WK02.?HF$L\"E+Z4\\S`1,X[)`+W&90!3%BA08PYL(D*DYYS#QQ\"'\nM4!/%B7K$Z4:];36:]835=P?[XO,C^,.W6WF@H-=N0$M/7MZCE]`#BSU,&+,`\nMF\"<Y<RXP'00A=^`=KD_T,J-%\"-!\\[-IC+>YP6_()YA*4]01E#\\E1EOLV!YN<\nMR&6ZRC@F@3,^?$$39@G.!;DX&(O0<R@7`4Y[J?@R**@=;EM\"M\\,;IH)00*=3\nMKI7K_T:W*%O=BG'K`K]9JUO->GL1^-^_PQ6R/,L,>6).#DEDM[J*3TXC%;K!\nM'\\QS'6,[0[>Q<AU3)ZK<MXSFKN\\J%X6_(F0V\\^$31R\\--6Z.\\'EB[#)U72WZ\nM_#D87=_A7\\[QUTN<,*'W]OC8C+E?)`AIZA6_H\"P,`>[]%(<CPI7C]`9YV`NB\nMLUW:#\";L`@:()C*`J5!R[P)L3P1(!0-KM\"TF$U>9%H*:KI\\0N&'`(:KJ8(=2\nM4B&*03/F2*F*\"4H@C3+VHH?2+5!<7L![.-V##Z5DM+QDPP&)=-/EQC;98T'E\nM2\\6\"J@6'_6[O3?WHP(H6Z0_RK4NE`ZL7(A!.KD2_F:3+*]C&4:DAZ0_XR/5I\nM/.L#/2.M6)`^QD`(SP34QPB1!]6FB8IEYK7&:;Z,'>&(V%]:#KLGG0)H[Q%7\nM/O^B5OV.I@HY9]+1;G<E(NX*RCJ41YL[.U:ST5[DT?MD,66RJP0K0';BQEZ+\nM<F:<JV_BM?N/H/]/G/SW'\"9W70+5+XB($!^#!%XY(^+LD*&H!<LA<+?\\7YO5\nMFTVKV:JMMG-XQA@[U/Z0*WN<I3<-2@O.?3;%\"J9Z8IXQ'HFYC0K&8NY4R4SI\nMU5+3(3;)!LK$M?B2BBH>9\\2G(M_2SCCF6I;*5'0G`R[1Y=BX8ZA18<3J153S\nM.3++53#$9M]C&)Y`)3TH4R^`H&D2X\"Z9^'TAIA?&MCZYN*OZ\").##EB)Y3AM\nMM!\"FQ:7BP6#:2M%)DH_6PWQXE-:2J]5`U\"H1N32,;1SH++K8B-0_#R4AN9I\\\nM,U@6XK2V$;4X-`'C5(R,@[.SDS,K'M&?Q]0G:HUPP9.,@ABNZ`_G<1+HI%'2\nM;Y7SZ^_D]:;5:K222\"-)?X#-XR<*:.SD&$4IIO1>2+X[&>[K.6,<D8)H0*8D\nM2W8Q$6H8(F/U%LR).\\&QI&8X$BQ\"-2)1VDIO6G':[QV\\T[(D/L5++<I$\\Z@6\nM77*S-YK84OR+6;<[U.VD1WVE7DB\\$P*\"\";:VE*_'V&J#_HYB#EUY`2_:2,1?\nMM6!T;6]6K%8K)<<=`!3?4FX,T35&;Y&/8R2OV7J!Y0TVS=$(M0`&B5K+DN?4\nM%,32V!8D:MRU`]8QMMUL6.U6,\\O8'_7(>L[F\\)J()MW$(XL+X*8U\"=3T^`-:\nM;\"W@T?(XP5)Z#:8V\"J4O.?&Z;\"+5K&RW:U:[T\\JR\\M8@)+?GF\\-PO5FZ3*20\nM77=\"`EKA1GMKJ92DL?@ZGL8K(IXFB?%'$%]'PTZK;G62YTQR2:?3L*J5RDY:\nM1C2^D:5;SY/TK[?,7N0S#\"#=L%#JD@!8'S%:?($W>7^$P3RC\"D&5$(4R=?\"<\nMJZX_V]#PZZY>!V2VIY8<^UX_?8Z@DR_7F]AIDT5+Q0&WPAYY&N*-/\"J0]*)'\nME1PU]4//BXMU;HS=,<*ZX641XJ?%1<H&YJ#!/@A*%G2UL9('Q=+,ZWOT:I+Y\nMC:DB/@>5T!I@G:83\"V!@U[CZ`%F(C.6:CXL5QR>]H^@E,QZ[6D+V/'I(4;J;\nMCPW&6P+]B,@V<P6%FC]*QX`-`B69[FY+FAC52JV&..[4LM%ZUS@64M'L\"RT-\n2/3#(677N#>9_`,<A2^HE&0``\n`\nend\n\n", "msg_date": "Mon, 08 Mar 1999 11:08:38 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": true, "msg_subject": "Re: Sunos 5.6 ,SUN C compiler 4.2,postgresql-6.4.2 " } ]
[ { "msg_contents": "which plaut\n\n", "msg_date": "Mon, 8 Mar 1999 08:57:56 +0100 (CET)", "msg_from": "Michael Reifenberger <[email protected]>", "msg_from_op": true, "msg_subject": "None" } ]
[ { "msg_contents": "Hi,\nWhen I try on current:\nabr=> select sum(endt-begt-pausen)/count(*)::float8 from tage;\nERROR: Unable to identify an operator '/' for types 'timespan' and 'float8'\n You will have to retype this query using an explicit cast \nbut:\nabr=> \\df\n...\ntimespan |timespan_div |timespan float8 |divide\n...\n\nWhy? Directly used I get:\nabr=> select timespan_div(sum(endt-begt-pausen),count(*)::float8) from tage;\ntimespan_div\n-----------------------------\n@ 11 hours 24 mins 34.79 secs\n(1 row) \n\nIt seems that an entry for '/' is defined:\ngrep timespan_div catalog/*.bki*\n...\ncatalog/local1_template1.bki.source:insert OID = 1585 ( \"/\" PGUID 0 b t f 1186\n1186 1186 0 0 0 0 timespan_div - - ) \n...\n\nOR are the three 1186's wrong? \nShouldn't at least the second input type be 701?\n\nAny clues?\n\nBye!\n----\nMichael Reifenberger\nPlaut Software GmbH, R/3 Basis\n\n\n", "msg_date": "Mon, 8 Mar 1999 09:33:05 +0100 (CET)", "msg_from": "Michael Reifenberger <[email protected]>", "msg_from_op": true, "msg_subject": "Timespan_div misbehaving?" }, { "msg_contents": "Michael Reifenberger <[email protected]> writes:\n> catalog/local1_template1.bki.source:insert OID = 1585 ( \"/\" PGUID 0 b t f 1186\n> 1186 1186 0 0 0 0 timespan_div - - ) \n\n> OR are the three 1186's wrong? \n> Shouldn't at least the second input type be 701?\n\nI think you are right --- the only timespan_div I can find in the source\ncode is TimeSpan *timespan_div(TimeSpan *span1, float8 *arg2) in adt/dt.c.\nIt looks like the entry for it in pg_proc is right, but the one in\npg_operator is wrong.\n\nHmm, a whole new class of cross-checks that opr_sanity ought to make ;-)\n\n\t\t\tregards, tom lane\n", "msg_date": "Mon, 08 Mar 1999 10:21:29 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Timespan_div misbehaving? " }, { "msg_contents": "I said:\n> It looks like the entry for it in pg_proc is right, but the one in\n> pg_operator is wrong.\n>\n> Hmm, a whole new class of cross-checks that opr_sanity ought to make ;-)\n\nIndeed, applying a mechanical cross-check was fruitful ... I found four\ndozen pg_operator entries that disagree with the corresponding pg_proc\nentries:\n\nQUERY: SELECT p1.oid, p1.oprname, p2.oid, p2.proname\nFROM pg_operator AS p1, pg_proc AS p2\nWHERE p1.oprcode = p2.oid AND\n p1.oprkind = 'b' AND\n (p2.pronargs != 2 OR\n p1.oprresult != p2.prorettype OR\n (p1.oprleft != p2.proargtypes[0] AND p2.proargtypes[0] != 0) OR\n (p1.oprright != p2.proargtypes[1] AND p2.proargtypes[1] != 0));\n oid|oprname| oid|proname \n----+-------+----+-------------\n 15|= | 852|int48eq \n 36|<> | 853|int48ne \n 37|< | 854|int48lt \n 76|> | 855|int48gt \n 80|<= | 856|int48le \n 82|>= | 857|int48ge \n 532|= | 158|int24eq \n 533|= | 159|int42eq \n 534|< | 160|int24lt \n 535|< | 161|int42lt \n 536|> | 162|int24gt \n 537|> | 163|int42gt \n 538|<> | 164|int24ne \n 539|<> | 165|int42ne \n 540|<= | 166|int24le \n 541|<= | 167|int42le \n 542|>= | 168|int24ge \n 543|>= | 169|int42ge \n 609|< | 66|int4lt \n 610|> | 147|int4gt \n 611|<= | 149|int4le \n 612|>= | 150|int4ge \n 626|!!= |1285|int4notin \n 974||| |1258|textcat \n 979||| |1258|textcat \n1055|~ |1254|textregexeq \n1056|!~ |1256|textregexne \n1063|~ |1254|textregexeq \n1064|!~ |1256|textregexne \n1211|~~ | 850|textlike \n1212|!~~ | 851|textnlike \n1213|~~ | 850|textlike \n1214|!~~ | 851|textnlike \n1232|~* |1238|texticregexeq\n1233|!~* |1239|texticregexne\n1234|~* |1238|texticregexeq\n1235|!~* |1239|texticregexne\n1522|<-> |1476|dist_pc \n1585|/ |1326|timespan_div \n 820|= | 920|network_eq \n 821|<> | 925|network_ne \n 822|< | 921|network_lt \n 823|<= | 922|network_le \n 824|> | 923|network_gt \n 825|>= | 924|network_ge \n 826|<< | 927|network_sub \n 827|<<= | 928|network_subeq\n 828|>> | 929|network_sup \n1004|>>= | 930|network_supeq\n(49 rows)\n\nSome of these are quasi-legitimate cases (like both \"text\" and \"varchar\"\nentries for one operator --- is that really necessary?). Quite a few\nseem to be real bugs. Working on fixing them now.\n\nIf I can figure out how to deal with the binary-equivalent cases\nautomatically, will commit an extension of opr_sanity regress test to\ndetect such mistakes in future.\n\n\t\t\tregards, tom lane\n", "msg_date": "Mon, 08 Mar 1999 20:55:20 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Timespan_div misbehaving? " }, { "msg_contents": "> I said:\n> > It looks like the entry for it in pg_proc is right, but the one in\n> > pg_operator is wrong.\n> >\n> > Hmm, a whole new class of cross-checks that opr_sanity ought to make ;-)\n> \n> Indeed, applying a mechanical cross-check was fruitful ... I found four\n> dozen pg_operator entries that disagree with the corresponding pg_proc\n> entries:\n\nTom, would you please take the pg/include/catalog/*.sql files and merge\nthem into the regression sanity checks? They should be in there too. \nYou can remove them once you have moved them over.\n\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 21:18:00 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Timespan_div misbehaving?" }, { "msg_contents": "Bruce Momjian <[email protected]> wrote a couple weeks ago:\n> Tom, would you please take the pg/include/catalog/*.sql files and merge\n> them into the regression sanity checks? They should be in there too. \n> You can remove them once you have moved them over.\n\nI made some progress on this last night, but still need to work on the\nother test file from include/catalog/.\n\nThere is a new regress test \"oidjoins\" that checks consistency of all\nOID and REGPROC columns in the standard tables. Basically it looks\nfor dangling links --- values that do not match any OID in the target\ntable that the column is supposed to link to. The test script can be\nrebuilt, if the set of columns to check changes, by running the code\nin contrib/findoidjoins. (I thought about just running findoidjoins\nas part of the regress test, but it seems unreasonably slow for that.\nAlso, you really want to eyeball findoidjoins' output before using it,\nsince the proggie will be misled by OIDs that join to the wrong table.)\n\nI found a dozen or so bogus tuples in several system tables (mostly\npg_amop and pg_aggregate IIRC). They had OID entries that referred\nto OIDs that actually exist ... but are in the wrong table. For\nexample some of the amopclaid fields had values that were OIDs of\nprocs or types instead of opclass rows. I believe these are leftovers\nfrom some ancient time when those OIDs were being used for something\nelse. (Most of the current assignments of those OIDs have to do with\nint8, so I know they are relatively new. Evidently no one ran an\noidjoins check while the OIDs were unused...)\n\nAll of the tuples in question were either unused or else referenced\nonly by other bogus tuples, so I just deleted 'em. (I'm new to\nadding/removing entries in system tables; is there anything else that\nneeds to be done besides editing include/catalog/pg_*.h?)\n\nNote the oidjoins test will fail until you do an initdb, because of\nthe aforementioned bogus tuples.\n\n\t\t\tregards, tom lane\n", "msg_date": "Fri, 26 Mar 1999 10:23:32 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "New regress test (was Re: Timespan_div misbehaving?)" }, { "msg_contents": "> I found a dozen or so bogus tuples in several system tables (mostly\n> pg_amop and pg_aggregate IIRC). They had OID entries that referred\n> to OIDs that actually exist ... but are in the wrong table. For\n> example some of the amopclaid fields had values that were OIDs of\n> procs or types instead of opclass rows. I believe these are leftovers\n> from some ancient time when those OIDs were being used for something\n> else. (Most of the current assignments of those OIDs have to do with\n> int8, so I know they are relatively new. Evidently no one ran an\n> oidjoins check while the OIDs were unused...)\n> \n> All of the tuples in question were either unused or else referenced\n> only by other bogus tuples, so I just deleted 'em. (I'm new to\n> adding/removing entries in system tables; is there anything else that\n> needs to be done besides editing include/catalog/pg_*.h?)\n\nGreat. Good job. No, I just edit the *.h files, and run initdb.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Fri, 26 Mar 1999 11:25:29 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] New regress test (was Re: Timespan_div misbehaving?)" } ]
[ { "msg_contents": "Hi all,\n\nThis seems like a bug...\n\nprova=> select date '1941-01-01' + interval '1 year';\n?column?\n----------------------\n1942-01-01 00:00:00+02\n(1 row)\n\nprova=> select date '1942-01-01' + interval '1 year';\n?column?\n----------------------\n1942-12-31 23:00:00+01 <---------????????????????\n(1 row)\n\nprova=> select date '1943-01-01' + interval '1 year';\n?column?\n----------------------\n1944-01-01 00:00:00+01\n(1 row)\n\nAny ideas?\n--\n- Jose' -\n\n\"No other success in life can compensate for failure in the home\" (David\nO. McKay)\n\n\n", "msg_date": "Mon, 08 Mar 1999 10:16:24 +0100", "msg_from": "\"jose' soares\" <[email protected]>", "msg_from_op": true, "msg_subject": "What's happened with 1942" }, { "msg_contents": "\"jose' soares\" <[email protected]> writes:\n> prova=> select date '1941-01-01' + interval '1 year';\n> ?column?\n> ----------------------\n> 1942-01-01 00:00:00+02\n> (1 row)\n\n> prova=> select date '1942-01-01' + interval '1 year';\n> ?column?\n> ----------------------\n> 1942-12-31 23:00:00+01 <---------????????????????\n> (1 row)\n\nNote the transition from timezone \"+02\" to \"+01\" above. You're probably\nlooking at an artifact of the year-round daylight savings time that (IIRC)\nsome nations used during the war. Or at least some kind of change in\nlocal timezone or DST rules that year.\n\nThomas would have a better idea about this than I do, but I'm betting that\nthe date-plus-interval operator thinks that a \"one year\" interval means\n\"365 (or 366 in leap year) times 24 hours\", rather than \"same local time\non the same nominal date next year\". I seem to recall discussions about\nexactly what \"date plus 1 day\" means when crossing a daylight-savings-time\nboundary, for example.\n\n\t\t\tregards, tom lane\n", "msg_date": "Mon, 08 Mar 1999 10:32:04 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] What's happened with 1942 " }, { "msg_contents": "> > prova=> select date '1942-01-01' + interval '1 year';\n> > ?column?\n> > ----------------------\n> > 1942-12-31 23:00:00+01 <---------????????????????\n> > (1 row)\n\nIt is as Tom says:\n\npostgres=> select date '1942-01-01' + interval '1 year';\n?column?\n----------------------------\nFri Jan 01 00:00:00 1943 GMT\n(1 row)\n\n$ setenv PGTZ PST8PDT\n\npostgres=> select date '1942-01-01' + interval '1 year';\n?column?\n----------------------------\nFri Jan 01 01:00:00 1943 PWT\n(1 row)\n\nWe seem to have had a timezone adjustment for that year here too. But it\nis in your timezone database, not inside of Postgres. Really...\n\n> Thomas would have a better idea about this than I do, but I'm betting \n> that the date-plus-interval operator thinks that a \"one year\" interval \n> means \"365 (or 366 in leap year) times 24 hours\", rather than \"same \n> local time on the same nominal date next year\". I seem to recall \n> discussions about exactly what \"date plus 1 day\" means when crossing a \n> daylight-savings-time boundary, for example.\n\nActually, I'm pretty sure that it adds one to the year in a tm\nstructure. But Jose's local time zone changed across that year, and the\ncalculations are done in GMT, then \"rotated\" back to local time, which\nreflects that new zone definition.\n\nYou can use zdump to look at the time zone definitions. What time zone\nare you using Jose'?\n\n - Tom\n", "msg_date": "Tue, 09 Mar 1999 03:45:57 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] What's happened with 1942" } ]
[ { "msg_contents": ">> When will updateable views be supported?\n> Since v6.4 :-)\n\nWhat about VIEWS with UNIONS? I hope someone will write it for 6.5...\n\n", "msg_date": "Mon, 08 Mar 1999 10:29:24 +0100", "msg_from": "Kaare Rasmussen <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Updatable views?" } ]
[ { "msg_contents": "On Mon, 8 Mar 1999, Vadim Valiakhmetov wrote:\n\n> Subject: [HACKERS] Postgress case insensitive search on FreeBSD3.0\n\n> I have encountered strange behaviour of Postgresql-6.4.2 server\n> (installed on FreeBSD3.0 from distributive) when I was trying to do\n> case-insensitive search in simplest text table (without indexes)\n\nTry to compile with unsigend chars, if the problem persists then you have\nthe same problem that I witnessed on BSDi.\n\nYou will need to manually edit regexp.c and cast all chars to unsigned chars\nin all the pg_upper and pg_lower function calls.\n\n\n\n\n-- \nIncredible Networks LTD Angelos Karageorgiou\n20 Karea st, +30.1.92.12.312 (voice)\n116 36 Athens, Greece. +30.1.92.12.314 (fax)\nhttp://www.incredible.com [email protected] (e-mail)\n\n", "msg_date": "Mon, 8 Mar 1999 15:22:35 +0200 (EET)", "msg_from": "Angelos Karageorgiou <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Postgress case insensitive search on FreeBSD3.0" }, { "msg_contents": "Hello!\n\nI have encountered strange behaviour of Postgresql-6.4.2 server\n(installed on FreeBSD3.0 from distributive) when I was trying to do\ncase-insensitive search in simplest text table (without indexes)\ncontaining sequence of strings of characters from national alphabet\n(Russian KOI8-R and CP1251 character sets). The following example query\ndoes not do case insensitive search for both KOI8-R and CP1251:\n\nselect firm from plan where firm ~* ('substring') order by firm asc\n\nHere 'substring' consists of russian characters, of course.\n\nI believe, that my russian locale properly configured and working, as\nwell as postmaster compiled with option --enable-locale. Environment\nvariables $LC_CTYPE and $LC_COLLATE also defined. For instance,\noperators upper('substring') and lower ('substring') both working\ncorrectly with russian characters.\n\nJust the same example query does case insensitive search under RedHat\nLinux 4.2 with glibc2 installed. All postmaster settings were the same\nas for FreeBSD.\n\nCan anybody clarify this situation ?\n\nThanks in advance, and sorry for bad english.\n\nVadim Valiakhmetov mailto:[email protected]\nSysadmin, SibLine Co.\n\n\n\n", "msg_date": "Mon, 08 Mar 1999 19:22:52 +0600", "msg_from": "Vadim Valiakhmetov <[email protected]>", "msg_from_op": false, "msg_subject": "Postgress case insensitive search on FreeBSD3.0" } ]
[ { "msg_contents": "The \"Integrated Document\" on the web page has the following paragraph in\nthe section \"CREATE TABLE\" in chapter 20.\n\nIn the current release (v6.4), Postgres evaluates all default expressions at\nthe time the table is defined. Hence, functions which are \"non-cacheable\"\nsuch as CURRENT_TIMESTAMP may not produce the desired effect. For the\nparticular case of date/time types, one can work around this behavior by\nusing \"DEFAULT TEXT 'now'\" instead of \"DEFAULT 'now'\" or \"DEFAULT\nCURRENT_TIMESTAMP\". This forces Postgres to consider the constant a string\ntype and then to convert the value to timestamp at runtime.\n\nThis appears to be untrue. Is this a change since 6.4 or is there\nsome cases where using CURRENT_TIMESTAMP will not do the expected thing?\n\nAlso, the title of the document (The PostgreSQL Development Team) seems\nto be incorrect.\n\n-- \nD'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\nhttp://www.druid.net/darcy/ | and a sheep voting on\n+1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.\n", "msg_date": "Mon, 8 Mar 1999 11:50:24 -0500 (EST)", "msg_from": "\"D'Arcy\" \"J.M.\" Cain <[email protected]>", "msg_from_op": true, "msg_subject": "Niladic functions" }, { "msg_contents": "> The \"Integrated Document\" on the web page has the following paragraph \n> in the section \"CREATE TABLE\" in chapter 20.\n> In the current release (v6.4), Postgres evaluates all default \n> expressions at the time the table is defined. Hence, functions which \n> are \"non-cacheable\" such as CURRENT_TIMESTAMP may not produce the \n> desired effect. For the particular case of date/time types, one can \n> work around this behavior by using \"DEFAULT TEXT 'now'\" instead of \n> \"DEFAULT 'now'\" or \"DEFAULT CURRENT_TIMESTAMP\". This forces Postgres \n> to consider the constant a string type and then to convert the value \n> to timestamp at runtime.\n> \n> This appears to be untrue. Is this a change since 6.4 or is there\n> some cases where using CURRENT_TIMESTAMP will not do the expected \n> thing?\n\nSorry for being slow, but exactly which of the several assertions in the\nabove paragraph are not true? Just the DEFAULT CURRENT_TIMESTAMP\nbehavior?\n\n> Also, the title of the document (The PostgreSQL Development Team) \n> seems to be incorrect.\n\nMy browser shows that as the second line, just under the title\n\"PostgreSQL\", as I intended. What browser are you using?\n\n - Tom\n", "msg_date": "Tue, 09 Mar 1999 03:58:50 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Niladic functions" }, { "msg_contents": "Thus spake Thomas G. Lockhart\n> > The \"Integrated Document\" on the web page has the following paragraph \n> > in the section \"CREATE TABLE\" in chapter 20.\n> > In the current release (v6.4), Postgres evaluates all default \n> > expressions at the time the table is defined. Hence, functions which \n> > are \"non-cacheable\" such as CURRENT_TIMESTAMP may not produce the \n> > desired effect. For the particular case of date/time types, one can \n> > work around this behavior by using \"DEFAULT TEXT 'now'\" instead of \n> > \"DEFAULT 'now'\" or \"DEFAULT CURRENT_TIMESTAMP\". This forces Postgres \n> > to consider the constant a string type and then to convert the value \n> > to timestamp at runtime.\n> > \n> > This appears to be untrue. Is this a change since 6.4 or is there\n> > some cases where using CURRENT_TIMESTAMP will not do the expected \n> > thing?\n> \n> Sorry for being slow, but exactly which of the several assertions in the\n> above paragraph are not true? Just the DEFAULT CURRENT_TIMESTAMP\n> behavior?\n\nYes. This paragraph says that CURRENT_TIMESTAMP is evaluated at the time\nthe table is defined yet I tested this with a current version and it\nactually gets evaluated at INSERT time.\n\n> > Also, the title of the document (The PostgreSQL Development Team) \n> > seems to be incorrect.\n> \n> My browser shows that as the second line, just under the title\n> \"PostgreSQL\", as I intended. What browser are you using?\n\nSorry, my misunderstanding. I didn't realize that PostgreSQL was the\ntitle of this particular document. Anyway, the link to this page is\n\"Integrated Document\" so there is a small inconsistency.\n\n-- \nD'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\nhttp://www.druid.net/darcy/ | and a sheep voting on\n+1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.\n", "msg_date": "Tue, 9 Mar 1999 10:42:50 -0500 (EST)", "msg_from": "\"D'Arcy\" \"J.M.\" Cain <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Niladic functions" }, { "msg_contents": "> > > This appears to be untrue. Is this a change since 6.4 or is there\n> > > some cases where using CURRENT_TIMESTAMP will not do the expected\n> > > thing?\n> This paragraph says that CURRENT_TIMESTAMP is evaluated at the time\n> the table is defined yet I tested this with a current version and it\n> actually gets evaluated at INSERT time.\n\nYou are right; I had changed the parser to force a string into the parse\ntree rather than a constant. Would you be interested in sending patches\nfor the docs? :)\n\n> > > Also, the title of the document (The PostgreSQL Development Team)\n> > > seems to be incorrect.\n> > My browser shows that as the second line, just under the title\n> > \"PostgreSQL\", as I intended. What browser are you using?\n> Sorry, my misunderstanding. I didn't realize that PostgreSQL was the\n> title of this particular document. Anyway, the link to this page is\n> \"Integrated Document\" so there is a small inconsistency.\n\nThe problem is that this is a document integrated from all of the other\ndocuments for purposes of HTML formatting and browsing. I wanted to make\nclear that all of the information appears somewhere in the other\ndocuments.\n\n - Tom\n", "msg_date": "Tue, 09 Mar 1999 17:37:55 +0000", "msg_from": "\"Thomas G. Lockhart\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Niladic functions" }, { "msg_contents": "Thus spake Thomas G. Lockhart\n> > This paragraph says that CURRENT_TIMESTAMP is evaluated at the time\n> > the table is defined yet I tested this with a current version and it\n> > actually gets evaluated at INSERT time.\n> \n> You are right; I had changed the parser to force a string into the parse\n> tree rather than a constant. Would you be interested in sending patches\n> for the docs? :)\n\nI can't find the source doc in the distribution but it's pretty simple.\nJust remove that whole paragraph.\n\n> > Sorry, my misunderstanding. I didn't realize that PostgreSQL was the\n> > title of this particular document. Anyway, the link to this page is\n> > \"Integrated Document\" so there is a small inconsistency.\n> \n> The problem is that this is a document integrated from all of the other\n> documents for purposes of HTML formatting and browsing. I wanted to make\n> clear that all of the information appears somewhere in the other\n> documents.\n\nHence the title on the referring page. How about changing the title on\nthe page itself to reflect that as well? Something like;\n\n \"PostgreSQL Integrated Development Document\"\n\nor some such. No biggie but it makes it easier to remember how you got\nto the document you are reading.\n\n-- \nD'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\nhttp://www.druid.net/darcy/ | and a sheep voting on\n+1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.\n", "msg_date": "Tue, 9 Mar 1999 14:00:59 -0500 (EST)", "msg_from": "\"D'Arcy\" \"J.M.\" Cain <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Niladic functions" }, { "msg_contents": "\nCan someone comment on this?\n\n\n> The \"Integrated Document\" on the web page has the following paragraph in\n> the section \"CREATE TABLE\" in chapter 20.\n> \n> In the current release (v6.4), Postgres evaluates all default expressions at\n> the time the table is defined. Hence, functions which are \"non-cacheable\"\n> such as CURRENT_TIMESTAMP may not produce the desired effect. For the\n> particular case of date/time types, one can work around this behavior by\n> using \"DEFAULT TEXT 'now'\" instead of \"DEFAULT 'now'\" or \"DEFAULT\n> CURRENT_TIMESTAMP\". This forces Postgres to consider the constant a string\n> type and then to convert the value to timestamp at runtime.\n> \n> This appears to be untrue. Is this a change since 6.4 or is there\n> some cases where using CURRENT_TIMESTAMP will not do the expected thing?\n> \n> Also, the title of the document (The PostgreSQL Development Team) seems\n> to be incorrect.\n> \n> -- \n> D'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\n> http://www.druid.net/darcy/ | and a sheep voting on\n> +1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.\n> \n> \n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sun, 9 May 1999 12:03:57 -0400 (EDT)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Niladic functions" }, { "msg_contents": "> Can someone comment on this?\n\nResolved.\n\n - Tom\n\n> > The \"Integrated Document\" on the web page has the following paragraph in\n> > the section \"CREATE TABLE\" in chapter 20.\n> >\n> > In the current release (v6.4), Postgres evaluates all default \n> > expressions...\n\n-- \nThomas Lockhart\t\t\t\[email protected]\nSouth Pasadena, California\n", "msg_date": "Sun, 09 May 1999 16:53:55 +0000", "msg_from": "Thomas Lockhart <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Niladic functions" }, { "msg_contents": "\nWas this addressed?\n\n\n> The \"Integrated Document\" on the web page has the following paragraph in\n> the section \"CREATE TABLE\" in chapter 20.\n> \n> In the current release (v6.4), Postgres evaluates all default expressions at\n> the time the table is defined. Hence, functions which are \"non-cacheable\"\n> such as CURRENT_TIMESTAMP may not produce the desired effect. For the\n> particular case of date/time types, one can work around this behavior by\n> using \"DEFAULT TEXT 'now'\" instead of \"DEFAULT 'now'\" or \"DEFAULT\n> CURRENT_TIMESTAMP\". This forces Postgres to consider the constant a string\n> type and then to convert the value to timestamp at runtime.\n> \n> This appears to be untrue. Is this a change since 6.4 or is there\n> some cases where using CURRENT_TIMESTAMP will not do the expected thing?\n> \n> Also, the title of the document (The PostgreSQL Development Team) seems\n> to be incorrect.\n> \n> -- \n> D'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\n> http://www.druid.net/darcy/ | and a sheep voting on\n> +1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.\n> \n> \n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 6 Jul 1999 22:19:34 -0400 (EDT)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Niladic functions" }, { "msg_contents": "> Was this addressed?\n\nYes.\n\n - Thomas\n\n> > The \"Integrated Document\" on the web page has the following paragraph in\n> > the section \"CREATE TABLE\" in chapter 20.\n> >\n> > In the current release (v6.4), Postgres evaluates all default expressions at\n> > the time the table is defined. Hence, functions which are \"non-cacheable\"\n> > such as CURRENT_TIMESTAMP may not produce the desired effect. For the\n> > particular case of date/time types, one can work around this behavior by\n> > using \"DEFAULT TEXT 'now'\" instead of \"DEFAULT 'now'\" or \"DEFAULT\n> > CURRENT_TIMESTAMP\". This forces Postgres to consider the constant a string\n> > type and then to convert the value to timestamp at runtime.\n> >\n> > This appears to be untrue. Is this a change since 6.4 or is there\n> > some cases where using CURRENT_TIMESTAMP will not do the expected thing?\n> >\n> > Also, the title of the document (The PostgreSQL Development Team) seems\n> > to be incorrect.\n\n\n-- \nThomas Lockhart\t\t\t\[email protected]\nSouth Pasadena, California\n", "msg_date": "Wed, 14 Jul 1999 05:30:52 +0000", "msg_from": "Thomas Lockhart <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Niladic functions" } ]
[ { "msg_contents": "OK, I have installed a rotating globe on the developers page. After I\nget all the developer locations, I will add red dots to the globe.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 12:34:27 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Developers globe" }, { "msg_contents": "> OK, I have installed a rotating globe on the developers page. After I\n> get all the developer locations, I will add red dots to the globe.\n\nAs an example, I have added Toronto to the globe, which I believe is the\nlocation of our server, and where D'Arcy lives.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 12:53:36 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Mon, 8 Mar 1999, Bruce Momjian wrote:\n\n> > OK, I have installed a rotating globe on the developers page. After I\n> > get all the developer locations, I will add red dots to the globe.\n> \n> As an example, I have added Toronto to the globe, which I believe is the\n> location of our server, and where D'Arcy lives.\n\nWould a 'flat-map' be more friendly? Just looking at it right now, and\nI'm on a T1 at work, and its \"stalled\"...and altho I can see the red dot\naround Toronto, its \"Faint\"...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 8 Mar 1999 14:49:42 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> On Mon, 8 Mar 1999, Bruce Momjian wrote:\n> \n> > > OK, I have installed a rotating globe on the developers page. After I\n> > > get all the developer locations, I will add red dots to the globe.\n> > \n> > As an example, I have added Toronto to the globe, which I believe is the\n> > location of our server, and where D'Arcy lives.\n> \n> Would a 'flat-map' be more friendly? Just looking at it right now, and\n> I'm on a T1 at work, and its \"stalled\"...and altho I can see the red dot\n> around Toronto, its \"Faint\"...\n\nThat may be better. I can make it larger, or can someone send me a flat\nearth map to use.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 13:57:40 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "Wht not collect the lon and lat for xearth? Then you can stick the\ndevelopers name around the red circle. \n\n-Egon\n\nOn Mon, 8 Mar 1999, The Hermit Hacker wrote:\n\n> On Mon, 8 Mar 1999, Bruce Momjian wrote:\n> \n> > > OK, I have installed a rotating globe on the developers page. After I\n> > > get all the developer locations, I will add red dots to the globe.\n> > \n> > As an example, I have added Toronto to the globe, which I believe is the\n> > location of our server, and where D'Arcy lives.\n> \n> Would a 'flat-map' be more friendly? Just looking at it right now, and\n> I'm on a T1 at work, and its \"stalled\"...and altho I can see the red dot\n> around Toronto, its \"Faint\"...\n> \n> Marc G. Fournier \n> Systems Administrator @ hub.org \n> primary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n> \n> \n> \n\n", "msg_date": "Mon, 8 Mar 1999 20:10:43 +0100 (MET)", "msg_from": "Egon Schmid <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Mon, 8 Mar 1999, Egon Schmid wrote:\n\n> Wht not collect the lon and lat for xearth? Then you can stick the\n> developers name around the red circle. \n\nWant to build up a sample of what you are proposing and let us see it?\n\n> \n> -Egon\n> \n> On Mon, 8 Mar 1999, The Hermit Hacker wrote:\n> \n> > On Mon, 8 Mar 1999, Bruce Momjian wrote:\n> > \n> > > > OK, I have installed a rotating globe on the developers page. After I\n> > > > get all the developer locations, I will add red dots to the globe.\n> > > \n> > > As an example, I have added Toronto to the globe, which I believe is the\n> > > location of our server, and where D'Arcy lives.\n> > \n> > Would a 'flat-map' be more friendly? Just looking at it right now, and\n> > I'm on a T1 at work, and its \"stalled\"...and altho I can see the red dot\n> > around Toronto, its \"Faint\"...\n> > \n> > Marc G. Fournier \n> > Systems Administrator @ hub.org \n> > primary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n> > \n> > \n> > \n> \n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 8 Mar 1999 15:40:32 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "\nOn 08-Mar-99 The Hermit Hacker wrote:\n> On Mon, 8 Mar 1999, Egon Schmid wrote:\n> \n>> Wht not collect the lon and lat for xearth? Then you can stick the\n>> developers name around the red circle. \n> \n> Want to build up a sample of what you are proposing and let us see it?\n\nLook at the national directory on camping-usa. The map was drawn with\nlon-lat coordinates for the US and each state and the red dots were put\nin place with lon-lat coords. (The clickable map translates the XY to\nlon-lat) I don't know how you'd put names in it, tho, without making\nthe image rather large.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Mon, 08 Mar 1999 14:44:51 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> On Mon, 8 Mar 1999, Egon Schmid wrote:\n> \n> > Wht not collect the lon and lat for xearth? Then you can stick the\n> > developers name around the red circle. \n> \n> Want to build up a sample of what you are proposing and let us see it?\n\nYes, I can do that with xearth. Very easy. But xearth produces a very\nlarge image, not really proper for a browser, unless we forget the\nrotation, and just grab western and eastern hemisphere images and put\nthe names on there.\n\nWith names, that is going to be a large image. Do we want that?\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 14:45:37 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "I have seen xearth on my Debian system with the most Debian maintainers\nspread around the globe. I think there is a flag for project so no htere\nis no need to put western and eastern hemispheres together. xearth is\ntrimmed to run in a X Windows background.\n\n-Egon\n\nOn Mon, 8 Mar 1999, Bruce Momjian wrote:\n\n> > On Mon, 8 Mar 1999, Egon Schmid wrote:\n> > \n> > > Wht not collect the lon and lat for xearth? Then you can stick the\n> > > developers name around the red circle. \n> > \n> > Want to build up a sample of what you are proposing and let us see it?\n> \n> Yes, I can do that with xearth. Very easy. But xearth produces a very\n> large image, not really proper for a browser, unless we forget the\n> rotation, and just grab western and eastern hemisphere images and put\n> the names on there.\n> \n> With names, that is going to be a large image. Do we want that?\n> \n> -- \n> Bruce Momjian | http://www.op.net/~candle\n> [email protected] | (610) 853-3000\n> + If your life is a hard drive, | 830 Blythe Avenue\n> + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n> \n\n", "msg_date": "Mon, 8 Mar 1999 20:57:28 +0100 (MET)", "msg_from": "Egon Schmid <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "\nYou don't really need nams associated iwth the dots, just make the dots\nreasonably accurate...put a long/lat beside each name in the list, and\ncountry or whatnot...\n\n On Mon, 8 Mar 1999, Vince Vielhaber wrote:\n\n> \n> On 08-Mar-99 The Hermit Hacker wrote:\n> > On Mon, 8 Mar 1999, Egon Schmid wrote:\n> > \n> >> Wht not collect the lon and lat for xearth? Then you can stick the\n> >> developers name around the red circle. \n> > \n> > Want to build up a sample of what you are proposing and let us see it?\n> \n> Look at the national directory on camping-usa. The map was drawn with\n> lon-lat coordinates for the US and each state and the red dots were put\n> in place with lon-lat coords. (The clickable map translates the XY to\n> lon-lat) I don't know how you'd put names in it, tho, without making\n> the image rather large.\n> \n> Vince.\n> -- \n> ==========================================================================\n> Vince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n> # include <std/disclaimers.h> TEAM-OS2\n> Online Campground Directory http://www.camping-usa.com\n> Online Giftshop Superstore http://www.cloudninegifts.com\n> ==========================================================================\n> \n> \n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 8 Mar 1999 16:06:10 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "\nOn 08-Mar-99 The Hermit Hacker wrote:\n> \n> You don't really need nams associated iwth the dots, just make the dots\n> reasonably accurate...put a long/lat beside each name in the list, and\n> country or whatnot...\n\n'cept xearth isn't using lon/lat - unless it's changed from the v1.0 source\nI have here (from the 2.2.8 ports collection CD).\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Mon, 08 Mar 1999 15:16:38 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "Thus spake Bruce Momjian\n> > OK, I have installed a rotating globe on the developers page. After I\n> > get all the developer locations, I will add red dots to the globe.\n> \n> As an example, I have added Toronto to the globe, which I believe is the\n> location of our server, and where D'Arcy lives.\n\nHmmm. I tried to go there but for some reason I keep getting bounced\nto Saskatchewan which I assume doesn't have the updates yet. Any idea\nwhy I don't go to the main site? After all, I'm on the same LAN basically.\nIt seems silly to go out to the net to a mirror.\n\n-- \nD'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\nhttp://www.druid.net/darcy/ | and a sheep voting on\n+1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.\n", "msg_date": "Mon, 8 Mar 1999 16:08:39 -0500 (EST)", "msg_from": "\"D'Arcy\" \"J.M.\" Cain <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "\"D'Arcy\" \"J.M.\" Cain <[email protected]> writes:\n> Hmmm. I tried to go there but for some reason I keep getting bounced\n> to Saskatchewan which I assume doesn't have the updates yet. Any idea\n> why I don't go to the main site? After all, I'm on the same LAN basically.\n> It seems silly to go out to the net to a mirror.\n\nI just figured this out yesterday:\n\nhttp://www.postgresql.org/ ---> redirection 'bot\nhttp://www.postgresql.org/index.html ---> master site\n\nThe 'bot is good for load sharing, but when you want to see the latest\nupdates it's a real pain. I'm glad it's possible to bypass it.\n\n\t\t\tregards, tom lane\n", "msg_date": "Mon, 08 Mar 1999 17:43:43 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe " }, { "msg_contents": "> Thus spake Bruce Momjian\n> > > OK, I have installed a rotating globe on the developers page. After I\n> > > get all the developer locations, I will add red dots to the globe.\n> > \n> > As an example, I have added Toronto to the globe, which I believe is the\n> > location of our server, and where D'Arcy lives.\n> \n> Hmmm. I tried to go there but for some reason I keep getting bounced\n> to Saskatchewan which I assume doesn't have the updates yet. Any idea\n> why I don't go to the main site? After all, I'm on the same LAN basically.\n> It seems silly to go out to the net to a mirror.\n\nYes. If you do:\n\n\thttp://www.postgresql.org/index.html\n\nyou are forced to the main site. Marc told me this a while ago.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 18:10:16 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Mon, 8 Mar 1999, D'Arcy J.M. Cain wrote:\n\n> Thus spake Bruce Momjian\n> > > OK, I have installed a rotating globe on the developers page. After I\n> > > get all the developer locations, I will add red dots to the globe.\n> > \n> > As an example, I have added Toronto to the globe, which I believe is the\n> > location of our server, and where D'Arcy lives.\n> \n> Hmmm. I tried to go there but for some reason I keep getting bounced\n> to Saskatchewan which I assume doesn't have the updates yet. Any idea\n> why I don't go to the main site? After all, I'm on the same LAN basically.\n> It seems silly to go out to the net to a mirror.\n\nJust the way the redirect software is setup...you can cheat it though:\n\n\thttp://www.postgresql.org/index.html\n\n:)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 8 Mar 1999 19:25:54 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> OK, I have installed a rotating globe on the developers page. After I\n> get all the developer locations, I will add red dots to the globe.\n\nOK, larger dots. I added a dot for myself and Tom Lane. Europe will be\nall red once I am done. Things are very imprecise at that size.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 18:37:08 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> OK, I have installed a rotating globe on the developers page. After I\n> get all the developer locations, I will add red dots to the globe.\n\nOK, now the dot is huge. Marc, Tom Lane, and I fit in the same dot. \nLooks good, though.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 22:21:25 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> > OK, I have installed a rotating globe on the developers page. After I\n> > get all the developer locations, I will add red dots to the globe.\n> \n> OK, now the dot is huge. Marc, Tom Lane, and I fit in the same dot. \n> Looks good, though.\n> \n\nNo, Toronto/D'Arch, Tom Lane, and I are in the same dot. Marc gets his\non more east than us.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 22:23:58 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> OK, I have installed a rotating globe on the developers page. After I\n> get all the developer locations, I will add red dots to the globe.\n\nOK, I have big dots for almost everyone. How does it look. The moving\nof the dots is because I was sloppy, but I just wanted to do it quick to\nsee how it looks.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 8 Mar 1999 22:48:37 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "Marc G. Fournier wrote:\n\n> You don't really need nams associated iwth the dots, just make the dots\n> reasonably accurate...put a long/lat beside each name in the list, and\n> country or whatnot...\n\n Alternatively you might put a map behind it and use some\n Jscript to write the names into the status line when the\n pointer is over the dots.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Tue, 9 Mar 1999 10:23:44 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Tue, 9 Mar 1999, Jan Wieck wrote:\n\n> Marc G. Fournier wrote:\n> \n> > You don't really need nams associated iwth the dots, just make the dots\n> > reasonably accurate...put a long/lat beside each name in the list, and\n> > country or whatnot...\n> \n> Alternatively you might put a map behind it and use some\n> Jscript to write the names into the status line when the\n> pointer is over the dots.\n\nAs long as we're thinking JScript, why not pop up the name on the \nmouseover? Maybe a little flag popping up on the map.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Tue, 9 Mar 1999 06:58:37 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Tue, 9 Mar 1999, Jan Wieck wrote:\n\n> Marc G. Fournier wrote:\n> \n> > You don't really need nams associated iwth the dots, just make the dots\n> > reasonably accurate...put a long/lat beside each name in the list, and\n> > country or whatnot...\n> \n> Alternatively you might put a map behind it and use some\n> Jscript to write the names into the status line when the\n> pointer is over the dots.\n\nOh, I like that one...anyone know Javascript? :)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Tue, 9 Mar 1999 09:32:58 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Tue, 9 Mar 1999, Vince Vielhaber wrote:\n\n> On Tue, 9 Mar 1999, Jan Wieck wrote:\n> \n> > Marc G. Fournier wrote:\n> > \n> > > You don't really need nams associated iwth the dots, just make the dots\n> > > reasonably accurate...put a long/lat beside each name in the list, and\n> > > country or whatnot...\n> > \n> > Alternatively you might put a map behind it and use some\n> > Jscript to write the names into the status line when the\n> > pointer is over the dots.\n> \n> As long as we're thinking JScript, why not pop up the name on the \n> mouseover? Maybe a little flag popping up on the map.\n\nHrmmmm...isn't that what he said? *grin*\n\nHow abuot this one...a world map that when you go over a country, it pops\nup an expanded view of that country so that the dots aren't so\nconcentrated? :)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Tue, 9 Mar 1999 09:34:13 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Tue, 9 Mar 1999, The Hermit Hacker wrote:\n\n> On Tue, 9 Mar 1999, Vince Vielhaber wrote:\n> \n> > On Tue, 9 Mar 1999, Jan Wieck wrote:\n> > \n> > > Marc G. Fournier wrote:\n> > > \n> > > > You don't really need nams associated iwth the dots, just make the dots\n> > > > reasonably accurate...put a long/lat beside each name in the list, and\n> > > > country or whatnot...\n> > > \n> > > Alternatively you might put a map behind it and use some\n> > > Jscript to write the names into the status line when the\n> > > pointer is over the dots.\n> > \n> > As long as we're thinking JScript, why not pop up the name on the \n> > mouseover? Maybe a little flag popping up on the map.\n> \n> Hrmmmm...isn't that what he said? *grin*\n\nNope. Status line is at the bottom of the browser. I meant for the name\nto pop up (or a flag to pop up) in the actual image.\n\n> \n> How abuot this one...a world map that when you go over a country, it pops\n> up an expanded view of that country so that the dots aren't so\n> concentrated? :)\n\nHmmmmm. *wheels turning*\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Tue, 9 Mar 1999 10:50:51 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Tue, 9 Mar 1999, The Hermit Hacker wrote:\n\n> On Tue, 9 Mar 1999, Jan Wieck wrote:\n> \n> > Marc G. Fournier wrote:\n> > \n> > > You don't really need nams associated iwth the dots, just make the dots\n> > > reasonably accurate...put a long/lat beside each name in the list, and\n> > > country or whatnot...\n> > \n> > Alternatively you might put a map behind it and use some\n> > Jscript to write the names into the status line when the\n> > pointer is over the dots.\n> \n> Oh, I like that one...anyone know Javascript? :)\n\nyep. Anyone know of a tool in unix that splits up a gif into many\nsmaller squares (as opposed to manually cropping)?\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Tue, 9 Mar 1999 11:11:46 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> On Tue, 9 Mar 1999, Jan Wieck wrote:\n> \n> > Marc G. Fournier wrote:\n> > \n> > > You don't really need nams associated iwth the dots, just make the dots\n> > > reasonably accurate...put a long/lat beside each name in the list, and\n> > > country or whatnot...\n> > \n> > Alternatively you might put a map behind it and use some\n> > Jscript to write the names into the status line when the\n> > pointer is over the dots.\n> \n> Oh, I like that one...anyone know Javascript? :)\n\nNo.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 9 Mar 1999 12:38:45 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> OK, I have installed a rotating globe on the developers page. After I\n> get all the developer locations, I will add red dots to the globe.\n\nNew map. Gimp Nova dots. You like?\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 00:25:21 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> > Hrmmmm...isn't that what he said? *grin*\n> \n> Nope. Status line is at the bottom of the browser. I meant for the name\n> to pop up (or a flag to pop up) in the actual image.\n\n Right - and popping up in the image is better :-)\n\n\nJan\n\n-- \n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n", "msg_date": "Wed, 10 Mar 1999 10:07:46 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> \n> On Tue, 9 Mar 1999, The Hermit Hacker wrote:\n> \n> > On Tue, 9 Mar 1999, Jan Wieck wrote:\n> > \n> > > Marc G. Fournier wrote:\n> > > \n> > > > You don't really need nams associated iwth the dots, just make the dots\n> > > > reasonably accurate...put a long/lat beside each name in the list, and\n> > > > country or whatnot...\n> > > \n> > > Alternatively you might put a map behind it and use some\n> > > Jscript to write the names into the status line when the\n> > > pointer is over the dots.\n> > \n> > Oh, I like that one...anyone know Javascript? :)\n> \n> yep. Anyone know of a tool in unix that splits up a gif into many\n> smaller squares (as opposed to manually cropping)?\n\n A little script with pbm tools should do.\n\n\nJan\n\n-- \n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n", "msg_date": "Wed, 10 Mar 1999 10:08:54 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": ">\n> > On Tue, 9 Mar 1999, Jan Wieck wrote:\n> >\n> > > Marc G. Fournier wrote:\n> > >\n> > > > You don't really need nams associated iwth the dots, just make the dots\n> > > > reasonably accurate...put a long/lat beside each name in the list, and\n> > > > country or whatnot...\n> > >\n> > > Alternatively you might put a map behind it and use some\n> > > Jscript to write the names into the status line when the\n> > > pointer is over the dots.\n> >\n> > Oh, I like that one...anyone know Javascript? :)\n>\n> No.\n\n Dunno if that's possible (popping something up in an image)\n with Javascript. Maybe that must be done in an applet. Or\n what Venice suggested, splitting up the image into hundred's\n of peaces, putting the names into the ALT attributes and let\n the browser do it. For sake not every browser show's the ALT\n text when you hold the pointer over an image.\n\n Maybe I should do some tests.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 10 Mar 1999 10:13:17 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Wed, 10 Mar 1999, Bruce Momjian wrote:\n\n> > OK, I have installed a rotating globe on the developers page. After I\n> > get all the developer locations, I will add red dots to the globe.\n> \n> New map. Gimp Nova dots. You like?\n> \n\nLooks great!!\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Wed, 10 Mar 1999 06:52:06 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "\nDefinitely doable...see 'http://www.nova-scotia.org', a site we played\nwith that used javascript...\n\nOn Wed, 10 Mar 1999, Jan Wieck wrote:\n\n> >\n> > > On Tue, 9 Mar 1999, Jan Wieck wrote:\n> > >\n> > > > Marc G. Fournier wrote:\n> > > >\n> > > > > You don't really need nams associated iwth the dots, just make the dots\n> > > > > reasonably accurate...put a long/lat beside each name in the list, and\n> > > > > country or whatnot...\n> > > >\n> > > > Alternatively you might put a map behind it and use some\n> > > > Jscript to write the names into the status line when the\n> > > > pointer is over the dots.\n> > >\n> > > Oh, I like that one...anyone know Javascript? :)\n> >\n> > No.\n> \n> Dunno if that's possible (popping something up in an image)\n> with Javascript. Maybe that must be done in an applet. Or\n> what Venice suggested, splitting up the image into hundred's\n> of peaces, putting the names into the ALT attributes and let\n> the browser do it. For sake not every browser show's the ALT\n> text when you hold the pointer over an image.\n> \n> Maybe I should do some tests.\n> \n> \n> Jan\n> \n> --\n> \n> #======================================================================#\n> # It's easier to get forgiveness for being wrong than for being right. #\n> # Let's break this rule - forgive me. #\n> #======================================== [email protected] (Jan Wieck) #\n> \n> \n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 10 Mar 1999 10:23:06 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Wed, 10 Mar 1999, Vince Vielhaber wrote:\n\n> On Wed, 10 Mar 1999, Bruce Momjian wrote:\n> \n> > > OK, I have installed a rotating globe on the developers page. After I\n> > > get all the developer locations, I will add red dots to the globe.\n> > \n> > New map. Gimp Nova dots. You like?\n> > \n> \n> Looks great!!\n\nHate to be the wet blanket...throw another one of those 'Nova dots' onto\nEurope, and Europe will sink :( It sort of reminds me of those old movies\ndealing with Nuclear WAr...\"New Brunswick, Canada...totally wiped' out,\nMaine, USA...like it never existed, England...WWI and WWII failed, where\nPostgreSQL succeseds\" *grin*\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 10 Mar 1999 10:26:07 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Wed, 10 Mar 1999, The Hermit Hacker wrote:\n\n> \n> Definitely doable...see 'http://www.nova-scotia.org', a site we played\n> with that used javascript...\n\nAnd I just found a nifty little js library for doing popups. Should have\na simple demo available shortly of index.html's frames.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Wed, 10 Mar 1999 09:51:32 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> On Wed, 10 Mar 1999, Vince Vielhaber wrote:\n> \n> > On Wed, 10 Mar 1999, Bruce Momjian wrote:\n> > \n> > > > OK, I have installed a rotating globe on the developers page. After I\n> > > > get all the developer locations, I will add red dots to the globe.\n> > > \n> > > New map. Gimp Nova dots. You like?\n> > > \n> > \n> > Looks great!!\n> \n> Hate to be the wet blanket...throw another one of those 'Nova dots' onto\n> Europe, and Europe will sink :( It sort of reminds me of those old movies\n> dealing with Nuclear WAr...\"New Brunswick, Canada...totally wiped' out,\n> Maine, USA...like it never existed, England...WWI and WWII failed, where\n> PostgreSQL succeseds\" *grin*\n\nOK, smaller stars, and fewer. Should I continue in this direction, or\ntry something else?\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 10:02:16 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "Thus spake Bruce Momjian\n> > OK, I have installed a rotating globe on the developers page. After I\n> > get all the developer locations, I will add red dots to the globe.\n> \n> New map. Gimp Nova dots. You like?\n\nNice. Now we just need to fill it in. I know you have mine but just to\nconfirm:\n\n 43 27 02 N / 79 20 44 W\n\n-- \nD'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\nhttp://www.druid.net/darcy/ | and a sheep voting on\n+1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.\n", "msg_date": "Wed, 10 Mar 1999 10:04:40 -0500 (EST)", "msg_from": "\"D'Arcy\" \"J.M.\" Cain <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Wed, 10 Mar 1999, Bruce Momjian wrote:\n\n> > On Wed, 10 Mar 1999, Vince Vielhaber wrote:\n> > \n> > > On Wed, 10 Mar 1999, Bruce Momjian wrote:\n> > > \n> > > > > OK, I have installed a rotating globe on the developers page. After I\n> > > > > get all the developer locations, I will add red dots to the globe.\n> > > > \n> > > > New map. Gimp Nova dots. You like?\n> > > > \n> > > \n> > > Looks great!!\n> > \n> > Hate to be the wet blanket...throw another one of those 'Nova dots' onto\n> > Europe, and Europe will sink :( It sort of reminds me of those old movies\n> > dealing with Nuclear WAr...\"New Brunswick, Canada...totally wiped' out,\n> > Maine, USA...like it never existed, England...WWI and WWII failed, where\n> > PostgreSQL succeseds\" *grin*\n> \n> OK, smaller stars, and fewer. Should I continue in this direction, or\n> try something else?\n\nI'd say continue, but go with \"points\" vs \"dabs\"...enough ppl get added to\nthat globe, there will be so much overlap of the 'dabs' that the country\nborders will be indistinguishable :)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 10 Mar 1999 11:41:35 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> > OK, smaller stars, and fewer. Should I continue in this direction, or\n> > try something else?\n> \n> I'd say continue, but go with \"points\" vs \"dabs\"...enough ppl get added to\n> that globe, there will be so much overlap of the 'dabs' that the country\n> borders will be indistinguishable :)\n> \n\nAmazing how much better the blurred dots look compared to the regular\ndots. Points are too small to see. If we get more people, I will\nshrink the dots.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 11:05:43 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Wed, 10 Mar 1999, Vince Vielhaber wrote:\n\n> On Wed, 10 Mar 1999, The Hermit Hacker wrote:\n> \n> > \n> > Definitely doable...see 'http://www.nova-scotia.org', a site we played\n> > with that used javascript...\n> \n> And I just found a nifty little js library for doing popups. Should have\n> a simple demo available shortly of index.html's frames.\n\nOk try this. I haven't tried it on a graphic yet, that'll be next. I\nwas just proving it out.\n\nhttp://www.postgresql.org/~vev/index.html\n\nMouseover on the left frame.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Wed, 10 Mar 1999 11:35:17 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> On Wed, 10 Mar 1999, Vince Vielhaber wrote:\n> \n> > On Wed, 10 Mar 1999, The Hermit Hacker wrote:\n> > \n> > > \n> > > Definitely doable...see 'http://www.nova-scotia.org', a site we played\n> > > with that used javascript...\n> > \n> > And I just found a nifty little js library for doing popups. Should have\n> > a simple demo available shortly of index.html's frames.\n> \n> Ok try this. I haven't tried it on a graphic yet, that'll be next. I\n> was just proving it out.\n> \n> http://www.postgresql.org/~vev/index.html\n> \n> Mouseover on the left frame.\n\nTotally cool.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 11:59:26 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": ">\n> > On Wed, 10 Mar 1999, Vince Vielhaber wrote:\n> >\n> > > On Wed, 10 Mar 1999, Bruce Momjian wrote:\n> > >\n> > > > > OK, I have installed a rotating globe on the developers page. After I\n> > > > > get all the developer locations, I will add red dots to the globe.\n> > > >\n> > > > New map. Gimp Nova dots. You like?\n> > > >\n> > >\n> > > Looks great!!\n> >\n> > Hate to be the wet blanket...throw another one of those 'Nova dots' onto\n> > Europe, and Europe will sink :( It sort of reminds me of those old movies\n> > dealing with Nuclear WAr...\"New Brunswick, Canada...totally wiped' out,\n> > Maine, USA...like it never existed, England...WWI and WWII failed, where\n> > PostgreSQL succeseds\" *grin*\n>\n> OK, smaller stars, and fewer. Should I continue in this direction, or\n> try something else?\n\n Actually I'm hacking a little around. Up to now I've\n extracted the coords of the world map out of xearth and\n produced a 600x250 GIF as a flat map. Having only a few\n colors it's about 6K small.\n\n Now I'm looking for spec's on animated GIF's. If I finally\n understand how they are created, I could build a little\n Tcl/Tk application where you can enter location (long/lat or\n lookup by city) and the text to popup and Tk will create an\n animated GIF where the locations are small blinking dot's\n plus the required image map etc.\n\n Just give me another day for an example.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 10 Mar 1999 18:45:46 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> > Ok try this. I haven't tried it on a graphic yet, that'll be next. I\n> > was just proving it out.\n> >\n> > http://www.postgresql.org/~vev/index.html\n> >\n> > Mouseover on the left frame.\n>\n> Totally cool.\n\n What about this one:\n\n http://www.postgresql.org/~wieck/test.html\n\n I've used xearth's coords for the map (a little Tcl/Tk script\n did it) and GIFmerge to combine two ones. Resulting gif is\n 6275 bytes!\n\n Then I've added overlib (thanks Venice). Only Bruce and me\n are on the map. Finally (if we choose it) I'll create the\n mentioned Tk application to manage all the dot's and the\n underlying map.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 10 Mar 1999 20:30:55 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> > > Ok try this. I haven't tried it on a graphic yet, that'll be next. I\n> > > was just proving it out.\n> > >\n> > > http://www.postgresql.org/~vev/index.html\n> > >\n> > > Mouseover on the left frame.\n> >\n> > Totally cool.\n> \n> What about this one:\n> \n> http://www.postgresql.org/~wieck/test.html\n> \n> I've used xearth's coords for the map (a little Tcl/Tk script\n> did it) and GIFmerge to combine two ones. Resulting gif is\n> 6275 bytes!\n> \n> Then I've added overlib (thanks Venice). Only Bruce and me\n> are on the map. Finally (if we choose it) I'll create the\n> mentioned Tk application to manage all the dot's and the\n> underlying map.\n\nI like the dots. Perhaps larger.\n\nCan you do longitute/latitude with that map. Is it linear in both\ndirections. Usually such linear maps look terrible because Greenland\nand Northern Canada are huge. Don't tell me you compensated for it in\nthe tcl code.\n\nCan you use the more attractive map I am using. Your blinking dots and\nhighlighted names are certainly better than my map.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 14:47:27 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "\nOn 10-Mar-99 Jan Wieck wrote:\n> What about this one:\n> \n> http://www.postgresql.org/~wieck/test.html\n> \n> I've used xearth's coords for the map (a little Tcl/Tk script\n> did it) and GIFmerge to combine two ones. Resulting gif is\n> 6275 bytes!\n> \n> Then I've added overlib (thanks Venice). Only Bruce and me\n> are on the map. Finally (if we choose it) I'll create the\n> mentioned Tk application to manage all the dot's and the\n> underlying map.\n\nLooks like you discovered overlib would do it's popups on top of an\nimage while I was trying to figure out what happened to Netscape's\nno longer showing the XY coords of an imagemap! Looks/works great,\ndots could be just a touch bigger, tho.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Wed, 10 Mar 1999 14:59:55 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> >\n> > What about this one:\n> >\n> > http://www.postgresql.org/~wieck/test.html\n> >\n>\n> I like the dots. Perhaps larger.\n\n Larger dots, larger image. It's 10595 bytes now.\n\n>\n> Can you do longitute/latitude with that map. Is it linear in both\n> directions. Usually such linear maps look terrible because Greenland\n> and Northern Canada are huge. Don't tell me you compensated for it in\n> the tcl code.\n\n The map was draw using the coordinates from xearth. So it is\n long/lat based and thus I can.\n\n It isn't linear in the x-direction. It uses a sine-function\n to shrink the X down the farer it's away from 0 meridian.\n\n Currently it isn't using the latitude to give it an ellipsian\n touch. But I could modify the algorithm that converts from\n xearth's x,y,z coordinates to the map's x,y.\n\n>\n> Can you use the more attractive map I am using. Your blinking dots and\n> highlighted names are certainly better than my map.\n\n That would loose the long/lat capability (except it already\n has it). And it would blow up the image size again, because\n for the animation at least two different gifs must be in it.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 10 Mar 1999 21:11:24 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> > directions. Usually such linear maps look terrible because Greenland\n> > and Northern Canada are huge. Don't tell me you compensated for it in\n> > the tcl code.\n>\n\n Finally I've increased the shrink factor a little. Looks\n better IMHO.\n\n Don't know how to put more color on it without adding tons of\n more coordinate data. Maybe something like a transparent\n looking \"Postgres\" text over it has more effect than any\n color. I'll check, but it's enough for today. I'm off\n catching some zzzzzzzz's.\n\n BTW: Did you notice that the point's are clickable?\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 10 Mar 1999 21:48:04 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> > > directions. Usually such linear maps look terrible because Greenland\n> > > and Northern Canada are huge. Don't tell me you compensated for it in\n> > > the tcl code.\n> >\n> \n> Finally I've increased the shrink factor a little. Looks\n> better IMHO.\n> \n> Don't know how to put more color on it without adding tons of\n> more coordinate data. Maybe something like a transparent\n> looking \"Postgres\" text over it has more effect than any\n> color. I'll check, but it's enough for today. I'm off\n> catching some zzzzzzzz's.\n> \n> BTW: Did you notice that the point's are clickable?\n\nI click on the point, and my Netscape 4.5 crashes.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 15:51:10 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": ">\n> > > > directions. Usually such linear maps look terrible because Greenland\n> > > > and Northern Canada are huge. Don't tell me you compensated for it in\n> > > > the tcl code.\n> > >\n> >\n> > Finally I've increased the shrink factor a little. Looks\n> > better IMHO.\n> >\n> > Don't know how to put more color on it without adding tons of\n> > more coordinate data. Maybe something like a transparent\n> > looking \"Postgres\" text over it has more effect than any\n> > color. I'll check, but it's enough for today. I'm off\n> > catching some zzzzzzzz's.\n> >\n> > BTW: Did you notice that the point's are clickable?\n>\n> I click on the point, and my Netscape 4.5 crashes.\n\n Uh?\n\n There's simply a mailto: href on them. My Netscape 4.05 pops\n up the Mail window then. Maybe there's something wrong with\n the href because the <> in it aren't escaped. But that\n shouldn't cause Netscape to crash.\n\n What happens if you use a normal mailto: link?\n\n\nPuzzled, Jan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 10 Mar 1999 21:57:59 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> > I click on the point, and my Netscape 4.5 crashes.\n> \n> Uh?\n> \n> There's simply a mailto: href on them. My Netscape 4.05 pops\n> up the Mail window then. Maybe there's something wrong with\n> the href because the <> in it aren't escaped. But that\n> shouldn't cause Netscape to crash.\n> \n> What happens if you use a normal mailto: link?\n\nThat explains it. Mailto crashes unless the mail window has already\nbeen opened earlier. Good idea.\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 16:08:05 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "\nOn 10-Mar-99 Jan Wieck wrote:\n>> > directions. Usually such linear maps look terrible because Greenland\n>> > and Northern Canada are huge. Don't tell me you compensated for it in\n>> > the tcl code.\n>>\n> \n> Finally I've increased the shrink factor a little. Looks\n> better IMHO.\n> \n> Don't know how to put more color on it without adding tons of\n> more coordinate data. Maybe something like a transparent\n> looking \"Postgres\" text over it has more effect than any\n> color. I'll check, but it's enough for today. I'm off\n> catching some zzzzzzzz's.\n> \n> BTW: Did you notice that the point's are clickable?\n\nYeah, for some reason mailto: crashes my browser (netscape 4.5). I think\nit's got something to do with nothing being defined for any servers or\nsomething like that.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Wed, 10 Mar 1999 16:23:38 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "It is a (un)known problem with NS 4.5. Marc had this pleasure a couple of\ndays before also. \n\nBTW: xearth rotates by itself in real time and there are a lot of flags to\ninfluence the behaviour and appearance.\n\n-Egon \n\nOn Wed, 10 Mar 1999, Vince Vielhaber wrote:\n\n> Yeah, for some reason mailto: crashes my browser (netscape 4.5). I think\n> it's got something to do with nothing being defined for any servers or\n> something like that.\n> \n> Vince.\n\n", "msg_date": "Wed, 10 Mar 1999 22:38:59 +0100 (MET)", "msg_from": "Egon Schmid <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "\nOn 10-Mar-99 Bruce Momjian wrote:\n>> > I click on the point, and my Netscape 4.5 crashes.\n>> \n>> Uh?\n>> \n>> There's simply a mailto: href on them. My Netscape 4.05 pops\n>> up the Mail window then. Maybe there's something wrong with\n>> the href because the <> in it aren't escaped. But that\n>> shouldn't cause Netscape to crash.\n>> \n>> What happens if you use a normal mailto: link?\n> \n> That explains it. Mailto crashes unless the mail window has already\n> been opened earlier. Good idea.\n\nI read on one of the FreeBSD lists today that 4.51 (linux version) still\ncrashes, so it's not fixed yet.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Wed, 10 Mar 1999 16:57:26 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> \n> On 10-Mar-99 Bruce Momjian wrote:\n> >> > I click on the point, and my Netscape 4.5 crashes.\n> >> \n> >> Uh?\n> >> \n> >> There's simply a mailto: href on them. My Netscape 4.05 pops\n> >> up the Mail window then. Maybe there's something wrong with\n> >> the href because the <> in it aren't escaped. But that\n> >> shouldn't cause Netscape to crash.\n> >> \n> >> What happens if you use a normal mailto: link?\n> > \n> > That explains it. Mailto crashes unless the mail window has already\n> > been opened earlier. Good idea.\n> \n> I read on one of the FreeBSD lists today that 4.51 (linux version) still\n> crashes, so it's not fixed yet.\n\nDidn't know 4.51 was out yet. Downloading now. Thanks.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 18:11:38 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "Bruce Momjian <[email protected]> writes:\n\n> Can you do longitute/latitude with that map. Is it linear in both\n> directions. Usually such linear maps look terrible because Greenland\n> and Northern Canada are huge. Don't tell me you compensated for it in\n> the tcl code.\n\nI also have a map program I wrote up for an IRC network.\n\nSee http://www.sorcery.net/maps/ for that implementation. It could\nuse a database backend easily, but right now it uses a combination of\nxearth, fly (a gif editing scripting language) and Perl script to\nwrite the fly script.\n\n--Michael\n", "msg_date": "10 Mar 1999 17:11:51 -0800", "msg_from": "Michael Graff <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Wed, 10 Mar 1999, Jan Wieck wrote:\n\n> > > Ok try this. I haven't tried it on a graphic yet, that'll be next. I\n> > > was just proving it out.\n> > >\n> > > http://www.postgresql.org/~vev/index.html\n> > >\n> > > Mouseover on the left frame.\n> >\n> > Totally cool.\n> \n> What about this one:\n> \n> http://www.postgresql.org/~wieck/test.html\n\nThis is *totally* cool...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 10 Mar 1999 22:28:28 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Wed, 10 Mar 1999, Jan Wieck wrote:\n\n> > > directions. Usually such linear maps look terrible because Greenland\n> > > and Northern Canada are huge. Don't tell me you compensated for it in\n> > > the tcl code.\n> >\n> \n> Finally I've increased the shrink factor a little. Looks\n> better IMHO.\n> \n> Don't know how to put more color on it without adding tons of\n> more coordinate data. Maybe something like a transparent\n> looking \"Postgres\" text over it has more effect than any\n> color. I'll check, but it's enough for today. I'm off\n> catching some zzzzzzzz's.\n> \n> BTW: Did you notice that the point's are clickable?\n\nErk...just tried that and it crashed my netscape :)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 10 Mar 1999 22:30:58 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Wed, 10 Mar 1999, Jan Wieck wrote:\n\n> Uh?\n> \n> There's simply a mailto: href on them. My Netscape 4.05 pops\n> up the Mail window then. Maybe there's something wrong with\n> the href because the <> in it aren't escaped. But that\n> shouldn't cause Netscape to crash.\n> \n> What happens if you use a normal mailto: link?\n\nCrashes also...a bug in 4.5...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 10 Mar 1999 22:31:26 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "The Hermit Hacker wrote:\n> \n> On Wed, 10 Mar 1999, Jan Wieck wrote:\n> \n> >\n> > What about this one:\n> >\n> > http://www.postgresql.org/~wieck/test.html\n> \n> This is *totally* cool...\n\nI like this too.\n\nVadim\n", "msg_date": "Thu, 11 Mar 1999 09:47:00 +0700", "msg_from": "Vadim Mikheev <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "Marc G. Fournier wrote:\n\n> On Wed, 10 Mar 1999, Jan Wieck wrote:\n>\n> > > > Ok try this. I haven't tried it on a graphic yet, that'll be next. I\n> > > > was just proving it out.\n> > > >\n> > > > http://www.postgresql.org/~vev/index.html\n> > > >\n> > > > Mouseover on the left frame.\n> > >\n> > > Totally cool.\n> >\n> > What about this one:\n> >\n> > http://www.postgresql.org/~wieck/test.html\n>\n> This is *totally* cool...\n\n And now DEEP FROZEN :-)\n\n http://www.postgresql.org/~wieck/index.html\n -------------------------------------------\n\n Well, the image is about 85K now (look at it before you\n reply). But I think it's one that is worth the bandwidth - a\n MUST SEE.\n\n Bruce: is that ATTRACTIVE enough for you?\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Thu, 11 Mar 1999 18:31:28 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "Jan Wieck wrote:\n> http://www.postgresql.org/~wieck/index.html\n\nIt's pretty neet. I'd remove the white flush by Flordia,\nperhaps moving it to a corner and introducing faint\nshadows on the lettering.\n\n:) Clark\n", "msg_date": "Thu, 11 Mar 1999 17:36:06 +0000", "msg_from": "Clark Evans <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> > This is *totally* cool...\n> \n> And now DEEP FROZEN :-)\n> \n> http://www.postgresql.org/~wieck/index.html\n> -------------------------------------------\n> \n> Well, the image is about 85K now (look at it before you\n> reply). But I think it's one that is worth the bandwidth - a\n> MUST SEE.\n> \n> Bruce: is that ATTRACTIVE enough for you?\n\nYes, certainly deep frozen. Winner.\n\nComments:\n\n\tUse as many colors as you can. No problem with many colors.\n\n\tMuch better than mine.\n\n\tCan you darken the PostgreSQL a little bit more, so the globe\nstands out a little more?\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Thu, 11 Mar 1999 12:42:53 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": ">\n> Jan Wieck wrote:\n> > http://www.postgresql.org/~wieck/index.html\n>\n> It's pretty neet. I'd remove the white flush by Flordia,\n> perhaps moving it to a corner and introducing faint\n> shadows on the lettering.\n>\n> :) Clark\n>\n\n I can't. I haven't created the background image and don't\n know who did or how.\n\n OTOH it is\n\n 1. At Florida - a sunny state.\n\n 2. Near the location of hub.org, where OUR sun is shining.\n\n So it is IMHO exactly where is should be.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Thu, 11 Mar 1999 18:46:18 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> >\n> > Jan Wieck wrote:\n> > > http://www.postgresql.org/~wieck/index.html\n> >\n> > It's pretty neet. I'd remove the white flush by Flordia,\n> > perhaps moving it to a corner and introducing faint\n> > shadows on the lettering.\n> >\n> > :) Clark\n> >\n> \n> I can't. I haven't created the background image and don't\n> know who did or how.\n> \n> OTOH it is\n> \n> 1. At Florida - a sunny state.\n> \n> 2. Near the location of hub.org, where OUR sun is shining.\n> \n> So it is IMHO exactly where is should be.\n\nActually, it is closest to me. Enough said. :-)\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Thu, 11 Mar 1999 12:48:55 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "On Thu, 11 Mar 1999, Jan Wieck wrote:\n\n> >\n> > Jan Wieck wrote:\n> > > http://www.postgresql.org/~wieck/index.html\n> >\n> > It's pretty neet. I'd remove the white flush by Flordia,\n> > perhaps moving it to a corner and introducing faint\n> > shadows on the lettering.\n> >\n> > :) Clark\n> >\n> \n> I can't. I haven't created the background image and don't\n> know who did or how.\n> \n> OTOH it is\n> \n> 1. At Florida - a sunny state.\n> \n> 2. Near the location of hub.org, where OUR sun is shining.\n> \n> So it is IMHO exactly where is should be.\n\nThe sun? Huh. I thought it was a hurricane hitting the east coast. :)\n\nLooks good, Jan!\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Thu, 11 Mar 1999 12:58:25 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": ">\n> > > This is *totally* cool...\n> >\n> > And now DEEP FROZEN :-)\n> >\n> > http://www.postgresql.org/~wieck/index.html\n> > -------------------------------------------\n> >\n> > Well, the image is about 85K now (look at it before you\n> > reply). But I think it's one that is worth the bandwidth - a\n> > MUST SEE.\n> >\n> > Bruce: is that ATTRACTIVE enough for you?\n>\n> Yes, certainly deep frozen. Winner.\n>\n> Comments:\n>\n> Use as many colors as you can. No problem with many colors.\n\n Tried to use 250 colors, but that blows the image up to 150K\n and more important, causes the image sometimes to flicker on\n my 200MHz Pentium notebook. So back to 60 colors.\n\n>\n> Much better than mine.\n\n Tnx.\n\n>\n> Can you darken the PostgreSQL a little bit more, so the globe\n> stands out a little more?\n\n Done - the other way round. Darken the map and lighten the\n PostgreSQL. You're right, look's better.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Thu, 11 Mar 1999 19:07:53 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "Jan Wieck wrote:\n> \n> \n> And now DEEP FROZEN :-)\n> \n> http://www.postgresql.org/~wieck/index.html\n> -------------------------------------------\n\nYes! Cool!!!\n\nVadim\n", "msg_date": "Fri, 12 Mar 1999 12:07:03 +0700", "msg_from": "Vadim Mikheev <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": ">\n> Jan Wieck wrote:\n> >\n> >\n> > And now DEEP FROZEN :-)\n> >\n> > http://www.postgresql.org/~wieck/index.html\n> > -------------------------------------------\n>\n> Yes! Cool!!!\n>\n> Vadim\n>\n\n Just put in another version. This time, all dot's are red but\n they flash frequently when displayed in a Netscape.\n\n The advantage is that the image is static and the flashes are\n done with JavaScript. So the image could get highest quality\n and become bigger and finally placed as jpeg.\n\n For sake I haven't found out how to determine the absolute\n position of an image unter MS IE4 yet. So they are simply\n static there. Does anyone know how to do that?\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 12 Mar 1999 22:53:32 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "\nOn 12-Mar-99 Jan Wieck wrote:\n>>\n> Just put in another version. This time, all dot's are red but\n> they flash frequently when displayed in a Netscape.\n> \n> The advantage is that the image is static and the flashes are\n> done with JavaScript. So the image could get highest quality\n> and become bigger and finally placed as jpeg.\n> \n> For sake I haven't found out how to determine the absolute\n> position of an image unter MS IE4 yet. So they are simply\n> static there. Does anyone know how to do that?\n\nNow that's slick!\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Fri, 12 Mar 1999 17:04:22 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": "> > For sake I haven't found out how to determine the absolute\n> > position of an image unter MS IE4 yet. So they are simply\n> > static there. Does anyone know how to do that?\n> \n> Now that's slick!\n> \n\nYes. Looks like stars. Also, I have added an office knick-knacks link\nto my personal web page. See the signature.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Fri, 12 Mar 1999 17:34:34 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Developers globe" }, { "msg_contents": ">\n>\n> On 12-Mar-99 Jan Wieck wrote:\n> >>\n> > Just put in another version. This time, all dot's are red but\n> > they flash frequently when displayed in a Netscape.\n> >\n> > The advantage is that the image is static and the flashes are\n> > done with JavaScript. So the image could get highest quality\n> > and become bigger and finally placed as jpeg.\n> >\n> > For sake I haven't found out how to determine the absolute\n> > position of an image unter MS IE4 yet. So they are simply\n> > static there. Does anyone know how to do that?\n>\n> Now that's slick!\n\n I've installed that one with different coloured flashes now\n on the main site developers page. The map image is now\n 700x270 in jpeg and with a good quality (95% - in lower\n quality the red dots get too diffuse).\n\n I checked that it works with Netscape 4.05 and MS-IE 4.0.\n\n Does anyone have problems with it?\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Mon, 15 Mar 1999 22:44:34 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Developers Globe (FINAL)" }, { "msg_contents": "> I've installed that one with different coloured flashes now\n> on the main site developers page. The map image is now\n> 700x270 in jpeg and with a good quality (95% - in lower\n> quality the red dots get too diffuse).\n> \n> I checked that it works with Netscape 4.05 and MS-IE 4.0.\n> \n> Does anyone have problems with it?\n\nGreat. Now I can see where everyone is.\n\nI don't think I like the fact the image is in its own frame. Makes the\nlist of developers very small. Was this necessary to get it working\nproperly? Also, what does it look like _without_ the PostgreSQL image\nmerged into it. Is it too plain?\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 15 Mar 1999 16:51:10 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: Developers Globe (FINAL)" }, { "msg_contents": ">\n> > I've installed that one with different coloured flashes now\n> > on the main site developers page. The map image is now\n> > 700x270 in jpeg and with a good quality (95% - in lower\n> > quality the red dots get too diffuse).\n> >\n> > I checked that it works with Netscape 4.05 and MS-IE 4.0.\n> >\n> > Does anyone have problems with it?\n>\n> Great. Now I can see where everyone is.\n>\n> I don't think I like the fact the image is in its own frame. Makes the\n> list of developers very small. Was this necessary to get it working\n> properly? Also, what does it look like _without_ the PostgreSQL image\n> merged into it. Is it too plain?\n\n I didn't like it too, but in it was necessary. Without\n dropping it into it's own frame I couldn't set\n SCROLLING=\"NO\". In MS InternetExplorer then a scrollbar\n appears/disappears very nervous with the size of the slider\n trying to follow the rightmost existing flash (totally\n irritating).\n\n Without the PostgreSQL image it is really very plain\n (spartanic).\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Mon, 15 Mar 1999 22:58:10 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "\nOn 15-Mar-99 Jan Wieck wrote:\n>>\n>>\n>> On 12-Mar-99 Jan Wieck wrote:\n>> >>\n>> > Just put in another version. This time, all dot's are red but\n>> > they flash frequently when displayed in a Netscape.\n>> >\n>> > The advantage is that the image is static and the flashes are\n>> > done with JavaScript. So the image could get highest quality\n>> > and become bigger and finally placed as jpeg.\n>> >\n>> > For sake I haven't found out how to determine the absolute\n>> > position of an image unter MS IE4 yet. So they are simply\n>> > static there. Does anyone know how to do that?\n>>\n>> Now that's slick!\n> \n> I've installed that one with different coloured flashes now\n> on the main site developers page. The map image is now\n> 700x270 in jpeg and with a good quality (95% - in lower\n> quality the red dots get too diffuse).\n> \n> I checked that it works with Netscape 4.05 and MS-IE 4.0.\n> \n> Does anyone have problems with it?\n\nLooks and works great here.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Mon, 15 Mar 1999 17:07:26 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "RE: [HACKERS] Developers Globe (FINAL)" }, { "msg_contents": "> I didn't like it too, but in it was necessary. Without\n> dropping it into it's own frame I couldn't set\n> SCROLLING=\"NO\". In MS InternetExplorer then a scrollbar\n> appears/disappears very nervous with the size of the slider\n> trying to follow the rightmost existing flash (totally\n> irritating).\n> \n> Without the PostgreSQL image it is really very plain\n> (spartanic).\n> \n\nJan, I accidentally deleted the page when updating the page via cvs. \nCan you put your copy on there again, or tell me how to get it?\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 15 Mar 1999 17:31:51 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> > I didn't like it too, but in it was necessary. Without\n> > dropping it into it's own frame I couldn't set\n> > SCROLLING=\"NO\". In MS InternetExplorer then a scrollbar\n> > appears/disappears very nervous with the size of the slider\n> > trying to follow the rightmost existing flash (totally\n> > irritating).\n> > \n> > Without the PostgreSQL image it is really very plain\n> > (spartanic).\n> > \n> \n> Jan, I accidentally deleted the page when updating the page via cvs. \n> Can you put your copy on there again, or tell me how to get it?\n> \n\nSorry for messing it up.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 15 Mar 1999 17:47:41 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "\nOn 15-Mar-99 Bruce Momjian wrote:\n>> I didn't like it too, but in it was necessary. Without\n>> dropping it into it's own frame I couldn't set\n>> SCROLLING=\"NO\". In MS InternetExplorer then a scrollbar\n>> appears/disappears very nervous with the size of the slider\n>> trying to follow the rightmost existing flash (totally\n>> irritating).\n>> \n>> Without the PostgreSQL image it is really very plain\n>> (spartanic).\n>> \n> \n> Jan, I accidentally deleted the page when updating the page via cvs. \n> Can you put your copy on there again, or tell me how to get it?\n\nIsn't it the same as in his $HOME/public_html dir? If not I still have\nit up on my browser, I can save it off easy enuf.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Mon, 15 Mar 1999 17:51:53 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> \n> On 15-Mar-99 Bruce Momjian wrote:\n> >> I didn't like it too, but in it was necessary. Without\n> >> dropping it into it's own frame I couldn't set\n> >> SCROLLING=\"NO\". In MS InternetExplorer then a scrollbar\n> >> appears/disappears very nervous with the size of the slider\n> >> trying to follow the rightmost existing flash (totally\n> >> irritating).\n> >> \n> >> Without the PostgreSQL image it is really very plain\n> >> (spartanic).\n> >> \n> > \n> > Jan, I accidentally deleted the page when updating the page via cvs. \n> > Can you put your copy on there again, or tell me how to get it?\n> \n> Isn't it the same as in his $HOME/public_html dir? If not I still have\n> it up on my browser, I can save it off easy enuf.\n\nDon't think so. It is framed for the other date in the page. Can you\nsend it over. I need devel-contrib.html\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 15 Mar 1999 17:54:23 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "\nOn 15-Mar-99 Bruce Momjian wrote:\n>> \n>> On 15-Mar-99 Bruce Momjian wrote:\n>> >> I didn't like it too, but in it was necessary. Without\n>> >> dropping it into it's own frame I couldn't set\n>> >> SCROLLING=\"NO\". In MS InternetExplorer then a scrollbar\n>> >> appears/disappears very nervous with the size of the slider\n>> >> trying to follow the rightmost existing flash (totally\n>> >> irritating).\n>> >> \n>> >> Without the PostgreSQL image it is really very plain\n>> >> (spartanic).\n>> >> \n>> > \n>> > Jan, I accidentally deleted the page when updating the page via cvs. \n>> > Can you put your copy on there again, or tell me how to get it?\n>> \n>> Isn't it the same as in his $HOME/public_html dir? If not I still have\n>> it up on my browser, I can save it off easy enuf.\n> \n> Don't think so. It is framed for the other date in the page. Can you\n> send it over. I need devel-contrib.html\n\nLook in my $HOME dir to make sure it's what you're looking for. If not I\nhave the rest as well.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Mon, 15 Mar 1999 17:57:35 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> > \n> > Don't think so. It is framed for the other date in the page. Can you\n> > send it over. I need devel-contrib.html\n> \n> Look in my $HOME dir to make sure it's what you're looking for. If not I\n> have the rest as well.\n> \n\nNo. It should have frames for the new globe. This is for the left\nmargin. \n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 15 Mar 1999 18:00:47 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> > Isn't it the same as in his $HOME/public_html dir? If not I still have\n> > it up on my browser, I can save it off easy enuf.\n> \n> Don't think so. It is framed for the other date in the page. Can you\n> send it over. I need devel-contrib.html\n\nVince got me the right file. Thanks. No need for you to get it Jan.\n\nI will check it into cvs. We have a 'www' branch for the web site.\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 15 Mar 1999 18:14:45 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> > > Isn't it the same as in his $HOME/public_html dir? If not I still have\n> > > it up on my browser, I can save it off easy enuf.\n> > \n> > Don't think so. It is framed for the other date in the page. Can you\n> > send it over. I need devel-contrib.html\n> \n> Vince got me the right file. Thanks. No need for you to get it Jan.\n> \n> I will check it into cvs. We have a 'www' branch for the web site.\n\nI have checked in Jan's html, but not the subdirectories, which are\nprobably generated by Jan's program. Probably will not change very\noften, so no need to cvs it. Do people agree?\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 15 Mar 1999 18:21:10 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "Jan Wieck wrote:\n> \n> I checked that it works with Netscape 4.05 and MS-IE 4.0.\n> \n> Does anyone have problems with it?\n\nI just got:\n\nJavaScript Error: http://www.postgresql.org/~wieck/index.html, line 31:\nie4 is not defined.\n\nusing the FreeBSD Netscape 4.05 binary.\n\nCheers,\n\nPatrick\n", "msg_date": "Mon, 15 Mar 1999 23:51:15 +0000 (GMT)", "msg_from": "\"Patrick Welche\" <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers Globe (FINAL)" }, { "msg_contents": "[email protected] (Jan Wieck) writes:\n> I checked that it works with Netscape 4.05 and MS-IE 4.0.\n> Does anyone have problems with it?\n\nIt fails rather spectacularly for me --- I'm watching a Netscape image\nplaceholder icon bounce around the empty frame as if it were a ping-pong\nball. Kinda fun to watch actually, but I'm sure it's not the intended\neffect. It's sucking up an unreasonable percentage of my CPU, too.\n\nThis is Netscape 4.08 for HPUX, auto image load off, and I currently\nhave Javascript but not Java enabled. With Javascript off (my more\nusual browsing setup, but I chanced to leave it on today) the page\nbehaves a lot more reasonably; I get one placeholder that I can click on\nif I feel like looking at the graphic.\n\nI realize that you spent a good deal of time on that \"flashing dot\"\neffect, but I'd counsel you to forget it. It doesn't add a darn thing\nto the useful content of the page --- the red dots are perfectly visible\nwithout flashing. (You could make them a tiny bit bigger and brighter,\nif you find them marginal.) What it *will* do is create all sorts of\nportability headaches, of which you've seen only the beginning; you\nhave not tried to get it to work across multiple browser versions or\nmultiple platforms, to say nothing of browsers other than the Big Two.\n\"I hacked it till it worked on the two browsers I use\" is just about the\ndefinition of bad Web design in my book. You'd be embarrassed to ship\nC code that unportable --- why are you willing to accept it in HTML?\n\n\t\t\tregards, tom lane\n", "msg_date": "Mon, 15 Mar 1999 22:15:49 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers Globe (FINAL) " }, { "msg_contents": "On Mon, 15 Mar 1999, Jan Wieck wrote:\n\n> >\n> >\n> > On 12-Mar-99 Jan Wieck wrote:\n> > >>\n> > > Just put in another version. This time, all dot's are red but\n> > > they flash frequently when displayed in a Netscape.\n> > >\n> > > The advantage is that the image is static and the flashes are\n> > > done with JavaScript. So the image could get highest quality\n> > > and become bigger and finally placed as jpeg.\n> > >\n> > > For sake I haven't found out how to determine the absolute\n> > > position of an image unter MS IE4 yet. So they are simply\n> > > static there. Does anyone know how to do that?\n> >\n> > Now that's slick!\n> \n> I've installed that one with different coloured flashes now\n> on the main site developers page. The map image is now\n> 700x270 in jpeg and with a good quality (95% - in lower\n> quality the red dots get too diffuse).\n> \n> I checked that it works with Netscape 4.05 and MS-IE 4.0.\n> \n> Does anyone have problems with it?\n\nMy only comment/problem...the underlying logo, IMHO, detracts from the\ndots ... some of them you really have to pick at to see, with the logo\novershadowing it...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Tue, 16 Mar 1999 00:01:36 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: Developers Globe (FINAL)" }, { "msg_contents": "> > \n> > I've installed that one with different coloured flashes now\n> > on the main site developers page. The map image is now\n> > 700x270 in jpeg and with a good quality (95% - in lower\n> > quality the red dots get too diffuse).\n> > \n> > I checked that it works with Netscape 4.05 and MS-IE 4.0.\n> > \n> > Does anyone have problems with it?\n> \n> My only comment/problem...the underlying logo, IMHO, detracts from the\n> dots ... some of them you really have to pick at to see, with the logo\n> overshadowing it...\n\nI have to agree on this. \n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 15 Mar 1999 23:09:21 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: Developers Globe (FINAL)" }, { "msg_contents": ">\n> > > > Isn't it the same as in his $HOME/public_html dir? If not I still have\n> > > > it up on my browser, I can save it off easy enuf.\n> > >\n> > > Don't think so. It is framed for the other date in the page. Can you\n> > > send it over. I need devel-contrib.html\n> >\n> > Vince got me the right file. Thanks. No need for you to get it Jan.\n> >\n> > I will check it into cvs. We have a 'www' branch for the web site.\n>\n> I have checked in Jan's html, but not the subdirectories, which are\n> probably generated by Jan's program. Probably will not change very\n> often, so no need to cvs it. Do people agree?\n\n I've just committed my last changes to the www module and put\n the files onto the site. Now the globe is on the page again\n without a separate frame (much better).\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Tue, 16 Mar 1999 10:47:24 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": ">\n> [email protected] (Jan Wieck) writes:\n> > I checked that it works with Netscape 4.05 and MS-IE 4.0.\n> > Does anyone have problems with it?\n>\n> It fails rather spectacularly for me --- I'm watching a Netscape image\n> placeholder icon bounce around the empty frame as if it were a ping-pong\n> ball. Kinda fun to watch actually, but I'm sure it's not the intended\n> effect. It's sucking up an unreasonable percentage of my CPU, too.\n\n I know that there is still a bunch of problems with it and be\n sure, I'll work on. Finally it should display a static map by\n default and only if it finds a known browser, fill in the\n flash images into empty layers and start blinking.\n\n What it is currently is only a point I'm starting from.\n\n>\n> This is Netscape 4.08 for HPUX, auto image load off, and I currently\n> have Javascript but not Java enabled. With Javascript off (my more\n> usual browsing setup, but I chanced to leave it on today) the page\n> behaves a lot more reasonably; I get one placeholder that I can click on\n> if I feel like looking at the graphic.\n\n Ech - auto load off. Thanks for the hint.\n\n>\n> I realize that you spent a good deal of time on that \"flashing dot\"\n> effect, but I'd counsel you to forget it. It doesn't add a darn thing\n> to the useful content of the page --- the red dots are perfectly visible\n> without flashing. (You could make them a tiny bit bigger and brighter,\n> if you find them marginal.) What it *will* do is create all sorts of\n> portability headaches, of which you've seen only the beginning; you\n> have not tried to get it to work across multiple browser versions or\n> multiple platforms, to say nothing of browsers other than the Big Two.\n> \"I hacked it till it worked on the two browsers I use\" is just about the\n> definition of bad Web design in my book. You'd be embarrassed to ship\n> C code that unportable --- why are you willing to accept it in HTML?\n\n But then you have to vote for removing overlib from the main\n index too since as far as I've understood the code it does\n too less checking about browser versions etc. We want such\n attracktive things even if only the big two support them -\n no?\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Tue, 16 Mar 1999 11:28:09 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers Globe (FINAL)" }, { "msg_contents": ">\n> > >\n> > > I've installed that one with different coloured flashes now\n> > > on the main site developers page. The map image is now\n> > > 700x270 in jpeg and with a good quality (95% - in lower\n> > > quality the red dots get too diffuse).\n> > >\n> > > I checked that it works with Netscape 4.05 and MS-IE 4.0.\n> > >\n> > > Does anyone have problems with it?\n> >\n> > My only comment/problem...the underlying logo, IMHO, detracts from the\n> > dots ... some of them you really have to pick at to see, with the logo\n> > overshadowing it...\n>\n> I have to agree on this.\n\n Man - first it is too flat, now the logo detracs, the flashes\n cause problems and ppl vote agains. Next someone want's it\n back rotating :-)\n\n OK, ok - but slow motion now. A version that works for NS4\n and IE4 is in place. Let's leave it there until we have\n something better. I really wanted to get my fingers back on\n raytracing for a long time and that time is now. I'll produce\n some different map's the next days and place them all into\n ~wieck/index.html, then we can have a voting.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Tue, 16 Mar 1999 11:37:18 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "On Tue, 16 Mar 1999, Jan Wieck wrote:\n\n> But then you have to vote for removing overlib from the main\n> index too since as far as I've understood the code it does\n> too less checking about browser versions etc. We want such\n> attracktive things even if only the big two support them -\n> no?\n\nOut of the last 150k different computers hitting all of the web sites\nhosted by Hub, 49.89% reported Netscape 4.x browers, while 31.84% reported\nIE4.x ...\n\nUse the Attractive Things...\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Tue, 16 Mar 1999 08:49:56 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Developers Globe (FINAL)" }, { "msg_contents": "> > I have to agree on this.\n> \n> Man - first it is too flat, now the logo detracs, the flashes\n> cause problems and ppl vote agains. Next someone want's it\n> back rotating :-)\n> \n> OK, ok - but slow motion now. A version that works for NS4\n> and IE4 is in place. Let's leave it there until we have\n> something better. I really wanted to get my fingers back on\n> raytracing for a long time and that time is now. I'll produce\n> some different map's the next days and place them all into\n> ~wieck/index.html, then we can have a voting.\n\nNow you know how I felt with the globe(too fast, hard to see), and\nstatic map(bad dots).\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 16 Mar 1999 13:41:19 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": ">\n> > > I have to agree on this.\n> >\n> > Man - first it is too flat, now the logo detracs, the flashes\n> > cause problems and ppl vote agains. Next someone want's it\n> > back rotating :-)\n> >\n> > OK, ok - but slow motion now. A version that works for NS4\n> > and IE4 is in place. Let's leave it there until we have\n> > something better. I really wanted to get my fingers back on\n> > raytracing for a long time and that time is now. I'll produce\n> > some different map's the next days and place them all into\n> > ~wieck/index.html, then we can have a voting.\n>\n> Now you know how I felt with the globe(too fast, hard to see), and\n> static map(bad dots).\n\n The first raytraced example (just a quick shot :-) is in\n place.\n\n http://www.postgresql.org/~wieck/index.html\n\n As you might expect, the locations of the nails are from the\n same data I already have. I just bent them a little so every\n location is visible. There is no map behind, but that's easy\n and could have it's hot areas under the knobs of the nails.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Tue, 16 Mar 1999 20:53:55 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> > Now you know how I felt with the globe(too fast, hard to see), and\n> > static map(bad dots).\n> \n> The first raytraced example (just a quick shot :-) is in\n> place.\n> \n> http://www.postgresql.org/~wieck/index.html\n> \n> As you might expect, the locations of the nails are from the\n> same data I already have. I just bent them a little so every\n> location is visible. There is no map behind, but that's easy\n> and could have it's hot areas under the knobs of the nails.\n\nWoh. I didn't know what ratracing was, but I know I like it. Certainly\njumps out at you. Is it in a paper tray?\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 16 Mar 1999 15:09:52 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> >\n> > > > I have to agree on this.\n> > >\n> > > Man - first it is too flat, now the logo detracs, the flashes\n> > > cause problems and ppl vote agains. Next someone want's it\n> > > back rotating :-)\n> > >\n> > > OK, ok - but slow motion now. A version that works for NS4\n> > > and IE4 is in place. Let's leave it there until we have\n> > > something better. I really wanted to get my fingers back on\n> > > raytracing for a long time and that time is now. I'll produce\n> > > some different map's the next days and place them all into\n> > > ~wieck/index.html, then we can have a voting.\n> >\n> > Now you know how I felt with the globe(too fast, hard to see), and\n> > static map(bad dots).\n\nWhen you put people's names on a page, amazing how interested they\nbecome in the page.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 16 Mar 1999 15:10:26 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "\nOn 16-Mar-99 Bruce Momjian wrote:\n>> > Now you know how I felt with the globe(too fast, hard to see), and\n>> > static map(bad dots).\n>> \n>> The first raytraced example (just a quick shot :-) is in\n>> place.\n>> \n>> http://www.postgresql.org/~wieck/index.html\n>> \n>> As you might expect, the locations of the nails are from the\n>> same data I already have. I just bent them a little so every\n>> location is visible. There is no map behind, but that's easy\n>> and could have it's hot areas under the knobs of the nails.\n> \n> Woh. I didn't know what ratracing was, but I know I like it. Certainly\n> jumps out at you. Is it in a paper tray?\n\nAlmost looks like a wooden box with water in it and the continents are\nfloating on the water. Jan, what are you using to create that?\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Tue, 16 Mar 1999 15:20:39 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": ">\n>\n> On 16-Mar-99 Bruce Momjian wrote:\n> >> > Now you know how I felt with the globe(too fast, hard to see), and\n> >> > static map(bad dots).\n> >>\n> >> The first raytraced example (just a quick shot :-) is in\n> >> place.\n> >>\n> >> http://www.postgresql.org/~wieck/index.html\n> >>\n> >> As you might expect, the locations of the nails are from the\n> >> same data I already have. I just bent them a little so every\n> >> location is visible. There is no map behind, but that's easy\n> >> and could have it's hot areas under the knobs of the nails.\n> >\n> > Woh. I didn't know what ratracing was, but I know I like it. Certainly\n> > jumps out at you. Is it in a paper tray?\n>\n> Almost looks like a wooden box with water in it and the continents are\n> floating on the water. Jan, what are you using to create that?\n>\n> Vince.\n\n It's a wooden box with some water looking ground - right. The\n continents sit on a glass surface above. Actually I'm playing\n around with some bump mapping on the glass so it will be\n better visible. And there where some errors in the images I\n used for mapping, better soon.\n\n I'm using rayshade and the utah raster toolkit for that.\n\n It's not the final version. Finally there will be some\n objects around the whole scene that produce reflections on\n the glass so it looks like the box is laying on a table\n inside of a room.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Tue, 16 Mar 1999 21:28:37 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": ">\n> > http://www.postgresql.org/~wieck/index.html\n> >\n> Woh. I didn't know what ratracing was, but I know I like it. Certainly\n> jumps out at you. Is it in a paper tray?\n\n Do a reload. Just to show you what \"HIGH QUALITY\" rendering\n means.\n\n It's still the same scene - just the border of the box a\n little higher, the glass better visible and sized to 600x300\n (to better fit on the page). This time rendered with a\n sampling rate of 5^2 per pixel (so per pixel 25 eye rays are\n shot into the scene). Including all shadowing, transparency\n and reflection a total of over 20 million rays have been\n traced (for a picture with 180,000 pixels)!\n\n Those pictures are CPU killers. A 333 MHz Celeron needs 20\n minutes to finish that one.\n\n Remember, it's a 100% virtual picture. Not one single detail\n is from a scanned photo or the like.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 17 Mar 1999 01:06:51 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> minutes to finish that one.\n> \n> Remember, it's a 100% virtual picture. Not one single detail\n> is from a scanned photo or the like.\n> \n\nYou realize that now that we know you can do these images, a new\nPostgreSQL logo request is coming! (Marc, ask him. Ask him!)\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 16 Mar 1999 19:29:48 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "<raytracing details snipped>\n> \n> Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> minutes to finish that one.\n> \n> Remember, it's a 100% virtual picture. Not one single detail\n> is from a scanned photo or the like.\n> \n\nAnd to make it just that little bit more life-like - \n\nHow about either:\n1 - bumpmap the continents to match the glass, then don't do a drop shadow\non the glass - so they look 'painted' on\n\nor\n\n2 - extrude the continents, so they have 'edges', and look like blocks of\nfoam pasted on the glass (with or w/o a finer grained bumpmap of their own)\n\nYou've got all the CPU in the world, right? ;-)\n\nRoss\n-- \nRoss J. Reedstrom, Ph.D., <[email protected]> \nNSBRI Research Scientist/Programmer\nComputer and Information Technology Institute\nRice University, 6100 S. Main St., Houston, TX 77005\n", "msg_date": "Tue, 16 Mar 1999 18:38:53 -0600 (CST)", "msg_from": "[email protected] (Ross J. Reedstrom)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "Jan, your picture looks great. I'm going to have a presentation at\nLinuxWorld Conference in Japan about PostgreSQL tomorrow and I would\nlike to show your picture to tell people how PostgreSQL developers\nwidely spread in the world. Can I use that?\n\nOf course I will put a copyright notice of you along with the picture.\n--\nTatsuo Ishii\n\n>> > http://www.postgresql.org/~wieck/index.html\n>> >\n>> Woh. I didn't know what ratracing was, but I know I like it. Certainly\n>> jumps out at you. Is it in a paper tray?\n>\n> Do a reload. Just to show you what \"HIGH QUALITY\" rendering\n> means.\n>\n> It's still the same scene - just the border of the box a\n> little higher, the glass better visible and sized to 600x300\n> (to better fit on the page). This time rendered with a\n> sampling rate of 5^2 per pixel (so per pixel 25 eye rays are\n> shot into the scene). Including all shadowing, transparency\n> and reflection a total of over 20 million rays have been\n> traced (for a picture with 180,000 pixels)!\n>\n> Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> minutes to finish that one.\n>\n> Remember, it's a 100% virtual picture. Not one single detail\n> is from a scanned photo or the like.\n>\n>\n>Jan\n>\n>--\n>\n>#======================================================================#\n># It's easier to get forgiveness for being wrong than for being right. #\n># Let's break this rule - forgive me. #\n>#======================================== [email protected] (Jan Wieck) #\n>\n>\n>\n\n", "msg_date": "Wed, 17 Mar 1999 11:08:23 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL) " }, { "msg_contents": "> It's a wooden box with some water looking ground - right. The\n> continents sit on a glass surface above. Actually I'm playing\n> around with some bump mapping on the glass so it will be\n> better visible. And there where some errors in the images I\n> used for mapping, better soon.\n\nJan, can I recommend you change the perspective of viewing the globe\nfrom southern to northern. Because we are mostly in the northern\nhemisphere, a perspective change would make that area larger, like\nEurope.\n\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 17 Mar 1999 00:02:14 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "\nJan...\n\n\tAny way of coming up with a suitable Logo to replace what we have\non the main screen? Something tasteful that jumps out at you? I like the\none we do have, but I don't think its very \"strong\"? How about it? :)\n\n\nOn Wed, 17 Mar 1999, Jan Wieck wrote:\n\n> >\n> > > http://www.postgresql.org/~wieck/index.html\n> > >\n> > Woh. I didn't know what ratracing was, but I know I like it. Certainly\n> > jumps out at you. Is it in a paper tray?\n> \n> Do a reload. Just to show you what \"HIGH QUALITY\" rendering\n> means.\n> \n> It's still the same scene - just the border of the box a\n> little higher, the glass better visible and sized to 600x300\n> (to better fit on the page). This time rendered with a\n> sampling rate of 5^2 per pixel (so per pixel 25 eye rays are\n> shot into the scene). Including all shadowing, transparency\n> and reflection a total of over 20 million rays have been\n> traced (for a picture with 180,000 pixels)!\n> \n> Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> minutes to finish that one.\n> \n> Remember, it's a 100% virtual picture. Not one single detail\n> is from a scanned photo or the like.\n> \n> \n> Jan\n> \n> --\n> \n> #======================================================================#\n> # It's easier to get forgiveness for being wrong than for being right. #\n> # Let's break this rule - forgive me. #\n> #======================================== [email protected] (Jan Wieck) #\n> \n> \n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 17 Mar 1999 01:11:49 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "Tatsuo Ishii wrote:\n>\n> Jan, your picture looks great. I'm going to have a presentation at\n> LinuxWorld Conference in Japan about PostgreSQL tomorrow and I would\n> like to show your picture to tell people how PostgreSQL developers\n> widely spread in the world. Can I use that?\n\n Yeah - use it!\n\n>\n> Of course I will put a copyright notice of you along with the picture.\n\n It is based on free informations (the map coords are from\n xearth). And free software (rayshade) made it. Do whatever\n you want with it - it's free.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 17 Mar 1999 09:47:36 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": ">\n> > It's a wooden box with some water looking ground - right. The\n> > continents sit on a glass surface above. Actually I'm playing\n> > around with some bump mapping on the glass so it will be\n> > better visible. And there where some errors in the images I\n> > used for mapping, better soon.\n>\n> Jan, can I recommend you change the perspective of viewing the globe\n> from southern to northern. Because we are mostly in the northern\n> hemisphere, a perspective change would make that area larger, like\n> Europe.\n\n I tried that, but the nails in europe are packed that close\n that they hide each other then. Wasn't easy to find a point\n of view where everyone is visible.\n\n I'll first create a magnifying glass over europe as sample 2\n :-)\n\n The next I'll try is a flat map directly looked at from above\n with all the names put with arrows onto a transparent layer\n above. That would make overlib/JavaScript obsolete! But it\n requires that the names are readable at last.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 17 Mar 1999 10:00:02 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "Marc G. Fournier wrote:\n\n> Jan...\n>\n> Any way of coming up with a suitable Logo to replace what we have\n> on the main screen? Something tasteful that jumps out at you? I like the\n> one we do have, but I don't think its very \"strong\"? How about it? :)\n\n Nothing's impossible with a raytracer and some time! And I\n expected this question while creating the map :-)\n\n But it would be nice if some others would share some ideas\n for it.\n\n My first thought was a big truck on a road in the desert with\n some cirrus clouds in the sky and \"PostgreSQL\" written on the\n trailer. Maybe some snow capped mountains in the background.\n\n Don't think that such a detailed scene is good for a small\n Powered By logo, and the new logo must definitely fit for\n that too.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 17 Mar 1999 10:10:51 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": ">Tatsuo Ishii wrote:\n>>\n>> Jan, your picture looks great. I'm going to have a presentation at\n>> LinuxWorld Conference in Japan about PostgreSQL tomorrow and I would\n>> like to show your picture to tell people how PostgreSQL developers\n>> widely spread in the world. Can I use that?\n>\n> Yeah - use it!\n\nThanks!\n\n>>\n>> Of course I will put a copyright notice of you along with the picture.\n>\n> It is based on free informations (the map coords are from\n> xearth). And free software (rayshade) made it. Do whatever\n> you want with it - it's free.\n\nOk.\n--\nTatsuo Ishii\n", "msg_date": "Wed, 17 Mar 1999 19:10:13 +0900", "msg_from": "Tatsuo Ishii <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL) " }, { "msg_contents": "Bruce Momjian wrote:\n\n>\n> > Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> > minutes to finish that one.\n> >\n> > Remember, it's a 100% virtual picture. Not one single detail\n> > is from a scanned photo or the like.\n> >\n>\n> You realize that now that we know you can do these images, a new\n> PostgreSQL logo request is coming! (Marc, ask him. Ask him!)\n\n A first one is on my index. I prefer the version with ORDBMS.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 19 Mar 1999 19:14:18 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "On Fri, 19 Mar 1999, Jan Wieck wrote:\n\n> Bruce Momjian wrote:\n> \n> >\n> > > Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> > > minutes to finish that one.\n> > >\n> > > Remember, it's a 100% virtual picture. Not one single detail\n> > > is from a scanned photo or the like.\n> > >\n> >\n> > You realize that now that we know you can do these images, a new\n> > PostgreSQL logo request is coming! (Marc, ask him. Ask him!)\n> \n> A first one is on my index. I prefer the version with ORDBMS.\n\nI agree! What are anyone's thoughts to having the web background the\nsame as the logo's background?\n\nOne suggestion on the logo, tho, can you dim the light that's shining\non it? It sorta takes away from the rest of it. (Yes I know it's\na virtual light)\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Fri, 19 Mar 1999 13:36:55 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "On Fri, 19 Mar 1999, Jan Wieck wrote:\n\n> Bruce Momjian wrote:\n> \n> >\n> > > Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> > > minutes to finish that one.\n> > >\n> > > Remember, it's a 100% virtual picture. Not one single detail\n> > > is from a scanned photo or the like.\n> > >\n> >\n> > You realize that now that we know you can do these images, a new\n> > PostgreSQL logo request is coming! (Marc, ask him. Ask him!)\n> \n> A first one is on my index. I prefer the version with ORDBMS.\n\nShocking to most, I like...nice, simple, not too overwhelming...but (ya\nya, always a but)...in the Powered By one, the left side is really faint,\nalmost to the point of unable to read the letters...?\n\nAs for preference...ya, Ilike the ORDBS one too...\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Fri, 19 Mar 1999 14:43:01 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "On Fri, 19 Mar 1999, Vince Vielhaber wrote:\n\n> On Fri, 19 Mar 1999, Jan Wieck wrote:\n> \n> > Bruce Momjian wrote:\n> > \n> > >\n> > > > Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> > > > minutes to finish that one.\n> > > >\n> > > > Remember, it's a 100% virtual picture. Not one single detail\n> > > > is from a scanned photo or the like.\n> > > >\n> > >\n> > > You realize that now that we know you can do these images, a new\n> > > PostgreSQL logo request is coming! (Marc, ask him. Ask him!)\n> > \n> > A first one is on my index. I prefer the version with ORDBMS.\n> \n> I agree! What are anyone's thoughts to having the web background the\n> same as the logo's background?\n\n\tUmmm...it depends on how it looks. ONe thing that I *really* hate\nabout backgrounds is they generally make reading teh page\ndifficult...there are *alot* of bad sites out there that are a result of\nit :(\n\n\tIf it can be done right, it would be nice if it could all look\nlike one solid piece, yes...\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Fri, 19 Mar 1999 14:44:11 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> > I agree! What are anyone's thoughts to having the web background the\n> > same as the logo's background?\n>\n> Ummm...it depends on how it looks. ONe thing that I *really* hate\n> about backgrounds is they generally make reading teh page\n> difficult...there are *alot* of bad sites out there that are a result of\n> it :(\n>\n> If it can be done right, it would be nice if it could all look\n> like one solid piece, yes...\n\n That's impossible. The background is a sky texture which\n isn't repeating. It is computed by a fractional brownian\n motion function.\n\n For a web page background, you need a repeating image that\n can be tiled.\n\n I like the sky in the background, because it carries a\n message. But I wouldn't mind too much removing it and making\n the background transparent.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 19 Mar 1999 20:00:04 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> Bruce Momjian wrote:\n> \n> >\n> > > Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> > > minutes to finish that one.\n> > >\n> > > Remember, it's a 100% virtual picture. Not one single detail\n> > > is from a scanned photo or the like.\n> > >\n> >\n> > You realize that now that we know you can do these images, a new\n> > PostgreSQL logo request is coming! (Marc, ask him. Ask him!)\n\nI think I like the 3rd developers globe. Can we try it without\nstickpins? I think it is looking confusing. Looks like there is enough\nroom in England on that image to show everyone as red dome, rather than\na stickpin.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Fri, 19 Mar 1999 14:01:30 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> same as the logo's background?\n> \n> One suggestion on the logo, tho, can you dim the light that's shining\n> on it? It sorta takes away from the rest of it. (Yes I know it's\n> a virtual light)\n> \n\nHere is an idea. I have thought about making the name stand out as a\npattern within a grid/table. \n\nIf you remember the end of the movie 2001, where HAL has those red bars\ncoming out as he is disconnected. We could have PostgreSQL spelled out\nas red bard sticking out in a grid:\n\n\n\t---\n\t- -\n\t---\n\t-\n\t_\n\nlike this. Of course, the grid would have to have lots of bars to look\ngood. I thought this would go well with an SQL/table motif.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Fri, 19 Mar 1999 14:04:36 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> That's impossible. The background is a sky texture which\n> isn't repeating. It is computed by a fractional brownian\n> motion function.\n> \n> For a web page background, you need a repeating image that\n> can be tiled.\n> \n> I like the sky in the background, because it carries a\n> message. But I wouldn't mind too much removing it and making\n> the background transparent.\n\nMaybe more of a sunset/sunrise sky, so it has some color, and looks less\nlike an MS-Windows background. The setting sun could be in front of the\nname, illuminating it with dramatic color.\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Fri, 19 Mar 1999 14:07:07 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": ">\n> > same as the logo's background?\n> >\n> > One suggestion on the logo, tho, can you dim the light that's shining\n> > on it? It sorta takes away from the rest of it. (Yes I know it's\n> > a virtual light)\n> >\n>\n> Here is an idea. I have thought about making the name stand out as a\n> pattern within a grid/table.\n>\n> If you remember the end of the movie 2001, where HAL has those red bars\n> coming out as he is disconnected. We could have PostgreSQL spelled out\n> as red bard sticking out in a grid:\n\n Hmmm - you mean his crystal storage which Dave was pushing\n and they smoothly came out - no?\n\n Nice idea.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 19 Mar 1999 20:16:41 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> >\n> > > same as the logo's background?\n> > >\n> > > One suggestion on the logo, tho, can you dim the light that's shining\n> > > on it? It sorta takes away from the rest of it. (Yes I know it's\n> > > a virtual light)\n> > >\n> >\n> > Here is an idea. I have thought about making the name stand out as a\n> > pattern within a grid/table.\n> >\n> > If you remember the end of the movie 2001, where HAL has those red bars\n> > coming out as he is disconnected. We could have PostgreSQL spelled out\n> > as red bard sticking out in a grid:\n> \n> Hmmm - you mean his crystal storage which Dave was pushing\n> and they smoothly came out - no?\n> \n> Nice idea.\n\nYes, that was my idea of somehow merging PostgreSQL into some table\nformat. Could look really cool, especially if viewed from an angle, or\nreally poor.\n\nWith someone who can do almost anything(you!), it becomes a harder\nquestion of what we want.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Fri, 19 Mar 1999 14:20:42 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "On Fri, 19 Mar 1999, Bruce Momjian wrote:\n\n> > Bruce Momjian wrote:\n> > \n> > >\n> > > > Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> > > > minutes to finish that one.\n> > > >\n> > > > Remember, it's a 100% virtual picture. Not one single detail\n> > > > is from a scanned photo or the like.\n> > > >\n> > >\n> > > You realize that now that we know you can do these images, a new\n> > > PostgreSQL logo request is coming! (Marc, ask him. Ask him!)\n> \n> I think I like the 3rd developers globe. Can we try it without\n> stickpins? I think it is looking confusing. Looks like there is enough\n> room in England on that image to show everyone as red dome, rather than\n> a stickpin.\n\n\tI agree about the 3rd globe, but disagree about the stick pins...I\nlike them, personally...reminds me about how one does military tactical\nmaps *grin*\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Fri, 19 Mar 1999 15:48:36 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> \n> On Fri, 19 Mar 1999, Bruce Momjian wrote:\n> \n> > > Bruce Momjian wrote:\n> > > \n> > > >\n> > > > > Those pictures are CPU killers. A 333 MHz Celeron needs 20\n> > > > > minutes to finish that one.\n> > > > >\n> > > > > Remember, it's a 100% virtual picture. Not one single detail\n> > > > > is from a scanned photo or the like.\n> > > > >\n> > > >\n> > > > You realize that now that we know you can do these images, a new\n> > > > PostgreSQL logo request is coming! (Marc, ask him. Ask him!)\n> > \n> > I think I like the 3rd developers globe. Can we try it without\n> > stickpins? I think it is looking confusing. Looks like there is enough\n> > room in England on that image to show everyone as red dome, rather than\n> > a stickpin.\n> \n> \tI agree about the 3rd globe, but disagree about the stick pins...I\n> like them, personally...reminds me about how one does military tactical\n> maps *grin*\n\n.PS\nAnd there isn't enough room in England. Andrew and Peter are that close to \neach other that they would have to share one. It's not fair if anyone else\nhas a stick for it's own.\n\nBut I'll try another type of stick.\n.PE\n\n\nJan\n\n-- \n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n", "msg_date": "Fri, 19 Mar 1999 20:54:58 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> And there isn't enough room in England. Andrew and Peter are that close to\n> each other that they would have to share one. It's not fair if anyone else\n> has a stick for it's own.\n>\n> But I'll try another type of stick.\n\n Hmmm - I'm absolutely unhappy with example #4. The conversion\n to JPG destroyed that one nearly. The rayshade output (517K)\n is absolutely nice. But it seems the drastic color changes\n from one pixel to the next aren't what JPG likes.\n\n I've noticed it earlier that all the images get a little\n dimmed when converted to JPG. They look brighter in the RLE\n format or when converted to PPM.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 19 Mar 1999 22:04:29 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "\nOn 19-Mar-99 Jan Wieck wrote:\n>> And there isn't enough room in England. Andrew and Peter are that close to\n>> each other that they would have to share one. It's not fair if anyone else\n>> has a stick for it's own.\n>>\n>> But I'll try another type of stick.\n> \n> Hmmm - I'm absolutely unhappy with example #4. The conversion\n> to JPG destroyed that one nearly. The rayshade output (517K)\n> is absolutely nice. But it seems the drastic color changes\n> from one pixel to the next aren't what JPG likes.\n> \n> I've noticed it earlier that all the images get a little\n> dimmed when converted to JPG. They look brighter in the RLE\n> format or when converted to PPM.\n\nI agree. Ick. Or were those pillars where stickpins used to be \nintentional? Either way, ick.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Fri, 19 Mar 1999 16:16:00 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> My first thought was a big truck on a road in the desert with\n> some cirrus clouds in the sky and \"PostgreSQL\" written on the\n> trailer. Maybe some snow capped mountains in the background.\n\nLet me give you an idea.\n\nInstead of a truck, a freight train.\n\nYou could have the engine with PostreSQL on the side, and the boxcars\ncould have PostgreSQL keywords on it, like SELECT, INSERT, UPDATE, RULE,\netc. The cars could trail off into the distance.\n\nOr you could picture a horseshoe curve, like we have in Pennsylvania. \nThe curve does a 180 degree turn, like an upside-down U, so you see more\nof the train in the picture, with the engine on the left, and caboose on\nthe right. (I can show you a picture if that helps.) You could put the\nname PostgreSQL on the mountain in the background of the curve too,\nagain with keywords on the boxcars.\n\nYou could even add a train to the bottom of the existing logo you have\nmade, with perhaps objects in the hopper cars. Would make a nice logo,\nand give it some movement. Smoke trailing out of the engine, etc.\n\nOf course, being from Pennsylvania, USA, I have to recommend a train.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Sat, 20 Mar 1999 08:26:29 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "Bruce Momjian wrote:\n\n>\n> > My first thought was a big truck on a road in the desert with\n> > some cirrus clouds in the sky and \"PostgreSQL\" written on the\n> > trailer. Maybe some snow capped mountains in the background.\n>\n> Let me give you an idea.\n>\n> Instead of a truck, a freight train.\n>\n> You could have the engine with PostreSQL on the side, and the boxcars\n> could have PostgreSQL keywords on it, like SELECT, INSERT, UPDATE, RULE,\n> etc. The cars could trail off into the distance.\n>\n> Or you could picture a horseshoe curve, like we have in Pennsylvania.\n> The curve does a 180 degree turn, like an upside-down U, so you see more\n> of the train in the picture, with the engine on the left, and caboose on\n> the right. (I can show you a picture if that helps.) You could put the\n> name PostgreSQL on the mountain in the background of the curve too,\n> again with keywords on the boxcars.\n>\n> You could even add a train to the bottom of the existing logo you have\n> made, with perhaps objects in the hopper cars. Would make a nice logo,\n> and give it some movement. Smoke trailing out of the engine, etc.\n>\n> Of course, being from Pennsylvania, USA, I have to recommend a train.\n\n Another nice idea - better than the truck.\n\n But (I known that many of my sentences begin with but) - but\n I thought about all that during the wieckend without creating\n pictures. And I came to the conclusion that we make a major\n mistake actually.\n\n We are talking about some nice image for the WEB banner, not\n about a LOGO.\n\n PostgreSQL doesn't have a logo. There is a wide range of\n kinds of logos out in the world. Some are simple graphics,\n some are self associative ways to instanciate something. Most\n car manufacturers use some simple graphic that could easily\n be made out of chrome (must be part of the concept) or with\n few colors:\n\n a lion (Peugeot)\n two upward pointing angles (Citroen)\n four circles (Audi)\n a circle in bavarian colors (BMW)\n a rhomb (Renault)\n a stylistic H (Honda)\n specially forms of stars (Chrysler, Mercedes)\n\n Most of them have interesting shading and reflectional\n effects when made out of chrome.\n\n One well known self associative thing is this: Write any\n combination of three capital letters in horizontal blue\n stripes. What a brand!\n\n Free software products also have logos already. Remember this\n coloured feather, that little cute penguin or the little\n devil.\n\n Except for the last two (Linux and BSD), all these logos can\n be easily drawn in any pixel editor and don't need a\n raytracer. If we come up with such a thing, it might be easy\n to create a raytraced WEB banner image where this logo stands\n out. But thinking of the image first is bottom up.\n\n Another advantage having such a logo would be, that the\n banner image could change from time to time without loosing\n anything. Only that the logo must be subject of all the\n images.\n\n I like the coloured feather of Apache, because it is simple\n and tells something about it. The message I see is \"free and\n light\".\n\n Sorry for taking us back to the start. But we need to make\n the first step first.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Mon, 22 Mar 1999 11:48:37 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "Marc G. Fournier wrote:\n\n> On Fri, 19 Mar 1999, Bruce Momjian wrote:\n> >\n> > I think I like the 3rd developers globe. Can we try it without\n> > stickpins? I think it is looking confusing. Looks like there is enough\n> > room in England on that image to show everyone as red dome, rather than\n> > a stickpin.\n>\n> I agree about the 3rd globe, but disagree about the stick pins...I\n> like them, personally...reminds me about how one does military tactical\n> maps *grin*\n\n I like them too :-). Maybe Bruce is right that they should\n look more like silver.\n\n I've changed them and placed a magnifying glass over europe\n as Example 5. Even if it hides some parts of the map, I\n think it's useful (not everyone has a real magnifying glass\n handy so a virtual one isn't a bad thing).\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Mon, 22 Mar 1999 14:35:31 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> Marc G. Fournier wrote:\n> \n> > On Fri, 19 Mar 1999, Bruce Momjian wrote:\n> > >\n> > > I think I like the 3rd developers globe. Can we try it without\n> > > stickpins? I think it is looking confusing. Looks like there is enough\n> > > room in England on that image to show everyone as red dome, rather than\n> > > a stickpin.\n> >\n> > I agree about the 3rd globe, but disagree about the stick pins...I\n> > like them, personally...reminds me about how one does military tactical\n> > maps *grin*\n> \n> I like them too :-). Maybe Bruce is right that they should\n> look more like silver.\n\nYes, better in silver.\n\n> \n> I've changed them and placed a magnifying glass over europe\n> as Example 5. Even if it hides some parts of the map, I\n> think it's useful (not everyone has a real magnifying glass\n> handy so a virtual one isn't a bad thing).\n\nWhy not set the image perspecive over Europe, that way all dots are\nvisible, and no need for a magnifying glass. The magnifying glass\nremoves the continuity of the image for me.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Mon, 22 Mar 1999 11:49:16 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "On Mon, 22 Mar 1999, Jan Wieck wrote:\n\n> PostgreSQL doesn't have a logo. There is a wide range of\n> kinds of logos out in the world. Some are simple graphics,\n> some are self associative ways to instanciate something. Most\n> car manufacturers use some simple graphic that could easily\n> be made out of chrome (must be part of the concept) or with\n> few colors:\n> \n> a lion (Peugeot)\n> two upward pointing angles (Citroen)\n> four circles (Audi)\n> a circle in bavarian colors (BMW)\n> a rhomb (Renault)\n> a stylistic H (Honda)\n> specially forms of stars (Chrysler, Mercedes)\n\nAgreed...Elephant still comes to mind...ability to carry heavy loads and\ngo the distance... :)\n\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 22 Mar 1999 13:11:13 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "On Mon, 22 Mar 1999, Bruce Momjian wrote:\n\n> > I've changed them and placed a magnifying glass over europe\n> > as Example 5. Even if it hides some parts of the map, I\n> > think it's useful (not everyone has a real magnifying glass\n> > handy so a virtual one isn't a bad thing).\n> \n> Why not set the image perspecive over Europe, that way all dots are\n> visible, and no need for a magnifying glass. The magnifying glass\n> removes the continuity of the image for me.\n\n\tAgreed on this point...some sort of floating magnifying glass\nwould be cool, that follows your mouse pointer, but, Ithink, unrealistic\n:(\n\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 22 Mar 1999 13:12:04 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "\nOn 22-Mar-99 The Hermit Hacker wrote:\n> On Mon, 22 Mar 1999, Jan Wieck wrote:\n> \n>> PostgreSQL doesn't have a logo. There is a wide range of\n>> kinds of logos out in the world. Some are simple graphics,\n>> some are self associative ways to instanciate something. Most\n>> car manufacturers use some simple graphic that could easily\n>> be made out of chrome (must be part of the concept) or with\n>> few colors:\n>> \n>> a lion (Peugeot)\n>> two upward pointing angles (Citroen)\n>> four circles (Audi)\n>> a circle in bavarian colors (BMW)\n>> a rhomb (Renault)\n>> a stylistic H (Honda)\n>> specially forms of stars (Chrysler, Mercedes)\n> \n> Agreed...Elephant still comes to mind...ability to carry heavy loads and\n> go the distance... :)\n\nnever forgets...\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Mon, 22 Mar 1999 12:13:52 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "Marc G. Fournier wrote:\n\n>\n> On Mon, 22 Mar 1999, Jan Wieck wrote:\n>\n> > PostgreSQL doesn't have a logo. There is a wide range of\n> > kinds of logos out in the world. Some are simple graphics,\n> > some are self associative ways to instanciate something. Most\n> > car manufacturers use some simple graphic that could easily\n> > be made out of chrome (must be part of the concept) or with\n> > few colors:\n> >\n> > a lion (Peugeot)\n> > two upward pointing angles (Citroen)\n> > four circles (Audi)\n> > a circle in bavarian colors (BMW)\n> > a rhomb (Renault)\n> > a stylistic H (Honda)\n> > specially forms of stars (Chrysler, Mercedes)\n>\n> Agreed...Elephant still comes to mind...ability to carry heavy loads and\n> go the distance... :)\n\n But is itself a bit fat - do we want that association? :-)\n\n The first I've created has a diamond. It's on my index.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Mon, 22 Mar 1999 18:14:46 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> \n> On Mon, 22 Mar 1999, Bruce Momjian wrote:\n> \n> > > I've changed them and placed a magnifying glass over europe\n> > > as Example 5. Even if it hides some parts of the map, I\n> > > think it's useful (not everyone has a real magnifying glass\n> > > handy so a virtual one isn't a bad thing).\n> > \n> > Why not set the image perspecive over Europe, that way all dots are\n> > visible, and no need for a magnifying glass. The magnifying glass\n> > removes the continuity of the image for me.\n> \n> \tAgreed on this point...some sort of floating magnifying glass\n> would be cool, that follows your mouse pointer, but, Ithink, unrealistic\n> :(\n\n Was only a test. Will remove it again.\n\n\nJan\n\n-- \n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n", "msg_date": "Mon, 22 Mar 1999 18:24:45 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "On Mon, 22 Mar 1999, Jan Wieck wrote:\n\n> Marc G. Fournier wrote:\n> \n> >\n> > On Mon, 22 Mar 1999, Jan Wieck wrote:\n> >\n> > > PostgreSQL doesn't have a logo. There is a wide range of\n> > > kinds of logos out in the world. Some are simple graphics,\n> > > some are self associative ways to instanciate something. Most\n> > > car manufacturers use some simple graphic that could easily\n> > > be made out of chrome (must be part of the concept) or with\n> > > few colors:\n> > >\n> > > a lion (Peugeot)\n> > > two upward pointing angles (Citroen)\n> > > four circles (Audi)\n> > > a circle in bavarian colors (BMW)\n> > > a rhomb (Renault)\n> > > a stylistic H (Honda)\n> > > specially forms of stars (Chrysler, Mercedes)\n> >\n> > Agreed...Elephant still comes to mind...ability to carry heavy loads and\n> > go the distance... :)\n> \n> But is itself a bit fat - do we want that association? :-)\n\nNot fat...heavy weight :)\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Mon, 22 Mar 1999 15:32:25 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "The Hermit Hacker wrote:\n> \n> On Mon, 22 Mar 1999, Jan Wieck wrote:\n> \n> > Marc G. Fournier wrote:\n> >\n> > >\n> > > On Mon, 22 Mar 1999, Jan Wieck wrote:\n> > >\n> > > > PostgreSQL doesn't have a logo. There is a wide range of\n> > > > kinds of logos out in the world. Some are simple graphics,\n> > > > some are self associative ways to instanciate something. Most\n> > > > car manufacturers use some simple graphic that could easily\n> > > > be made out of chrome (must be part of the concept) or with\n> > > > few colors:\n> > > >\n> > > > a lion (Peugeot)\n> > > > two upward pointing angles (Citroen)\n> > > > four circles (Audi)\n> > > > a circle in bavarian colors (BMW)\n> > > > a rhomb (Renault)\n> > > > a stylistic H (Honda)\n> > > > specially forms of stars (Chrysler, Mercedes)\n> > >\n> > > Agreed...Elephant still comes to mind...ability to carry heavy loads and\n> > > go the distance... :)\n\nAint elephant associated mostly with some political party in US ?\n\n> > But is itself a bit fat - do we want that association? :-)\n\nI once did some crocodiles, \n\nhttp://sid.trust.ee/Info/PostgreSQL.figs/logo/page2.html\n\nthe reasoning is at the beginning of:\n\nhttp://sid.trust.ee/Info/PostgreSQL.figs/logo/page.html\n\nThis discussion has escaped me until now as this went \nunder the heading Developers Globe ;)\n\n-------------------\nHannu\n", "msg_date": "Tue, 23 Mar 1999 00:26:35 +0200", "msg_from": "Hannu Krosing <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "[email protected] (Jan Wieck) writes:\n\n> But is itself a bit fat - do we want that association? :-)\n> \n> The first I've created has a diamond. It's on my index.\n\nI think this is the absolute best set of logo offerings yet.\n\nI like #2 best.\n\n#1 reminds me of the dopey I<heart>NY bumper stickers that were all\nthe rage some years back.\n\nIn #3, centering \"Powered by\" weakens the effect somehow.\n\nBut all three look very good.\n\n\nWould it look less cluttered without the ORDBMS?\n\n\nI've been speaking to one of the artists at work about logos. His\ndescription of a good logo includes the following:\n\n1. Should be simple enough that it's recognizable when scaled down to\n small icon size.\n\n2. Should be recognizable from the shape of its silhouette. We should\n know what it stands for even without standard colors and with\n extreme contrast.\n\n3. The shape should somehow suggest the entity that is represented.\n\n4. Does not necessarily include the company name. This can be attached\n beside the logo.\n\nHis examples of good logos: AT&T \"death star\", the Apple logo. I'd add\nthe FreeBSD logo as a third.\n\n", "msg_date": "23 Mar 1999 03:05:12 -0600", "msg_from": "Hal Snyder <[email protected]>", "msg_from_op": false, "msg_subject": "Dimaond Logo (was Globe...)" }, { "msg_contents": "Hal Snyder wrote:\n\n>\n> [email protected] (Jan Wieck) writes:\n>\n> > The first I've created has a diamond. It's on my index.\n>\n> I think this is the absolute best set of logo offerings yet.\n\n Thanks\n\n>\n> I like #2 best.\n\n And I #1 - maybe because I created that first.\n\n> In #3, centering \"Powered by\" weakens the effect somehow.\n\n If we choose the diamond as our logo, I'll create a bunch of\n permutations where \"Powered by\" and the diamond are left,\n center/as-o, right and we should vote on.\n\n> Would it look less cluttered without the ORDBMS?\n\n Don't know - would be part of the permutations. Would make\n it consume less space though :-).\n\n>\n> I've been speaking to one of the artists at work about logos. His\n> description of a good logo includes the following:\n>\n> 1. Should be simple enough that it's recognizable when scaled down to\n> small icon size.\n>\n> 2. Should be recognizable from the shape of its silhouette. We should\n> know what it stands for even without standard colors and with\n> extreme contrast.\n>\n> 3. The shape should somehow suggest the entity that is represented.\n>\n> 4. Does not necessarily include the company name. This can be attached\n> beside the logo.\n>\n> His examples of good logos: AT&T \"death star\", the Apple logo. I'd add\n> the FreeBSD logo as a third.\n\n Rules 1 and 2 are matched. Rule 3 - a diamond is hard, sharp,\n clear bright and more worthy than gold - matched.\n\n Rule number 4 I would not want to match. The spelling\n PostgreSQL (not Postgres-SQL or all the other spellings\n sometimes used) is somewhat special and thus should be part\n of the logo.\n\n My absolute favorites of logos are the ones where the\n font/colors+outline associate with the name. Write anything\n in that special curly way white on a red background and you\n have Coke. Any 3 capitals written in blue stripes reminds on\n IBM. And my wife bought NIVEA in Tunis - it was just that\n blue round cover with some arabic letters in white :-) - you\n can read that even if you can't identify one single letter.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Tue, 23 Mar 1999 11:59:23 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Dimaond Logo (was Globe...)" }, { "msg_contents": "On Tue, 23 Mar 1999, Jan Wieck wrote:\n\n> My absolute favorites of logos are the ones where the\n> font/colors+outline associate with the name. Write anything\n> in that special curly way white on a red background and you\n> have Coke. Any 3 capitals written in blue stripes reminds on\n> IBM. And my wife bought NIVEA in Tunis - it was just that\n> blue round cover with some arabic letters in white :-) - you\n> can read that even if you can't identify one single letter.\n\nSpeaking of color, how about a different color for \"Postgre\"? That\none's a bit boring.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Tue, 23 Mar 1999 06:23:13 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Dimaond Logo (was Globe...)" }, { "msg_contents": ">\n> On Tue, 23 Mar 1999, Jan Wieck wrote:\n>\n> > My absolute favorites of logos are the ones where the\n> > font/colors+outline associate with the name. Write anything\n> > in that special curly way white on a red background and you\n> > have Coke. Any 3 capitals written in blue stripes reminds on\n> > IBM. And my wife bought NIVEA in Tunis - it was just that\n> > blue round cover with some arabic letters in white :-) - you\n> > can read that even if you can't identify one single letter.\n>\n> Speaking of color, how about a different color for \"Postgre\"? That\n> one's a bit boring.\n\n Tell RGB...\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Tue, 23 Mar 1999 12:25:09 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Dimaond Logo (was Globe...)" }, { "msg_contents": "On Tue, 23 Mar 1999, Jan Wieck wrote:\n\n> >\n> > On Tue, 23 Mar 1999, Jan Wieck wrote:\n> >\n> > > My absolute favorites of logos are the ones where the\n> > > font/colors+outline associate with the name. Write anything\n> > > in that special curly way white on a red background and you\n> > > have Coke. Any 3 capitals written in blue stripes reminds on\n> > > IBM. And my wife bought NIVEA in Tunis - it was just that\n> > > blue round cover with some arabic letters in white :-) - you\n> > > can read that even if you can't identify one single letter.\n> >\n> > Speaking of color, how about a different color for \"Postgre\"? That\n> > one's a bit boring.\n> \n> Tell RGB...\n\nDunno. That's why I didn't suggest one. What works with blue and gray?\nIf I can find it I'll dig out the color wheel and take a look.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Tue, 23 Mar 1999 06:31:16 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Dimaond Logo (was Globe...)" }, { "msg_contents": "On 23 Mar 1999, Hal Snyder wrote:\n\n> [email protected] (Jan Wieck) writes:\n> \n> > But is itself a bit fat - do we want that association? :-)\n> > \n> > The first I've created has a diamond. It's on my index.\n> \n> I think this is the absolute best set of logo offerings yet.\n> \n> I like #2 best.\n> \n> #1 reminds me of the dopey I<heart>NY bumper stickers that were all\n> the rage some years back.\n\nGeez, and I liked that one best :(\n\n\n", "msg_date": "Tue, 23 Mar 1999 10:09:54 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Dimaond Logo (was Globe...)" }, { "msg_contents": "Hannu Krosing wrote:\n> Marc G. Fournier wrote:\n> > On Mon, 22 Mar 1999, Jan Wieck wrote:\n> > > PostgreSQL doesn't have a logo. \n> > Agreed...Elephant still comes to mind...ability to \n> > carry heavy loads and go the distance... :)\n\nI really like the Elephant idea. \n\n> Aint elephant associated mostly with some political party in US ?\n\nSure, but the first thing which came to my mind was Dumbo \nthe flying elephant and the Jungle book's kernel Harting.\n\n> I once did some crocodiles,\n> http://sid.trust.ee/Info/PostgreSQL.figs/logo/page.html\n \nThe Crocodile is cool, but I'd vote for an Elephant, \nor perhaps a bear.\n\n:) Clark\n", "msg_date": "Tue, 23 Mar 1999 22:58:03 +0000", "msg_from": "Clark Evans <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "> Hannu Krosing wrote:\n> > Marc G. Fournier wrote:\n> > > On Mon, 22 Mar 1999, Jan Wieck wrote:\n> > > > PostgreSQL doesn't have a logo. \n> > > Agreed...Elephant still comes to mind...ability to \n> > > carry heavy loads and go the distance... :)\n> \n> I really like the Elephant idea. \n> \n> > Aint elephant associated mostly with some political party in US ?\n> \n> Sure, but the first thing which came to my mind was Dumbo \n> the flying elephant and the Jungle book's kernel Harting.\n> \n> > I once did some crocodiles,\n> > http://sid.trust.ee/Info/PostgreSQL.figs/logo/page.html\n> \n> The Crocodile is cool, but I'd vote for an Elephant, \n> or perhaps a bear.\n\nElephant with a diamond on his back, or a train pulling a diamond?\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 23 Mar 1999 18:03:33 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": ">\n> Hannu Krosing wrote:\n> > Marc G. Fournier wrote:\n> > > On Mon, 22 Mar 1999, Jan Wieck wrote:\n> > > > PostgreSQL doesn't have a logo.\n> > > Agreed...Elephant still comes to mind...ability to\n> > > carry heavy loads and go the distance... :)\n>\n> I really like the Elephant idea.\n>\n> > Aint elephant associated mostly with some political party in US ?\n>\n> Sure, but the first thing which came to my mind was Dumbo\n> the flying elephant and the Jungle book's kernel Harting.\n\n :-)\n\n>\n> > I once did some crocodiles,\n> > http://sid.trust.ee/Info/PostgreSQL.figs/logo/page.html\n>\n> The Crocodile is cool, but I'd vote for an Elephant,\n> or perhaps a bear.\n\n I've just finished the first raytraces of the diamond logo\n and the according WEB banner.\n\n http://www.postgresql.org/~wieck\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 24 Mar 1999 00:08:30 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": ">\n> Elephant with a diamond on his back, or a train pulling a diamond?\n>\n\n Such a scene would look nice as a WEB banner for some page.\n But not for a \"Powered by\" logo. When sized to something\n below 200x100 you would only see some dust, not a train\n anymore. Or you can identify the elephant but the diamond\n shrinks down to a dot.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 24 Mar 1999 00:11:17 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "Bruce Momjian wrote:\n> Elephant with a diamond on his back, or a train pulling a diamond?\n\nHow about a dimond with an elephant in the middle.\nKinda like the pink panther.\n\n:) Clark\n", "msg_date": "Tue, 23 Mar 1999 23:19:51 +0000", "msg_from": "Clark Evans <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: Developers Globe (FINAL)" }, { "msg_contents": "\nOn 23-Mar-99 Jan Wieck wrote:\n>>\n>> Hannu Krosing wrote:\n>> > Marc G. Fournier wrote:\n>> > > On Mon, 22 Mar 1999, Jan Wieck wrote:\n>> > > > PostgreSQL doesn't have a logo.\n>> > > Agreed...Elephant still comes to mind...ability to\n>> > > carry heavy loads and go the distance... :)\n>>\n>> I really like the Elephant idea.\n>>\n>> > Aint elephant associated mostly with some political party in US ?\n>>\n>> Sure, but the first thing which came to my mind was Dumbo\n>> the flying elephant and the Jungle book's kernel Harting.\n> \n> :-)\n> \n>>\n>> > I once did some crocodiles,\n>> > http://sid.trust.ee/Info/PostgreSQL.figs/logo/page.html\n>>\n>> The Crocodile is cool, but I'd vote for an Elephant,\n>> or perhaps a bear.\n> \n> I've just finished the first raytraces of the diamond logo\n> and the according WEB banner.\n> \n> http://www.postgresql.org/~wieck\n\nWhere's the elephant? :)\n\nLooks good, but the diamond needs the benefit of a star filter on the \ncamera (so to speak). If you need I'll try to explain it in a bit more\ndetail.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Tue, 23 Mar 1999 18:23:58 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": ">\n> Bruce Momjian wrote:\n> > Elephant with a diamond on his back, or a train pulling a diamond?\n>\n> How about a dimond with an elephant in the middle.\n> Kinda like the pink panther.\n\n Not easy to design an elephant in the geometric figure world\n of a raytracer. But I like the idea of putting the elephant\n inside the diamond.\n\n Hmmm - many diamonds have some (dunno how it's called really)\n faults. The one of our diamond could have a shape like an\n elephant.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 24 Mar 1999 00:38:05 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": ">\n> Where's the elephant? :)\n\n Still in mind and the last idea of Clark was totally nice -\n putting the elephant into the diamond. Need's a complete\n crystal diamond (that will eat CPU - man).\n\n>\n> Looks good, but the diamond needs the benefit of a star filter on the\n> camera (so to speak). If you need I'll try to explain it in a bit more\n> detail.\n\n Outch - yes I know what you mean - I have such a one for my\n camera too (along with the other usual ones like the 2 diop\n hole lense etc.). That's one of the limits of raytracing\n AFAIK. I never had any success on such a thing like a filter.\n Also it's impossible to build spectral effects. Spectral\n effects are based on the fact that the refractional index of\n a material varies on the frequency of the light (that's what\n makes the fire in diamonds). But a raytracer cannot do this -\n or at least the trace would run infinitely long - thus no\n tracer I know of supports it.\n\n But on the star filter I have already an idea. Will try out\n some things.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 24 Mar 1999 00:50:10 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "On Tue, 23 Mar 1999, you wrote:\n\n> Outch - yes I know what you mean - I have such a one for my\n> camera too (along with the other usual ones like the 2 diop\n> hole lense etc.). That's one of the limits of raytracing\n> AFAIK. I never had any success on such a thing like a filter.\n> Also it's impossible to build spectral effects. Spectral\n> effects are based on the fact that the refractional index of\n> a material varies on the frequency of the light (that's what\n> makes the fire in diamonds). But a raytracer cannot do this -\n> or at least the trace would run infinitely long - thus no\n> tracer I know of supports it.\n\nDoesn't blender do spectral effects?\n\nTaral\n", "msg_date": "Tue, 23 Mar 1999 18:01:31 -0600", "msg_from": "Taral <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "> >\n> > Elephant with a diamond on his back, or a train pulling a diamond?\n> >\n> \n> Such a scene would look nice as a WEB banner for some page.\n> But not for a \"Powered by\" logo. When sized to something\n> below 200x100 you would only see some dust, not a train\n> anymore. Or you can identify the elephant but the diamond\n> shrinks down to a dot.\n> \n\nYes. Absolutely.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 23 Mar 1999 22:53:59 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "\nI kinda liked the ORDBMS tack'd onto the bottom...\n\n\nOn Wed, 24 Mar 1999, Jan Wieck wrote:\n\n> >\n> > Hannu Krosing wrote:\n> > > Marc G. Fournier wrote:\n> > > > On Mon, 22 Mar 1999, Jan Wieck wrote:\n> > > > > PostgreSQL doesn't have a logo.\n> > > > Agreed...Elephant still comes to mind...ability to\n> > > > carry heavy loads and go the distance... :)\n> >\n> > I really like the Elephant idea.\n> >\n> > > Aint elephant associated mostly with some political party in US ?\n> >\n> > Sure, but the first thing which came to my mind was Dumbo\n> > the flying elephant and the Jungle book's kernel Harting.\n> \n> :-)\n> \n> >\n> > > I once did some crocodiles,\n> > > http://sid.trust.ee/Info/PostgreSQL.figs/logo/page.html\n> >\n> > The Crocodile is cool, but I'd vote for an Elephant,\n> > or perhaps a bear.\n> \n> I've just finished the first raytraces of the diamond logo\n> and the according WEB banner.\n> \n> http://www.postgresql.org/~wieck\n> \n> \n> Jan\n> \n> --\n> \n> #======================================================================#\n> # It's easier to get forgiveness for being wrong than for being right. #\n> # Let's break this rule - forgive me. #\n> #======================================== [email protected] (Jan Wieck) #\n> \n> \n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Wed, 24 Mar 1999 01:08:55 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": ">\n>\n> I kinda liked the ORDBMS tack'd onto the bottom...\n\n Someone said it makes it confuse. So I tried without.\n\n> > > The Crocodile is cool, but I'd vote for an Elephant,\n> > > or perhaps a bear.\n\n Are we still talking about a PG LOGO or do we plan to open a\n zoo :-)\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 24 Mar 1999 17:02:50 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "I still like #2 in the set of three the best. The 3D highlighting done\nto the letters in the top set of logos makes them harder to read on my\nmonitor than the flat letters. And I can't tell what that blue thingie\nis inside the diamond.\n\nMaybe I agree with Vince that the straw color of the letters isn't the\ngreatest. We can always change the color scheme of the web pages to\nmatch a new logo if we ever agree on the perfect logo.\n\nI was one a complainer about the ORDBMS fine print. Trying to keep it\nsimple. (Also preferred the diamond outside the word, *still* feel a\nshape in place of a letter is a cliche but guess I'm outnumbered on\nthat point.)\n\nJan, thanks for putting up with this silliness.\n\[email protected] (Jan Wieck) writes:\n\n> >\n> >\n> > I kinda liked the ORDBMS tack'd onto the bottom...\n> \n> Someone said it makes it confuse. So I tried without.\n> \n> > > > The Crocodile is cool, but I'd vote for an Elephant,\n> > > > or perhaps a bear.\n> \n> Are we still talking about a PG LOGO or do we plan to open a\n> zoo :-)\n\n", "msg_date": "25 Mar 1999 02:22:23 -0600", "msg_from": "Hal Snyder <[email protected]>", "msg_from_op": false, "msg_subject": "Re: PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "Marc G. Fournier wrote:\n\n> I kinda liked the ORDBMS tack'd onto the bottom...\n>\n>\n\n Reload!\n\n You wanted an elephant - hopefully that's elephant enough\n (it's one of these majestic, african bulls). I like jewels,\n so that's still there :-)\n\n Haven't created the Powerd by ones now, because I think\n Powered by isn't the right wording for an elephant logo. What\n about \"Carried by\"? I Think for the small ones, the\n \"PostgreSQL\" might move down to make the room for the\n \"Carried by\" above. In that case, there's no room left for\n the ORDBMS, so I left that out - sorry.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Thu, 25 Mar 1999 20:05:50 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "Jan Wieck wrote:\n> You wanted an elephant - hopefully that's elephant enough\n> (it's one of these majestic, african bulls). I like jewels,\n> so that's still there :-)\n\nWay Cool. (watch out... I'm starting to expect\nmagic from you *smile* )\n\nIt seems that the elephant is 'on top of' the diamond, I\nwas wondering if it's possible to make it seem like the\nelephant is 'inside' of the jewel? \n\nThe opening credits of the pink panther have a \ncloudy jewel where the pink panther sorta emerges\nfrom the cloudy haze inside of the jewel. I guess\nthis is what I had dreamed.\n\nI really like the idea of an elephant, but perhaps\nit could be made less lifelike, more cartoonish.\nHmm. That's just a thought.\n\nWow. You must have spent forever getting it\nto look like that. \n\n:) Clark\n", "msg_date": "Thu, 25 Mar 1999 20:14:08 +0000", "msg_from": "Clark Evans <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "\nOn 25-Mar-99 Jan Wieck wrote:\n> Marc G. Fournier wrote:\n> \n>> I kinda liked the ORDBMS tack'd onto the bottom...\n>>\n>>\n> \n> Reload!\n> \n> You wanted an elephant - hopefully that's elephant enough\n> (it's one of these majestic, african bulls). I like jewels,\n> so that's still there :-)\n> \n> Haven't created the Powerd by ones now, because I think\n> Powered by isn't the right wording for an elephant logo. What\n> about \"Carried by\"? I Think for the small ones, the\n> \"PostgreSQL\" might move down to make the room for the\n> \"Carried by\" above. In that case, there's no room left for\n> the ORDBMS, so I left that out - sorry.\n\nThere ya go. Any luck with the star filter?\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Thu, 25 Mar 1999 15:20:54 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "> Wow. You must have spent forever getting it\n> to look like that.\n\n 1.5 hours including all the test traces :-)\n\n That's why I still use that old, outdated rayshade - I'm\n soooo familiar with it. And it's happy with a vi(1) as scene\n editor as long as you can keep enough objects in a virtual\n room in mind.\n\n Well, I can higher the reflections of the peaces of the\n jewel. Then it would look more like he's inside. But I have\n to do it in this fashion when running for the little \"Carried\n by\" images because you wouldn't see the elephant any more.\n All that is easy done, rayshade's default preprocessor is\n cpp, So it's simply one difference in the commandline options\n and some #ifdef's inside the scene description. And I don't\n run it from the commandline - I use make.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Thu, 25 Mar 1999 23:51:40 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "> There ya go. Any luck with the star filter?\n> \n> Vince.\n\n Not yet - still trying things...\n\n\nJan\n\n-- \n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n", "msg_date": "Thu, 25 Mar 1999 23:52:58 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "On Thu, 25 Mar 1999, Jan Wieck wrote:\n\n> Marc G. Fournier wrote:\n> \n> > I kinda liked the ORDBMS tack'd onto the bottom...\n> >\n> >\n> \n> Reload!\n> \n> You wanted an elephant - hopefully that's elephant enough\n> (it's one of these majestic, african bulls). I like jewels,\n> so that's still there :-)\n\nSold!! You've hit it soooooo unbelievably perfectly :)\n\nCan you make the Diamond/Elephant a seperate image easily? It makes a\n*perfect* stand alone logo... :)\n\n> Haven't created the Powerd by ones now, because I think\n> Powered by isn't the right wording for an elephant logo. What\n> about \"Carried by\"? I Think for the small ones, the\n> \"PostgreSQL\" might move down to make the room for the\n> \"Carried by\" above. In that case, there's no room left for\n> the ORDBMS, so I left that out - sorry.\n\nS'alright, I think I can forgive that :) \n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 25 Mar 1999 19:21:22 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "On Thu, 25 Mar 1999, Clark Evans wrote:\n\n> Jan Wieck wrote:\n> > You wanted an elephant - hopefully that's elephant enough\n> > (it's one of these majestic, african bulls). I like jewels,\n> > so that's still there :-)\n> \n> Way Cool. (watch out... I'm starting to expect\n> magic from you *smile* )\n> \n> It seems that the elephant is 'on top of' the diamond, I\n> was wondering if it's possible to make it seem like the\n> elephant is 'inside' of the jewel? \n> \n> The opening credits of the pink panther have a \n> cloudy jewel where the pink panther sorta emerges\n> from the cloudy haze inside of the jewel. I guess\n> this is what I had dreamed.\n> \n> I really like the idea of an elephant, but perhaps\n> it could be made less lifelike, more cartoonish.\n> Hmm. That's just a thought.\n\nNo no...perfect just as it is!! :)\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 25 Mar 1999 19:22:10 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "Marc G. Fournier wrote:\n\n> Can you make the Diamond/Elephant a seperate image easily? It makes a\n> *perfect* stand alone logo... :)\n\n /* the characters */ and change the viewpoint - that's all.\n Maybe I have to change some light positions too to get the\n same reflections from the other viewpoint, but that's not\n that hard to do.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 26 Mar 1999 00:58:03 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "> > Way Cool. (watch out... I'm starting to expect\n> > magic from you *smile* )\n\n After all it's still just a bunch of bit's - no magic :-)\n\n> >\n> > It seems that the elephant is 'on top of' the diamond, I\n> > was wondering if it's possible to make it seem like the\n> > elephant is 'inside' of the jewel?\n\n Done. But personally I prefer the first one.\n\n What I try to do is to have the jewel and the elephant\n wheighted equally - mean that noone has the majority of the\n entire symbol. I can surely cover it up to invisibility in\n the jewel, but that's not the target.\n\n> >\n> > The opening credits of the pink panther have a\n> > cloudy jewel where the pink panther sorta emerges\n> > from the cloudy haze inside of the jewel. I guess\n> > this is what I had dreamed.\n\n And I understood it that way on first call. But that has some\n problems. In the case of the pink panther it's a movie, and\n if I remember right the perspective is rotating slowly around\n it. If you have a crystal material with a refractional index,\n a modelled elephant inside would only become visible if\n viewed nearly from the top. Especially in the case of the\n wanted little powered by logos, because the elephant must be\n big enough to not to shrink down to 3-4 pixels. But then you\n don't identify the entire thing any longer as a jewel, it's\n just some round crystal, nothing else.\n\n> >\n> > I really like the idea of an elephant, but perhaps\n> > it could be made less lifelike, more cartoonish.\n> > Hmm. That's just a thought.\n>\n> No no...perfect just as it is!! :)\n\n PostgreSQL isn't a joke. It's IMHO closer to the one I choose\n than to flying DUMBO or any other silly one in comic strips.\n\n Never underestimate the power of such a majestic creature! So\n never let it look cute and harmless.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 26 Mar 1999 02:10:37 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "> On Thu, 25 Mar 1999, Jan Wieck wrote:\n> \n> > Marc G. Fournier wrote:\n> > \n> > > I kinda liked the ORDBMS tack'd onto the bottom...\n> > >\n> > >\n> > \n> > Reload!\n> > \n> > You wanted an elephant - hopefully that's elephant enough\n> > (it's one of these majestic, african bulls). I like jewels,\n> > so that's still there :-)\n> \n> Sold!! You've hit it soooooo unbelievably perfectly :)\n> \n\nI agree also.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Thu, 25 Mar 1999 20:17:00 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "\nOn 26-Mar-99 Jan Wieck wrote:\n\n> What I try to do is to have the jewel and the elephant\n> wheighted equally - mean that noone has the majority of the\n> entire symbol. I can surely cover it up to invisibility in\n> the jewel, but that's not the target.\n\nCan you make a stand alone diamond with the elephant where the elephant\ndoesn't look like he has jauntis? IOW, more like his natural color.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Thu, 25 Mar 1999 20:32:16 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "On Thu, 25 Mar 1999, you wrote:\n>On 26-Mar-99 Jan Wieck wrote:\n>\n>> What I try to do is to have the jewel and the elephant\n>> wheighted equally - mean that noone has the majority of the\n>> entire symbol. I can surely cover it up to invisibility in\n>> the jewel, but that's not the target.\n>\n>Can you make a stand alone diamond with the elephant where the elephant\n>doesn't look like he has jauntis? IOW, more like his natural color.\n\nWhere ARE these pictures? I missed a big chunk of list traffic...\n\nTaral\n", "msg_date": "Thu, 25 Mar 1999 19:55:59 -0600", "msg_from": "Taral <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "\nOn 26-Mar-99 Taral wrote:\n> On Thu, 25 Mar 1999, you wrote:\n>>On 26-Mar-99 Jan Wieck wrote:\n>>\n>>> What I try to do is to have the jewel and the elephant\n>>> wheighted equally - mean that noone has the majority of the\n>>> entire symbol. I can surely cover it up to invisibility in\n>>> the jewel, but that's not the target.\n>>\n>>Can you make a stand alone diamond with the elephant where the elephant\n>>doesn't look like he has jauntis? IOW, more like his natural color.\n> \n> Where ARE these pictures? I missed a big chunk of list traffic...\n\nhttp://www.postgresql.org/~wieck\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Thu, 25 Mar 1999 21:06:26 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "On Thu, 25 Mar 1999, you wrote:\n>http://www.postgresql.org/~wieck\n\nOh, well... put my opinion in. I like:\n\n* the first big logo [pg_banner3.jpg]\n* the standalone, but can it be half-size?\n* Map #5, but with the pov of the first two. Maybe the magnifying glass can be\ntipped?\n\nTaral\n", "msg_date": "Thu, 25 Mar 1999 20:17:21 -0600", "msg_from": "Taral <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "On Fri, 26 Mar 1999, Jan Wieck wrote:\n\n> > > Way Cool. (watch out... I'm starting to expect\n> > > magic from you *smile* )\n> \n> After all it's still just a bunch of bit's - no magic :-)\n> \n> > >\n> > > It seems that the elephant is 'on top of' the diamond, I\n> > > was wondering if it's possible to make it seem like the\n> > > elephant is 'inside' of the jewel?\n> \n> Done. But personally I prefer the first one.\n\nI prefer the original one too...the elephant's \"color\" looks more healthy\n:) Can we say \"jaundice elephant\"? :)\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Thu, 25 Mar 1999 22:31:01 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "Marc G. Fournier wrote:\n\n> I prefer the original one too...the elephant's \"color\" looks more healthy\n> :) Can we say \"jaundice elephant\"? :)\n\n I've removed that jaundice one and slightly polished up the\n jewel. There's now a \"Carried by\" logo too. And I've added\n another idea to the whole thing - just take a look.\n\n If we ever choose one of them to be it, I would like to add\n the raytracing sources and the image it uses along with a\n description which tools are required and how to build it to\n the www module of CVS. I think a new directory ./www/logo\n would be sufficient.\n\n Not that I think many of us could actually do anything with\n them, but just for the case of some accident, another one has\n a chance to take over.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 26 Mar 1999 17:17:16 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "> Marc G. Fournier wrote:\n> \n> > I prefer the original one too...the elephant's \"color\" looks more healthy\n> > :) Can we say \"jaundice elephant\"? :)\n> \n> I've removed that jaundice one and slightly polished up the\n> jewel. There's now a \"Carried by\" logo too. And I've added\n> another idea to the whole thing - just take a look.\n> \n> If we ever choose one of them to be it, I would like to add\n> the raytracing sources and the image it uses along with a\n> description which tools are required and how to build it to\n> the www module of CVS. I think a new directory ./www/logo\n> would be sufficient.\n> \n> Not that I think many of us could actually do anything with\n> them, but just for the case of some accident, another one has\n> a chance to take over.\n\n\nI like it. Can we add color back into the elephant in the jewel?\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Fri, 26 Mar 1999 11:59:01 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "Bruce Momjian wrote:\n\n> > I've removed that jaundice one and slightly polished up the\n> > jewel. There's now a \"Carried by\" logo too. And I've added\n> > another idea to the whole thing - just take a look.\n>\n> I like it. Can we add color back into the elephant in the jewel?\n\n No. The golden one (actually it was more yellow) looked\n really sick, and it's contrast to the jewel was too low for\n the \"Carried by\" size.\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 26 Mar 1999 18:16:25 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "Thus spake Jan Wieck\n> I've removed that jaundice one and slightly polished up the\n> jewel. There's now a \"Carried by\" logo too. And I've added\n> another idea to the whole thing - just take a look.\n\nI know that \"Powered by\" isn't quite right but \"Carried by\" doesn't\nseem right either. I understand the elephant reference in \"Carried\nby\" but I bet most people won't get it right off. How does this\none sound? \"Empowered by.\"\n\n-- \nD'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\nhttp://www.druid.net/darcy/ | and a sheep voting on\n+1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.\n", "msg_date": "Fri, 26 Mar 1999 12:45:25 -0500 (EST)", "msg_from": "\"D'Arcy\" \"J.M.\" Cain <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "On Fri, 26 Mar 1999, D'Arcy J.M. Cain wrote:\n\n> Thus spake Jan Wieck\n> > I've removed that jaundice one and slightly polished up the\n> > jewel. There's now a \"Carried by\" logo too. And I've added\n> > another idea to the whole thing - just take a look.\n> \n> I know that \"Powered by\" isn't quite right but \"Carried by\" doesn't\n> seem right either. I understand the elephant reference in \"Carried\n> by\" but I bet most people won't get it right off. How does this\n> one sound? \"Empowered by.\"\n\n*groan* That word and it's derivitives are insanely overused.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Fri, 26 Mar 1999 12:55:04 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "D'Arcy J.M. Cain wrote:\n\n> Thus spake Jan Wieck\n> > I've removed that jaundice one and slightly polished up the\n> > jewel. There's now a \"Carried by\" logo too. And I've added\n> > another idea to the whole thing - just take a look.\n>\n> I know that \"Powered by\" isn't quite right but \"Carried by\" doesn't\n> seem right either. I understand the elephant reference in \"Carried\n> by\" but I bet most people won't get it right off. How does this\n> one sound? \"Empowered by.\"\n\n In the context of computers (and thus any WEB server), the\n word \"Power\" only reminds me to something measured in Volts\n and Amperes. So I just wanted to get rid of it.\n\n I really liked the \"PostgreSQL underneath\" for the\n crocodiles. But that doesn't fit for an elephant.\n\n What about these:\n\n PostgreSQL\n in background\n\n PostgreSQL\n accelerated\n\n Supported by\n PostgreSQL\n\n PostgreSQL\n driven\n\n Or just without any such wording - why not?\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Fri, 26 Mar 1999 18:59:34 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "\nPerfect...I like the 'sample home page' you built also...it works for me\nwithout a doubt :)\n\n\nOn Fri, 26 Mar 1999, Jan Wieck wrote:\n\n> Marc G. Fournier wrote:\n> \n> > I prefer the original one too...the elephant's \"color\" looks more healthy\n> > :) Can we say \"jaundice elephant\"? :)\n> \n> I've removed that jaundice one and slightly polished up the\n> jewel. There's now a \"Carried by\" logo too. And I've added\n> another idea to the whole thing - just take a look.\n> \n> If we ever choose one of them to be it, I would like to add\n> the raytracing sources and the image it uses along with a\n> description which tools are required and how to build it to\n> the www module of CVS. I think a new directory ./www/logo\n> would be sufficient.\n> \n> Not that I think many of us could actually do anything with\n> them, but just for the case of some accident, another one has\n> a chance to take over.\n> \n> \n> Jan\n> \n> --\n> \n> #======================================================================#\n> # It's easier to get forgiveness for being wrong than for being right. #\n> # Let's break this rule - forgive me. #\n> #======================================== [email protected] (Jan Wieck) #\n> \n> \n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Fri, 26 Mar 1999 14:36:32 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "On Fri, 26 Mar 1999, Jan Wieck wrote:\n\n> D'Arcy J.M. Cain wrote:\n> \n> > Thus spake Jan Wieck\n> > > I've removed that jaundice one and slightly polished up the\n> > > jewel. There's now a \"Carried by\" logo too. And I've added\n> > > another idea to the whole thing - just take a look.\n> >\n> > I know that \"Powered by\" isn't quite right but \"Carried by\" doesn't\n> > seem right either. I understand the elephant reference in \"Carried\n> > by\" but I bet most people won't get it right off. How does this\n> > one sound? \"Empowered by.\"\n> \n> In the context of computers (and thus any WEB server), the\n> word \"Power\" only reminds me to something measured in Volts\n> and Amperes. So I just wanted to get rid of it.\n> \n> I really liked the \"PostgreSQL underneath\" for the\n> crocodiles. But that doesn't fit for an elephant.\n> \n> What about these:\n> \n> PostgreSQL\n> in background\n> \n> PostgreSQL\n> accelerated\n> \n> Supported by\n> PostgreSQL\n> \n> PostgreSQL\n> driven\n\nI'd go more:\n\n\tdriven by\n\tPostgreSQL\n\n\n> \n> Or just without any such wording - why not?\n> \n> \n> Jan\n> \n> --\n> \n> #======================================================================#\n> # It's easier to get forgiveness for being wrong than for being right. #\n> # Let's break this rule - forgive me. #\n> #======================================== [email protected] (Jan Wieck) #\n> \n> \n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Fri, 26 Mar 1999 14:37:16 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": ">> What about these:\n\n PostgreSQL\n\t used here!\n\n\n", "msg_date": "Fri, 26 Mar 1999 19:59:36 +0100 (CET)", "msg_from": "Kaare Rasmussen <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "> Bruce Momjian wrote:\n> \n> > > I've removed that jaundice one and slightly polished up the\n> > > jewel. There's now a \"Carried by\" logo too. And I've added\n> > > another idea to the whole thing - just take a look.\n> >\n> > I like it. Can we add color back into the elephant in the jewel?\n> \n> No. The golden one (actually it was more yellow) looked\n> really sick, and it's contrast to the jewel was too low for\n> the \"Carried by\" size.\n\nActually, I liked the naturally-colored elephant.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Fri, 26 Mar 1999 14:00:08 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "On Fri, 26 Mar 1999, Kaare Rasmussen wrote:\n\n> >> What about these:\n> \n> PostgreSQL\n> \t used here!\n\nthat one works too...\n\nMarc G. Fournier ICQ#7615664 IRC Nick: Scrappy\nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Fri, 26 Mar 1999 15:48:36 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" }, { "msg_contents": "Kaare Rasmussen <[email protected]> writes:\n\n> >> What about these:\n> \n> PostgreSQL\n> \t used here!\n\nGets my vote.\n", "msg_date": "27 Mar 1999 13:26:59 -0600", "msg_from": "Hal Snyder <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] PostgreSQL LOGO (was: Developers Globe (FINAL))" } ]
[ { "msg_contents": "\n============================================================================\n POSTGRESQL BUG REPORT TEMPLATE\n============================================================================\n\n\nYour name\t\t: Theo Kramer\nYour email address\t: [email protected]\n\nCategory\t\t: runtime: back-end\nSeverity\t\t: serious\n\nSummary: psql commands\n\nSystem Configuration\n--------------------\n Operating System : Redhat 5.0 Kernel 2.0.35\n\n PostgreSQL version : 6.5\n\n Compiler used : gcc 2.7.2.3-8\n\nHardware:\n---------\nPentium PRO 200 MHz 96 MByte Ram\n\nVersions of other tools:\n------------------------\nmake-3.76.1-2 \n\n--------------------------------------------------------------------------\nProblem Description:\n--------------------\nCommands in psql that return nothing cause backend to drop.\n\n--------------------------------------------------------------------------\n\nTest Case:\n----------\n\nThe following illustrates the problem. Note that this is when no tables\nhave been created. Creating a table (no indexes) then \\d works fine,\nhowever, a command like \\di then again causes the backend to drop.\n\n\n[postgres@flame postgres]$ psql template1\nWelcome to the POSTGRESQL interactive sql monitor:\n Please read the file COPYRIGHT for copyright terms of POSTGRESQL\n\n type \\? for help on slash commands\n type \\q to quit\n type \\g or terminate with semicolon to execute query\n You are currently connected to the database: template1\n\ntemplate1=> \\d\npqReadData() -- backend closed the channel unexpectedly.\n This probably means the backend terminated abnormally\n before or while processing the request.\ntemplate1=> \\q \n[postgres@flame postgres]$ \n\n--------------------------------------------------------------------------\n\nSolution:\n---------\n\nFunction pqReadData() closes the connection when nread == 0. Is this correct\nwhen there is no result set??\n\n--------------------------------------------------------------------------\n\nRegards\nTheo\n", "msg_date": "Mon, 8 Mar 1999 22:22:08 +0200 (SAST)", "msg_from": "Theo Kramer <[email protected]>", "msg_from_op": true, "msg_subject": "psql command drops connection to backend" } ]
[ { "msg_contents": "Hello, postgresqlers!\n\nI announce the creation of a project to build an LGPL-licensed transaction\nserver. I plan initially on supporting the CORBA transaction service\nspec, but I would also like to support the X/Open Distributed Transaction\nProcessing spec, a copy of which is on its way to me from the opengroup.\n(over $300 for a 1000-page spec! It's really outrageous.)\n\nAnyway, I am very interested in working with the PostgreSQL team. I would\nlike to use PostgreSQL as a backend for storing transaction state data\nand logs, and I would like for the PostgreSQL team to be able to use GTS\nto extend PostgreSQL to support distributed and/or replicated databases.\n\nEnclosed is the README for the project. Interested parties are encouraged\nto subscribe to the mailing list and share your thoughts.\n\n\n The G Transaction Server\n README by Todd Lewis\n begun 7 March 1999\n\nABOUT GTS\n=========\nGTS is intended to implement the CORBA transaction services spec.\nGTS has its origins in the ORBit project, and we plan on using ORBit\nas our ORB; hopefully ORBit will use us for their transaction services.\nGTS is roughly similar in function to the Microsoft Transaction Server\n(MTS), authored primarily by Jim Gray.\n\nGTS is licensed under the terms of the LGPL; see the file \"LICENSE\" for\nexact details.\n\nMOTTO\n=====\n\nThe motto of the Gnu Transaction Server is a quote from General George\nS. Patton, Jr:\n\n\t\"Rommel, you magnificent bastard, I read your book!\"\n\nINFO ON TRANSACTIONS\n====================\n\nPeople interested in playing with GTS should probably get a copy of the\nfollowing very awesome book:\n\n \"Transaction Processing : Concepts and Techniques\" \n by Jim Gray and Andreas Reuter\n ISBN 1-55860-190-2\n ($85 list price, unfortunately)\n\nAlternatively, you can just read the CORBA transaction service spec,\nwhich is included in the doc/ subdirectory.\n\nGETTING GTS\n===========\n\nYou can get GTS via anonymous CVS as follows:\n\n\t$ export CVSROOT=\":pserver:[email protected]:/cvsroot\"\n\t$ echo \"just press <enter> for the password\"; cvs login\n\t$ cvs checkout gts\n\nIf you write code for GTS, then send me mail and I'll give you an account\non the CVS server.\n\nPARTICIPATING\n=============\n\nAnyone can join the GTS mailing list. If you would like to participate in\nGTS development, then please join the list.\n\n\t$ echo \"subscribe gts-list\" | mail [email protected]\n\nTo post, send mail to \"[email protected]\". Only list members\nmay post.\n\nThe development environment for GTS is, roughly:\n\n\t- ANSI C/gcc\n\t- posix OS features/gnu libc\n\t- CORBA 2.2/ORBit\n\nIdeally, I would like to be platform-agnostic. I work on Linux, and so I\nimagine that GTS will run very well under Linux. If you want to get GTS\nrunning on a marginal system like MVS or DG/UX or Windows, then go for it\nand send us patches.\n\nFUTURE PLANS\n============\n\nEventually, I hope that we can implement the X/Open Distributed\nTransaction Processing spec as well. If we can, then we should even be\ninteroperable with MTS. I have a copy of the XA spec on the way soon;\npeople interested in this part should probably also look at the Free\nDCE project, which has (non-transactional) DCE-RPC already implemented.\n\n\tproject: http://www.bu.edu/~jrd/FreeDCE/\n\tFAQ: http://jrr.ne.mediaone.net/FAQ/FAQ.html\n\n\n\n--\nTodd Graham Lewis 32�49'N,83�36'W (800) 719-4664, x22804\n******Linux****** MindSpring Enterprises [email protected]\n\n\"A pint of sweat will save a gallon of blood.\" -- George S. Patton\n\n", "msg_date": "Mon, 8 Mar 1999 20:46:01 -0500 (EST)", "msg_from": "Todd Graham Lewis <[email protected]>", "msg_from_op": true, "msg_subject": "Announcing the G Transaction Server!" } ]
[ { "msg_contents": "Hi all,\n\nI have played with dynamic loading of modules to the backend and it works.\nIt needs only to add \"DLLIMPORT\" to all data structures, that should be\nexported.\n\nThe main thing is done in the src/include/c.h - it defines DLLIMPORT to\nnothing on platforms other then Cygwin and __declspec(...) on Cygwin, and in\nsrc/makefiles/Makefile.win - it defines the symbol BUILDING_DLL only when\nthe make process is in the src/backend subtree.\n\nI have tested it with trigger regression test - it works, and with some\nplpgsql stuff - it doesn't work and it probably needs some other data to be\nexported from the backend.\n\nI think the changes are minimal so they can by applied to the development\nsource tree for 6.5.\n\n\t\t\tDan\n\n----------------------------------------------\nDaniel Horak\nnetwork and system administrator\ne-mail: [email protected]\nprivat e-mail: [email protected]\n----------------------------------------------", "msg_date": "Tue, 9 Mar 1999 09:29:36 +0100 ", "msg_from": "Horak Daniel <[email protected]>", "msg_from_op": true, "msg_subject": "dynamic loading on NT works!" }, { "msg_contents": "\nApplied...\n\nOn Tue, 9 Mar 1999, Horak Daniel wrote:\n\n> Hi all,\n> \n> I have played with dynamic loading of modules to the backend and it works.\n> It needs only to add \"DLLIMPORT\" to all data structures, that should be\n> exported.\n> \n> The main thing is done in the src/include/c.h - it defines DLLIMPORT to\n> nothing on platforms other then Cygwin and __declspec(...) on Cygwin, and in\n> src/makefiles/Makefile.win - it defines the symbol BUILDING_DLL only when\n> the make process is in the src/backend subtree.\n> \n> I have tested it with trigger regression test - it works, and with some\n> plpgsql stuff - it doesn't work and it probably needs some other data to be\n> exported from the backend.\n> \n> I think the changes are minimal so they can by applied to the development\n> source tree for 6.5.\n> \n> \t\t\tDan\n> \n> ----------------------------------------------\n> Daniel Horak\n> network and system administrator\n> e-mail: [email protected]\n> privat e-mail: [email protected]\n> ----------------------------------------------\n> \n> \n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Tue, 9 Mar 1999 09:39:09 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] dynamic loading on NT works!" } ]
[ { "msg_contents": "Hi,\n\nAt some time I started to receive error message:\n\"ERROR: Tuple is too big: size 14800\"\nwhen I run VACUUM ANALYZE command from psql. I know that I can't insert\nrow more than 8K in size, so what this means?\n\nMike\n", "msg_date": "Tue, 09 Mar 1999 10:30:08 -0500", "msg_from": "Mikhail Terekhov <[email protected]>", "msg_from_op": true, "msg_subject": "vacuum analyze" } ]
[ { "msg_contents": "Hi all,\nI am new at this, and I have a simple (to you all) question:\n\nIs there anyway to do a bitwise AND in a SQL statement?\n\nI have a field of type \"int4\". If I want to check to see if a certain bit\nin the int4 is set or not, How can I do it?\n\nThanks in advance.\n\n- Eddie Choong\n\n\t\t-----Original Message-----\n\t\tFrom:\tVince Vielhaber [mailto:[email protected]]\n\t\tSent:\tTuesday, March 09, 1999 11:12 AM\n\t\tTo:\tThe Hermit Hacker\n\t\tCc:\tJan Wieck; [email protected]\n\t\tSubject:\tRe: [HACKERS] Developers globe\n\n\t\tOn Tue, 9 Mar 1999, The Hermit Hacker wrote:\n\n\t\t> On Tue, 9 Mar 1999, Jan Wieck wrote:\n\t\t> \n\t\t> > Marc G. Fournier wrote:\n\t\t> > \n\t\t> > > You don't really need nams associated iwth the dots,\njust make the dots\n\t\t> > > reasonably accurate...put a long/lat beside each name\nin the list, and\n\t\t> > > country or whatnot...\n\t\t> > \n\t\t> > Alternatively you might put a map behind it and\nuse some\n\t\t> > Jscript to write the names into the status line\nwhen the\n\t\t> > pointer is over the dots.\n\t\t> \n\t\t> Oh, I like that one...anyone know Javascript? :)\n\n\t\tyep. Anyone know of a tool in unix that splits up a gif\ninto many\n\t\tsmaller squares (as opposed to manually cropping)?\n\n\t\tVince.\n\t\t-- \n\t\n==========================================================================\n\t\tVince Vielhaber -- KA8CSH email: [email protected]\nflame-mail: /dev/null\n\t\t # include <std/disclaimers.h>\nTEAM-OS2\n\t\t Online Campground Directory\nhttp://www.camping-usa.com\n\t\t Online Giftshop Superstore\nhttp://www.cloudninegifts.com\n\t\n==========================================================================\n\n\n\t\t\n", "msg_date": "Tue, 9 Mar 1999 11:44:06 -0500 ", "msg_from": "\"Choong, Eddie\" <[email protected]>", "msg_from_op": true, "msg_subject": "Bitwise AND" } ]
[ { "msg_contents": "I have a function:\n\n\tvoid *nz(void *bp) {\n\t text *new_text;\n\t new_text = (text *) palloc(sizeof(text)+2);\n\t memset((void *) new_text, 0, sizeof(text)+2); \n\t strcpy(VARDATA(new_text), \":\");\n\t VARSIZE(new_text) = sizeof(text)+2;\n\t return (void *)(new_text);\n\t}\n\nI define the function in psql:\n\nCREATE FUNCTION nz(text) RETURNS text AS\n'/usr/src/pgsql/contrib/spi/psql_ext.so' LANGUAGE 'c';\n\nI call the function:\n\n\tselect addressid, nz(state) from addresses where addressid = 3870;\n\nIf state is NULL, nz(state) still returns NULL! How can I get my nz()\nfunction to reflect something other than NULL? The purpose behind this is\nto have a concatenated string return a value even if any value is null. For\nexample:\n\nselect addressid, nz(state) || nz(country) from addresses where addressid =\n3870;\n\nReturns NULL if state is null or if country is null. I want it to return\nthe country if state is null.\n", "msg_date": "Tue, 9 Mar 1999 13:53:48 -0600 ", "msg_from": "Michael Davis <[email protected]>", "msg_from_op": true, "msg_subject": "Nulls" } ]
[ { "msg_contents": "\n> \n> Bruce, how's this for a map? It'd probably be trivial if you wanted one\n> color for the land and make the water blue.\n> \n> http://www.pop4.net/~vev/bigworld.gif\n> \n\nIt is nice looking. Would people prefer this?\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 9 Mar 1999 18:04:27 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: map" }, { "msg_contents": "On Tue, 9 Mar 1999, Bruce Momjian wrote:\n\n> \n> > \n> > Bruce, how's this for a map? It'd probably be trivial if you wanted one\n> > color for the land and make the water blue.\n> > \n> > http://www.pop4.net/~vev/bigworld.gif\n> > \n> \n> It is nice looking. Would people prefer this?\n\nI like the rotating globe, since its realy \"cool\" to look at...but its too\nsmall...we're going to end up crowding things...\n\nI prefer the 'flat-globe' for what you are attempting to do...its also\neasier to see...you don't have to \"follow the dot\"...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n", "msg_date": "Tue, 9 Mar 1999 19:18:31 -0400 (AST)", "msg_from": "The Hermit Hacker <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "> I like the rotating globe, since its realy \"cool\" to look at...but its too\n> small...we're going to end up crowding things...\n> \n> I prefer the 'flat-globe' for what you are attempting to do...its also\n> easier to see...you don't have to \"follow the dot\"...\n\nDone. I can play with the colors and stuff.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Tue, 9 Mar 1999 18:20:52 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "Thus spake Bruce Momjian\n> > Bruce, how's this for a map? It'd probably be trivial if you wanted one\n> > color for the land and make the water blue.\n> > \n> > http://www.pop4.net/~vev/bigworld.gif\n> \n> It is nice looking. Would people prefer this?\n\nI vote yes. The spinning globe doesn't really do much for me.\n\n-- \nD'Arcy J.M. Cain <darcy@{druid|vex}.net> | Democracy is three wolves\nhttp://www.druid.net/darcy/ | and a sheep voting on\n+1 416 424 2871 (DoD#0082) (eNTP) | what's for dinner.\n", "msg_date": "Tue, 9 Mar 1999 22:04:48 -0500 (EST)", "msg_from": "\"D'Arcy\" \"J.M.\" Cain <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "How difficult would it be to allow attribute\nnames to be aliased when inheriting:\n\nCREATE TABLE base \n( oldtag TEXT );\n\nCREATE TABLE derived \n( more_attrib INT4 )\nINHERITS(base)\n--\nWITH base.oldtag \n AS derived.newtag\n\nSo that\n\nSELECT * from derived\n\nnewtag | more_attrib\n-------+-----------\nder#1 | derived value #1\n\n\nand\n\nSELECT * from base*\n\noldtag \n------\nbase #1\nder #1\n\nThoughts? Is this _that_ bad of an idea. I think\nit would really help to manage aggregation complexity.\n\nI suppose this is a big project hunh? \n\nThanks!\n\nClark Evans\n", "msg_date": "Wed, 10 Mar 1999 04:57:06 +0000", "msg_from": "Clark Evans <[email protected]>", "msg_from_op": false, "msg_subject": "Re-Name Attributes on Inheritance" }, { "msg_contents": "> On Tue, 9 Mar 1999, Bruce Momjian wrote:\n> \n> > \n> > > \n> > > Bruce, how's this for a map? It'd probably be trivial if you wanted one\n> > > color for the land and make the water blue.\n> > > \n> > > http://www.pop4.net/~vev/bigworld.gif\n> > > \n> > \n> > It is nice looking. Would people prefer this?\n> \n> I like the rotating globe, since its realy \"cool\" to look at...but its too\n> small...we're going to end up crowding things...\n> \n> I prefer the 'flat-globe' for what you are attempting to do...its also\n> easier to see...you don't have to \"follow the dot\"...\n\nOK, new map, older dots, but they now have fuzzy edges. Much better.\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 10:36:41 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "Clark Evans <[email protected]> writes:\n> How difficult would it be to allow attribute\n> names to be aliased when inheriting:\n> ...\n> Thoughts? Is this _that_ bad of an idea.\n\nIt seems like a horrible idea to me ;-). The point of inheritance\nis that whatever else your derived class may be, it *IS A* parent-\nclass object as well, and anything that works on the parent class\nwill work on the subclass. In your example,\n\tSELECT more_attrib FROM base where oldtag = something;\nwould work but\n\tSELECT more_attrib FROM derived where oldtag = something;\nwould fail. That's not my idea of a derived class.\n\nIt's not clear exactly what you want to accomplish here, but I\nwonder whether inheritance is the right model for the relationships\namong the tables in your database at all. It seems like you are\ntrying to force-fit some other kind of relationship into the\ninheritance model.\n\n\t\t\tregards, tom lane\n", "msg_date": "Wed, 10 Mar 1999 10:47:05 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re-Name Attributes on Inheritance " }, { "msg_contents": "Bruce Momjian <[email protected]> writes:\n>> I prefer the 'flat-globe' for what you are attempting to do...its also\n>> easier to see...you don't have to \"follow the dot\"...\n\n> OK, new map, older dots, but they now have fuzzy edges. Much better.\n\nI like the look better, but (putting on my graphics-guy hat) that GIF\nis *huge*. 175K is a tad of an excessive download for a web-page\nframmish. Also it doesn't look very good if your browser's page\nbackground is not white --- there's a white ring around each\ncontinent thanks to poor antialiasing.\n\nThe extent of the dithering makes me think the image was made from a\ntruecolor original. If so JPEG at a moderately high quality setting\nwould work better (and look much better on truecolor displays), though\nyou'd lose the ability to have a transparent background. PNG is also a\npossibility if you want to assume that visitors have recent browsers.\n\n\t\t\tregards, tom lane\n\t\t\torganizer, Independent JPEG Group\n\t\t\tmember, PNG development group\n", "msg_date": "Wed, 10 Mar 1999 11:02:02 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map " }, { "msg_contents": "On Wed, 10 Mar 1999, Tom Lane wrote:\n\n> Bruce Momjian <[email protected]> writes:\n> >> I prefer the 'flat-globe' for what you are attempting to do...its also\n> >> easier to see...you don't have to \"follow the dot\"...\n> \n> > OK, new map, older dots, but they now have fuzzy edges. Much better.\n> \n> I like the look better, but (putting on my graphics-guy hat) that GIF\n> is *huge*. 175K is a tad of an excessive download for a web-page\n> frammish. \n\n175K? It started at 24K, a load/save in xv brings it back to 24K.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n\n", "msg_date": "Wed, 10 Mar 1999 11:11:23 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map " }, { "msg_contents": "Vince Vielhaber <[email protected]> writes:\n> 175K? It started at 24K, a load/save in xv brings it back to 24K.\n\nWhy, so it does --- didn't think to try that. Bruce, are you using\nsomething that outputs \"uncompressed GIF\" to avoid the LZW patent?\n\nMy remarks about dithering and the antialias ring still stand though.\n\n\t\t\tregards, tom lane\n", "msg_date": "Wed, 10 Mar 1999 11:15:05 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map " }, { "msg_contents": "> Bruce Momjian <[email protected]> writes:\n> >> I prefer the 'flat-globe' for what you are attempting to do...its also\n> >> easier to see...you don't have to \"follow the dot\"...\n> \n> > OK, new map, older dots, but they now have fuzzy edges. Much better.\n> \n> I like the look better, but (putting on my graphics-guy hat) that GIF\n> is *huge*. 175K is a tad of an excessive download for a web-page\n> frammish. Also it doesn't look very good if your browser's page\n\nNot sure about that. You can read the text below as it is loading. The\nbrowser seems to know the image size.\n\n> background is not white --- there's a white ring around each\n> continent thanks to poor antialiasing.\n\nWe force a cream background for that page, no?\n\n> \n> The extent of the dithering makes me think the image was made from a\n> truecolor original. If so JPEG at a moderately high quality setting\n> would work better (and look much better on truecolor displays), though\n> you'd lose the ability to have a transparent background. PNG is also a\n> possibility if you want to assume that visitors have recent browsers.\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 11:54:57 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "> On Wed, 10 Mar 1999, Tom Lane wrote:\n> \n> > Bruce Momjian <[email protected]> writes:\n> > >> I prefer the 'flat-globe' for what you are attempting to do...its also\n> > >> easier to see...you don't have to \"follow the dot\"...\n> > \n> > > OK, new map, older dots, but they now have fuzzy edges. Much better.\n> > \n> > I like the look better, but (putting on my graphics-guy hat) that GIF\n> > is *huge*. 175K is a tad of an excessive download for a web-page\n> > frammish. \n> \n> 175K? It started at 24K, a load/save in xv brings it back to 24K.\n\nLet me look at that. Seems to have grown somehow. It is the\ntransparency that is causing it. I am using 'convert' to add the\ntransparency. Have other people seen 'transparancy' increasing the file\nsize?\n\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 11:57:53 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "On Wed, 10 Mar 1999, Bruce Momjian wrote:\n> Let me look at that. Seems to have grown somehow. It is the\n> transparency that is causing it. I am using 'convert' to add the\n> transparency. Have other people seen 'transparancy' increasing the file\n> size?\n\n Convert (if you meant ImageMagic's convert) definetely does not use LZW to\ncompress GIFs :(\n\n> -- \n> Bruce Momjian | http://www.op.net/~candle\n> [email protected] | (610) 853-3000\n> + If your life is a hard drive, | 830 Blythe Avenue\n> + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n> \n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Wed, 10 Mar 1999 20:06:25 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "> On Wed, 10 Mar 1999, Bruce Momjian wrote:\n> > Let me look at that. Seems to have grown somehow. It is the\n> > transparency that is causing it. I am using 'convert' to add the\n> > transparency. Have other people seen 'transparancy' increasing the file\n> > size?\n> \n> Convert (if you meant ImageMagic's convert) definetely does not use LZW to\n> compress GIFs :(\n\nMan, that stinks. OK, I am now using giftrans, and that works and\nproduces small files.\n\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 12:26:03 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "On Wed, 10 Mar 1999, Bruce Momjian wrote:\n\n> Date: Wed, 10 Mar 1999 12:26:03 -0500 (EST)\n> From: Bruce Momjian <[email protected]>\n> To: [email protected]\n> Cc: [email protected], [email protected], [email protected]\n> Subject: Re: [HACKERS] Re: map\n> \n> > On Wed, 10 Mar 1999, Bruce Momjian wrote:\n> > > Let me look at that. Seems to have grown somehow. It is the\n> > > transparency that is causing it. I am using 'convert' to add the\n> > > transparency. Have other people seen 'transparancy' increasing the file\n> > > size?\n> > \n> > Convert (if you meant ImageMagic's convert) definetely does not use LZW to\n> > compress GIFs :(\n> \n> Man, that stinks. OK, I am now using giftrans, and that works and\n> produces small files.\n\nI see a big thread in mailing list about 'map', so let me also\npost a reference to gifsicle. This is a great tool !\n\nhttp://www.pdos.lcs.mit.edu/~eddietwo/gifsicle/\n\n\tRegards,\n\n\t\tOleg\n> \n> \n> -- \n> Bruce Momjian | http://www.op.net/~candle\n> [email protected] | (610) 853-3000\n> + If your life is a hard drive, | 830 Blythe Avenue\n> + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n> \n\n_____________________________________________________________\nOleg Bartunov, sci.researcher, hostmaster of AstroNet,\nSternberg Astronomical Institute, Moscow University (Russia)\nInternet: [email protected], http://www.sai.msu.su/~megera/\nphone: +007(095)939-16-83, +007(095)939-23-83\n\n", "msg_date": "Wed, 10 Mar 1999 20:43:09 +0300 (MSK)", "msg_from": "Oleg Bartunov <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "> On Wed, 10 Mar 1999, Bruce Momjian wrote:\n> > Let me look at that. Seems to have grown somehow. It is the\n> > transparency that is causing it. I am using 'convert' to add the\n> > transparency. Have other people seen 'transparancy' increasing the file\n> > size?\n> \n> Convert (if you meant ImageMagic's convert) definetely does not use LZW to\n> compress GIFs :(\n\nI now see ImageMagick 3.9 has:\n\n\tEnable LZW compression with -HasLZW. See Magick.tmpl or Makefile.in. \n\nI am running 3.8.9. Oh, well, have to wait for a BSD/OS upgrade, or do\nit myself.\n\nActually, changelog for 3.7.* say:\n\n\tRather than a separate LZW kit, Image Magic now comes as two\n\tdistributions. One with LZW compression\n\t and one without (suggested by [email protected]). \n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 13:08:00 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "Bruce Momjian <[email protected]> writes:\n>> background is not white --- there's a white ring around each\n>> continent thanks to poor antialiasing.\n\n> We force a cream background for that page, no?\n\nIt's gray on my browser (Netscape 4.0something). Even if it were\ncream, the background of the gif is *white* not cream, so the ring\nwould still be there (maybe a little harder to see though).\n\nSince GIF hasn't got partial transparency, you can't do real\nantialiasing without making the background colors match.\nYou're probably best off not using transparency at all, unless\nyou forget about antialiasing. (Or is it antialiasing that's\nleaving the near-white pixels there? Maybe it's just sloppy drawing\nof the border of the transparent region?)\n\n\t\t\tregards, tom lane\n", "msg_date": "Wed, 10 Mar 1999 15:03:20 -0500", "msg_from": "Tom Lane <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map " }, { "msg_contents": ">\n> Bruce Momjian <[email protected]> writes:\n> >> background is not white --- there's a white ring around each\n> >> continent thanks to poor antialiasing.\n>\n> > We force a cream background for that page, no?\n>\n> It's gray on my browser (Netscape 4.0something). Even if it were\n> cream, the background of the gif is *white* not cream, so the ring\n> would still be there (maybe a little harder to see though).\n>\n> Since GIF hasn't got partial transparency, you can't do real\n> antialiasing without making the background colors match.\n> You're probably best off not using transparency at all, unless\n> you forget about antialiasing. (Or is it antialiasing that's\n> leaving the near-white pixels there? Maybe it's just sloppy drawing\n> of the border of the transparent region?)\n\n Did you take a look at my one? It's using transparency too,\n but since I created the image by program, there are only 4\n colors at all (plus tranparent). Has it noise too or is it\n O.K.?\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 10 Mar 1999 21:33:50 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "> Bruce Momjian <[email protected]> writes:\n> >> background is not white --- there's a white ring around each\n> >> continent thanks to poor antialiasing.\n> \n> > We force a cream background for that page, no?\n> \n> It's gray on my browser (Netscape 4.0something). Even if it were\n> cream, the background of the gif is *white* not cream, so the ring\n> would still be there (maybe a little harder to see though).\n> \n> Since GIF hasn't got partial transparency, you can't do real\n> antialiasing without making the background colors match.\n> You're probably best off not using transparency at all, unless\n> you forget about antialiasing. (Or is it antialiasing that's\n> leaving the near-white pixels there? Maybe it's just sloppy drawing\n> of the border of the transparent region?)\n\nI am confused. I don't even know what anti-aliasing is. Nor do I\nunderstand why you get a grey background. I am running Netscape 4.5,\nand I see the background of all the pages as off-white.\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 15:45:53 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": ">\n> > Bruce Momjian <[email protected]> writes:\n> > >> background is not white --- there's a white ring around each\n> > >> continent thanks to poor antialiasing.\n> >\n> > > We force a cream background for that page, no?\n> >\n> > It's gray on my browser (Netscape 4.0something). Even if it were\n> > cream, the background of the gif is *white* not cream, so the ring\n> > would still be there (maybe a little harder to see though).\n> >\n> > Since GIF hasn't got partial transparency, you can't do real\n> > antialiasing without making the background colors match.\n> > You're probably best off not using transparency at all, unless\n> > you forget about antialiasing. (Or is it antialiasing that's\n> > leaving the near-white pixels there? Maybe it's just sloppy drawing\n> > of the border of the transparent region?)\n>\n> I am confused. I don't even know what anti-aliasing is. Nor do I\n> understand why you get a grey background. I am running Netscape 4.5,\n> and I see the background of all the pages as off-white.\n\n If the <BODY> doesn't define a background, Netscape uses it's\n (configurable) default one. So your config say's white (why\n not slategray?).\n\n\nJan\n\n--\n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n\n", "msg_date": "Wed, 10 Mar 1999 21:50:33 +0100 (MET)", "msg_from": "[email protected] (Jan Wieck)", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "> > I am confused. I don't even know what anti-aliasing is. Nor do I\n> > understand why you get a grey background. I am running Netscape 4.5,\n> > and I see the background of all the pages as off-white.\n> \n> If the <BODY> doesn't define a background, Netscape uses it's\n> (configurable) default one. So your config say's white (why\n> not slategray?).\n\nBut there is a style sheet, no?\n\n-- \n Bruce Momjian | http://www.op.net/~candle\n [email protected] | (610) 853-3000\n + If your life is a hard drive, | 830 Blythe Avenue\n + Christ can be your backup. | Drexel Hill, Pennsylvania 19026\n", "msg_date": "Wed, 10 Mar 1999 16:07:23 -0500 (EST)", "msg_from": "Bruce Momjian <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Re: map" }, { "msg_contents": "\nOn 10-Mar-99 Jan Wieck wrote:\n>>\n>> > Bruce Momjian <[email protected]> writes:\n>> I am confused. I don't even know what anti-aliasing is. Nor do I\n>> understand why you get a grey background. I am running Netscape 4.5,\n>> and I see the background of all the pages as off-white.\n> \n> If the <BODY> doesn't define a background, Netscape uses it's\n> (configurable) default one. So your config say's white (why\n> not slategray?).\n\nIt's defined in the style sheet for the page as: #fffdec\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Campground Directory http://www.camping-usa.com\n Online Giftshop Superstore http://www.cloudninegifts.com\n==========================================================================\n\n\n", "msg_date": "Wed, 10 Mar 1999 16:20:55 -0500 (EST)", "msg_from": "Vince Vielhaber <[email protected]>", "msg_from_op": false, "msg_subject": "Re: [HACKERS] Re: map" } ]
[ { "msg_contents": "Hello!\n\n I rewrote my 4-tables join to use subselects:\n\nSELECT DISTINCT subsec_id FROM positions\n WHERE pos_id IN\n (SELECT DISTINCT pos_id\n FROM central\n WHERE shop_id IN\n (SELECT shop_id FROM shops\n WHERE distr_id IN\n (SELECT distr_id FROM districts\n WHERE city_id = 2)\n )\n )\n;\n\n This does not work, either - postgres loops forever, until I cancel\npsql.\n\n I splitted it - I ran\n\n (SELECT DISTINCT pos_id\n FROM central\n WHERE shop_id IN\n (SELECT shop_id FROM shops\n WHERE distr_id IN\n (SELECT distr_id FROM districts\n WHERE city_id = 2)\n )\n )\n\n and stored result in a file. Then I substituted the subselect with the\nfile:\n\nSELECT DISTINCT subsec_id FROM positions\n WHERE pos_id IN\n(1, 2, 3, 6, 22, 25, 26, 27, 28, 29, 31, 33, 34, 35, 38, 41, 42, 44, 45,\n46, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 60, 61, 62, 63, 64)\n\n and got desired result within a second.\n\n This finally solves my problem, but I need to pass a long way to find\nthat postgres cannot handle such not too complex joins and subselects.\n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Wed, 10 Mar 1999 12:33:34 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Bug on complex subselect (was: Bug on complex join)" }, { "msg_contents": "Hello!\n\n Now for those who said he cannot beleive there is a bug in join - the\nfollowing query produced two different results on two different systems\nwith identical data loaded:\n\nSELECT DISTINCT p.subsec_id\n FROM central cn, shops sh, districts d, positions p\n WHERE cn.shop_id = sh.shop_id AND sh.distr_id = d.distr_id\n AND d.city_id = 2 AND cn.pos_id = p.pos_id\n;\nsubsec_id\n---------\n 1\n 2\n 10\n 11\n 12\n 13\n 14\n 15\n(8 rows)\n\n on one system and\n\nsubsec_id\n---------\n(0 rows)\n\n on the other.\n\n First system is sparc-solrais and second is intel-linux (Debian 2.0,\nglibc2). Another glibc2-realted bug?\n\nOleg.\n---- \n Oleg Broytmann http://members.xoom.com/phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n", "msg_date": "Mon, 15 Mar 1999 16:36:06 +0300 (MSK)", "msg_from": "Oleg Broytmann <[email protected]>", "msg_from_op": true, "msg_subject": "Re: [HACKERS] Bug on complex join" } ]
[ { "msg_contents": "\n-- \nI get the following when I try to pg_dump a database which contains\nlarge\nobjects:\n\n\"failed sanity check, table xinv162689 was not found\"\n\nis this a known bug ? \n\n\n\n~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\n David O'Farrell AerSoft Limited \n mailto:[email protected] 2 Northumberland Avenue,\n Dun Laoghaire,Co. Dublin\n\tDirect Phone 353-1-2145950\n Phone: 01-2301166 Fax: 01-2301167\n~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\n", "msg_date": "Wed, 10 Mar 1999 09:57:23 +0000", "msg_from": "\"David O'Farrell\" <[email protected]>", "msg_from_op": true, "msg_subject": "pg_dump and large objects ,sunos 5.6 , SC4.2" } ]
[ { "msg_contents": "Coming into this discussion late but...\n\nI seem to remember being able to get the browser to display popup labels\nover parts of an image using the alt html tag with each area of a client\nside image map.\n\n--\nPeter T Mount, IT Section\[email protected]\nAnything I write here are my own views, and cannot be taken as the\nofficial words of Maidstone Borough Council\n\n\n-----Original Message-----\nFrom: [email protected] [mailto:[email protected]]\nSent: Wednesday, March 10, 1999 9:08 AM\nTo: [email protected]\nCc: [email protected]; [email protected]; [email protected]\nSubject: Re: [HACKERS] Developers globe\n\n\n> > Hrmmmm...isn't that what he said? *grin*\n> \n> Nope. Status line is at the bottom of the browser. I meant for the\nname\n> to pop up (or a flag to pop up) in the actual image.\n\n Right - and popping up in the image is better :-)\n\n\nJan\n\n-- \n\n#======================================================================#\n# It's easier to get forgiveness for being wrong than for being right. #\n# Let's break this rule - forgive me. #\n#======================================== [email protected] (Jan Wieck) #\n\n", "msg_date": "Wed, 10 Mar 1999 11:46:26 -0000", "msg_from": "Peter Mount <[email protected]>", "msg_from_op": true, "msg_subject": "RE: [HACKERS] Developers globe" } ]