threads
listlengths 1
2.99k
|
|---|
[
{
"msg_contents": "Hi, all.\n\nI've just sent a patch to the pgsql-patches list that (among other fixes\nspecific for Digital Unix) contains a patch for\nsrc/backent/ports/snprintf.c. That file contained two bugs: it incorrectly\nused HAVE_LONG_INT_64 to check if it could use the \"long long int\" type\n(changed to HAVE_LONG_LONG_INT_64), and it misused the variable argument\nmechanism defined in stdarg.h (vsnprintf was incorrectly defined, for\nexample).\n\nThe patch fixes the type int8 under Digital Unix, and it should fix it\nalso for systems that don't have native snprintf/vsnprintf.\n\nPlease take a look at it, and let me know if you see something wrong.\n\n-- \n-------------------------------------------------------------------\nPedro Jos� Lobo Perea Tel: +34 91 336 78 19\nCentro de C�lculo Fax: +34 91 331 92 29\nEUIT Telecomunicaci�n - UPM e-mail: [email protected]\n\n",
"msg_date": "Thu, 17 Dec 1998 17:09:57 +0100 (MET)",
"msg_from": "\"Pedro J. Lobo\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Bug fix for src/backend/ports/snprintf.c sent to the patches list"
}
] |
[
{
"msg_contents": "I am working on doing the HISTORY file for the 6.4.1 release.\n\nI can't figure out how to generate a cvs log for only the REL6_4 cvs\ntree.\n\nCan anyone tell me how to do 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": "Thu, 17 Dec 1998 12:32:21 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "CVS log for a specific tag"
},
{
"msg_contents": "Bruce asked:\n\n>\n> I am working on doing the HISTORY file for the 6.4.1 release.\n>\n> I can't figure out how to generate a cvs log for only the REL6_4 cvs\n> tree.\n>\n> Can anyone tell me how to do it?\n\n#!/bin/sh\n\negrep -e '/[0-9]+\\.[0-9]+\\.2\\.[0-9]+/' `find . -name Entries -print` | \\\n grep '[^:]*CVS/Entries:' | \\\n sed -e 's/\\(Entries:\\/[^\\/]*\\).*/\\1/' | \\\n sed -e 's/CVS\\/Entries:\\///' | \\\nwhile read f ; do\n cvs log -rREL6_4: $f\ndone\n\n Use this script in the working directory where you checked\n out the REL6_4 branch. I depends on the fact that every file,\n touched in that branch, has a 2 in it's third element of the\n revision number.\n\n The egrep-sed party just selects all file names from the\n CVS/Entries which have such a revision number. Then cvs is\n called for each to print out the log from REL6_4 to the last\n revision in the branch.\n\n Make sure there is no blank between -r and REL6_4: on the cvs\n log call, or the called rlog will misinterpret it as a\n filename and print out the complete logging for the trunk.\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, 17 Dec 1998 23:44:42 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] CVS log for a specific tag"
},
{
"msg_contents": "> Bruce asked:\n> \n> >\n> > I am working on doing the HISTORY file for the 6.4.1 release.\n> >\n> > I can't figure out how to generate a cvs log for only the REL6_4 cvs\n> > tree.\n> >\n> > Can anyone tell me how to do it?\n> \n> #!/bin/sh\n> \n> egrep -e '/[0-9]+\\.[0-9]+\\.2\\.[0-9]+/' `find . -name Entries -print` | \\\n> grep '[^:]*CVS/Entries:' | \\\n> sed -e 's/\\(Entries:\\/[^\\/]*\\).*/\\1/' | \\\n> sed -e 's/CVS\\/Entries:\\///' | \\\n> while read f ; do\n> cvs log -rREL6_4: $f\n> done\n> \n> Use this script in the working directory where you checked\n> out the REL6_4 branch. I depends on the fact that every file,\n> touched in that branch, has a 2 in it's third element of the\n> revision number.\n> \n> The egrep-sed party just selects all file names from the\n> CVS/Entries which have such a revision number. Then cvs is\n> called for each to print out the log from REL6_4 to the last\n> revision in the branch.\n> \n> Make sure there is no blank between -r and REL6_4: on the cvs\n> log call, or the called rlog will misinterpret it as a\n> filename and print out the complete logging for the trunk.\n\nI have attached my pgcvslog, modified to allow revisions to be specified:\n\n\tpgcvslog -r '\\.2\\.[0-9]*$'\n\nIt is a shame that we have to rely on one of the revisions numbers to\npull out the REL6_4 branch of the tree.\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# Usage $0 [-r 'revision pattern']\n\n# pgcvslog -r REL6_4\n\nif [ \"X$1\" = \"X-r\" ]\nthen\tREV=\"$2\"\n\tshift 2\nelse\tREV=\".*\"\nfi\n\ncat \"$@\" |\n\n# mark each line with a datetime and line number, for sorting and merging\n# we don't print anything from the -- or == line and the date:\n\nawk '\n\t$0 ~ /^Working file:/\t{workingfile = $0}\n\n\t$1 == \"revision\" && $2 !~ /'\"$REV\"'/ {skip = \"Y\"}\n\n\t($0 ~ /^====*$/ || $0 ~ /^----*$/) && skip == \"N\" \\\n\t{\n\t\t/* print blank line separating entries */\n\t\tif (datetime != \"\")\n\t\t{\n\t\t\tprintf (\"%s| %10d|%s\\n\", datetime, NR, \"\");\n\t\t\tprintf (\"%s| %10d|%s\\n\", datetime, NR, \"---\");\n\t\t\tprintf (\"%s| %10d|%s\\n\", datetime, NR+1, \"\");\n\t\t}\n\t}\n\n\t$0 ~ /^====*$/ || $0 ~ /^----*$/ \\\n\t{\n\t\tdatetime=\"\";\n\t\tskip=\"N\";\n\t}\n\n\tdatetime != \"\" && skip == \"N\" \\\n\t\t{printf (\"%s| %10d| %s\\n\", datetime, NR, $0);}\n\n\t$1 == \"date:\"\t\\\n\t{\n\t\t/* get entry date */\n\t\tdatetime=$2\"-\"$3\n\t\tif (workingfile != \"\" && skip == \"N\")\n\t\t{\n\t\t\tprintf (\"%s| %10d|%s\\n\", datetime, NR-1, workingfile);\n\t\t\tprintf (\"%s| %10d|%s\\n\", datetime, NR, $0);\n\t\t\tprintf (\"%s| %10d|%s\\n\", datetime, NR+1, \"\");\n\t\t}\n\t}\n\n\t$0 ~ /^====*$/ \t\t\t{workingfile=\"\"}' |\n\nsort | cut -d'|' -f3 | cat |\n\n# collect duplicate narratives\nawk '\tBEGIN\t{ slot = 0;}\n\t{\n\t\tif ($0 ~ /^Working file:/)\n\t\t{\n\t\t\tif (slot != oldslot)\n\t\t\t\tsame = 0;\n\t\t\telse\n\t\t\t{\n\t\t\t\tsame = 1;\n\t\t\t\tfor (i=1; i <= slot; i++)\n\t\t\t\t{\n\t\t\t\t\tif (oldnarr[i] != narr[i])\n\t\t\t\t\t\tsame = 0;\n\t\t\t\t}\n\t\t\t}\n\n\t\t\tif (oldslot && !same)\n\t\t\t\tfor (i=1; i <= oldslot; i++)\n\t\t\t\t\tprint oldnarr[i];\n\t\t\tfor (i=1; i <= slot; i++)\n\t\t\t\toldnarr[i] = narr[i];\n\t\t\toldslot = slot;\n\t\t\tslot = 0;\n\t\t\tprint save_working;\n\t\t\tsave_working = $0;\n \t\t}\n\t\telse if ($1 != \"date:\")\n\t\t\tnarr[++slot] = $0;\n\t}\n\tEND\t{\n\t\t\tprint save_working;\n\t\t\tfor (i=1; i <= slot; i++)\n\t\t\t\tprint narr[i];\n\t\t}'",
"msg_date": "Thu, 17 Dec 1998 22:08:32 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] CVS log for a specific tag"
}
] |
[
{
"msg_contents": "Appologies,\n\nIt looks like this was not the fix, as I'm still getting the\nBad node message....\n\n\nWill continue investigating.\n\nKeith.\n\nKeith Parks <[email protected]>\n> \n> Vadim Mikheev <[email protected]>\n> > \n> > Keith Parks wrote:\n> > > \n> > > Vadim Mikheev <[email protected]>\n> > > >\n> > > > ... in SELECT * FROM street;\n> > > >\n> > > \n> > > No crash just:-\n> > > \n> > > regression=> select * from street;\n> > > ERROR: nodeRead: Bad type 0\n> > > regression=>\n> > \n> > It seems platform dependent...\n> > Bugs are in readfuncs.c\n> > \n> \n> I think I've found it, a simple typo in outfuncs.c.\n> \n> Looks like :vartypmod got transmuted to %vartypmod in an editing session.\n> \n> Here's the patch,\n> \n> Keith.\n> \n> *** src/backend/nodes/outfuncs.c.orig Thu Dec 17 12:01:02 1998\n> --- src/backend/nodes/outfuncs.c Thu Dec 17 12:01:22 1998\n> ***************\n> *** 634,640 ****\n> _outVar(StringInfo str, Var *node)\n> {\n> appendStringInfo(str,\n> ! \" VAR :varno %d :varattno %d :vartype %u %vartypmod %d \",\n> node->varno,\n> node->varattno,\n> node->vartype,\n> --- 634,640 ----\n> _outVar(StringInfo str, Var *node)\n> {\n> appendStringInfo(str,\n> ! \" VAR :varno %d :varattno %d :vartype %u :vartypmod %d \",\n> node->varno,\n> node->varattno,\n> node->vartype, \n> \n> \n\n",
"msg_date": "Thu, 17 Dec 1998 22:04:00 +0000 (GMT)",
"msg_from": "Keith Parks <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] CURRENT: crash in select_view regression test..."
}
] |
[
{
"msg_contents": "Hi,\n\nI think I need some help here\n\nTaking part of the pg_views rule shows it thinks the views\nare NOT views.\n \npostgres=> SELECT relname AS viewname, pg_get_userbyid(relowner) AS viewowner, \npg_get_viewdef(relname) AS definition FROM pg_class WHERE relhasrules;\nviewname |viewowner|definition\n----------+---------+----------\npg_user |postgres |Not a view\npg_rules |postgres |Not a view\npg_views |postgres |Not a view\npg_tables |postgres |Not a view\npg_indexes|postgres |Not a view\n(5 rows)\n\nIf I select from the view I get a BE crash.\n\n\npostgres=> select * from pg_views;\npqReadData() -- backend closed the channel unexpectedly.\n This probably means the backend terminated abnormally before or while processing the \nrequest.\nWe have lost the connection to the backend, so further processing is impossible. Terminating. \n\n\nHere's the backtrace...\n\nProgram received signal SIGSEGV, Segmentation fault.\n0xbcd64 in ApplyRetrieveRule (parsetree=0x1f0790, rule=0xefffaf20, rt_index=1, relation_level=1,\n relation=0x1f0690, modified=0xefffaf1c) at rewriteHandler.c:1817\n1817 rtable = nconc(rtable, copyObject(rule_action->rtable));\n(gdb) bt\n#0 0xbcd64 in ApplyRetrieveRule (parsetree=0x1f0790, rule=0xefffaf20, rt_index=1, \nrelation_level=1,\n relation=0x1f0690, modified=0xefffaf1c) at rewriteHandler.c:1817\n#1 0xbd250 in fireRIRrules (parsetree=0x1f0790) at rewriteHandler.c:2079\n#2 0xbd950 in QueryRewrite (parsetree=0x198a50) at rewriteHandler.c:2615\n#3 0xceeb4 in pg_parse_and_plan (query_string=0xefffd1a0 \"select * from pg_views;\", typev=0x0, \nnargs=0,\n queryListP=0xefffd094, dest=Remote, aclOverride=0 '\\000') at postgres.c:505\n#4 0xcf290 in pg_exec_query_dest (query_string=0xefffd1a0 \"select * from pg_views;\", \ndest=Remote,\n aclOverride=0 '\\000') at postgres.c:722\n#5 0xcf244 in pg_exec_query (query_string=0xefffd1a0 \"select * from pg_views;\") at \npostgres.c:699\n#6 0xd0588 in PostgresMain (argc=-8192, argv=0x15dc00, real_argc=10, real_argv=0xeffffd84) at \npostgres.c:1646\n#7 0xb2fd4 in DoBackend (port=0x135800) at postmaster.c:1532\n#8 0xb2a60 in BackendStartup (port=0x199c00) at postmaster.c:1303\n#9 0xb1ec8 in ServerLoop () at postmaster.c:757\n#10 0xb1a10 in PostmasterMain (argc=0, argv=0xeffffd84) at postmaster.c:563\n#11 0x83f14 in main (argc=10, argv=0xeffffd84) at main.c:93 \n\nIf I start looking around in ruleutils.c why can't I see \"spirc\"??\n\nBreakpoint 2, pg_get_viewdef (rname=0xe01d56a0) at ruleutils.c:277\n277 if (spirc != SPI_OK_SELECT)\n(gdb) print spirc\nNo symbol \"spirc\" in current context.\n(gdb) list\n272 args[1] = PointerGetDatum(name2);\n273 nulls[0] = ' ';\n274 nulls[1] = ' ';\n275 nulls[2] = '\\0';\n276 spirc = SPI_execp(plan_getview, args, nulls, 1);\n277 if (spirc != SPI_OK_SELECT)\n278 elog(ERROR, \"failed to get pg_rewrite tuple for view %s\", rulename);\n279 if (SPI_processed != 1)\n280 tmp = \"Not a view\";\n281 else \n\nJan in particular, any idea what's happening here?\n\nKeith.\n\n\n\nKeith Parks <[email protected]>\n> Appologies,\n> \n> It looks like this was not the fix, as I'm still getting the\n> Bad node message....\n> \n> \n> Will continue investigating.\n> \n> Keith.\n> \n> Keith Parks <[email protected]>\n> > \n> > Vadim Mikheev <[email protected]>\n> > > \n> > > Keith Parks wrote:\n> > > > \n> > > > Vadim Mikheev <[email protected]>\n> > > > >\n> > > > > ... in SELECT * FROM street;\n> > > > >\n> > > > \n> > > > No crash just:-\n> > > > \n> > > > regression=> select * from street;\n> > > > ERROR: nodeRead: Bad type 0\n> > > > regression=>\n> > > \n> > > It seems platform dependent...\n> > > Bugs are in readfuncs.c\n> > > \n> > \n> > I think I've found it, a simple typo in outfuncs.c.\n> > \n> > Looks like :vartypmod got transmuted to %vartypmod in an editing session.\n> > \n> > Here's the patch,\n> > \n> > Keith.\n> > \n> > *** src/backend/nodes/outfuncs.c.orig Thu Dec 17 12:01:02 1998\n> > --- src/backend/nodes/outfuncs.c Thu Dec 17 12:01:22 1998\n> > ***************\n> > *** 634,640 ****\n> > _outVar(StringInfo str, Var *node)\n> > {\n> > appendStringInfo(str,\n> > ! \" VAR :varno %d :varattno %d :vartype %u %vartypmod %d \",\n> > node->varno,\n> > node->varattno,\n> > node->vartype,\n> > --- 634,640 ----\n> > _outVar(StringInfo str, Var *node)\n> > {\n> > appendStringInfo(str,\n> > ! \" VAR :varno %d :varattno %d :vartype %u :vartypmod %d \",\n> > node->varno,\n> > node->varattno,\n> > node->vartype, \n> > \n> > \n> \n\n",
"msg_date": "Fri, 18 Dec 1998 01:16:41 +0000 (GMT)",
"msg_from": "Keith Parks <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] CURRENT: crash in select_view regression test..."
},
{
"msg_contents": "fwiw, I was mucking around in the rules/views part of the code recently,\nadding support for the CASE construct. I didn't fully understand the\nareas I was modifying, but the changes shouldn't be visible if you\naren't dealing with the new construct. At least that was my\nimpression...\n\nHow long have you been seeing problems? It looks like I made changes on\nDec 14. But all regression tests passed on my system at that time with\nmy source tree, which had some of Vadim's changes already.\n\n - Tom\n",
"msg_date": "Fri, 18 Dec 1998 05:37:15 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] CURRENT: crash in select_view regression test..."
},
{
"msg_contents": "\"Thomas G. Lockhart\" wrote:\n> \n> fwiw, I was mucking around in the rules/views part of the code recently,\n> adding support for the CASE construct. I didn't fully understand the\n> areas I was modifying, but the changes shouldn't be visible if you\n> aren't dealing with the new construct. At least that was my\n> impression...\n> \n> How long have you been seeing problems? It looks like I made changes on\n> Dec 14. But all regression tests passed on my system at that time with\n> my source tree, which had some of Vadim's changes already.\n\nI don't remember. Problems are in readfuncs/outfuncs area...\n\nBTW, right now I get:\n\nvac=> select * from test where x = 1;\nERROR: type id lookup of 0 failed\n\n- from parser...\nI'm trying to fix it now...\n\nVadim\n",
"msg_date": "Fri, 18 Dec 1998 12:43:59 +0700",
"msg_from": "Vadim Mikheev <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] CURRENT: crash in select_view regression test..."
},
{
"msg_contents": "Vadim Mikheev wrote:\n> \n> BTW, right now I get:\n> \n> vac=> select * from test where x = 1;\n> ERROR: type id lookup of 0 failed\n> \n> - from parser...\n> I'm trying to fix it now...\n\nOps, sorry - I had to gmake clean in parser dir after\nadding SET TRANSACTION ISOLATION LEVEL...\n\nVadim\n",
"msg_date": "Fri, 18 Dec 1998 13:01:33 +0700",
"msg_from": "Vadim Mikheev <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] CURRENT: crash in select_view regression test..."
},
{
"msg_contents": "> If I select from the view I get a BE crash.\n\n That was a typo in nodes/read.c that is fixed.\n\n> If I start looking around in ruleutils.c why can't I see \"spirc\"??\n>\n> Breakpoint 2, pg_get_viewdef (rname=0xe01d56a0) at ruleutils.c:277\n> 277 if (spirc != SPI_OK_SELECT)\n> (gdb) print spirc\n> No symbol \"spirc\" in current context.\n> (gdb) list\n> 272 args[1] = PointerGetDatum(name2);\n> 273 nulls[0] = ' ';\n> 274 nulls[1] = ' ';\n> 275 nulls[2] = '\\0';\n> 276 spirc = SPI_execp(plan_getview, args, nulls, 1);\n> 277 if (spirc != SPI_OK_SELECT)\n> 278 elog(ERROR, \"failed to get pg_rewrite tuple for view %s\", rulename);\n> 279 if (SPI_processed != 1)\n> 280 tmp = \"Not a view\";\n> 281 else\n>\n> Jan in particular, any idea what's happening here?\n\n Must take a look at it anyway. Thomas added new node types\n which must be handled there too (CASE).\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, 18 Dec 1998 09:28:39 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] CURRENT: crash in select_view regression test..."
},
{
"msg_contents": "> Must take a look at it anyway. Thomas added new node types\n> which must be handled there too (CASE).\n\nWell, while you are looking... :)\n\nI've enclosed case.sql, which is closer to a complete test of the CASE\nstatement. There are a few queries which are commented out because they\ncrash the backend on my machine. If you happen to see why, that would be\ngreat.\n\nThe crashing statements involve multiple tables. afaik single-table\nqueries work pretty well.\n\nI haven't given up on looking for the problem, but was giving my eyes a\nfew days rest before getting back to it.\n\n - Tom",
"msg_date": "Fri, 18 Dec 1998 15:00:13 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] CURRENT: crash in select_view regression test..."
},
{
"msg_contents": "> Hi,\n> \n> I think I need some help here\n> \n> Taking part of the pg_views rule shows it thinks the views\n> are NOT views.\n> \n> postgres=> SELECT relname AS viewname, pg_get_userbyid(relowner) AS viewowner, \n> pg_get_viewdef(relname) AS definition FROM pg_class WHERE relhasrules;\n> viewname |viewowner|definition\n> ----------+---------+----------\n> pg_user |postgres |Not a view\n> pg_rules |postgres |Not a view\n> pg_views |postgres |Not a view\n> pg_tables |postgres |Not a view\n> pg_indexes|postgres |Not a view\n> (5 rows)\n\nWorks in the current development tree.\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": "Sat, 16 Jan 1999 22:40:55 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] CURRENT: crash in select_view regression test..."
}
] |
[
{
"msg_contents": "Attention PC dealers and Integrators!\n\nGet the following incredible Multimedia functions for just $49 (dealer cost):\n\n- Industry's best DVD Decode \n- 2D/3D Graphics with 6M video RAM and output to TV\n- Real-time MPEG1 compression and Home Movie Editing (software included)\n- Wave table Audio with 3D positioning, \n- 33.3K Fax / Modem, (upgradable to 56K)\n- LAN and Internet based Videoconferencing\n- Telephony (speaker phone, voice mail)\n- See www.st-blazer.com for details\n\nPentium 166 is now priced BELOW $30!. See www.st-blazer.com/6x86\n\nPlease call 888-249-5331 for quote\n\nHappy holidays!\n",
"msg_date": "Fri, 18 Dec 1998 02:34:58",
"msg_from": "\"PCblazer\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Total Multimedia Board & CPU deal"
}
] |
[
{
"msg_contents": "I have made the needed changes for 6.4.1.\n\nWould people please review the attached TODO list, and tell me what TODO\nitems can be removed, and if the attached 6.4.1 changes are correct.\n\nI suspect a few of the TODO items I added after 6.4 was release have\nbeen fixed, but am not sure.\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\nTODO list for PostgreSQL\n========================\nLast updated:\t\tFri Dec 18 00:15:58 EST 1998\n\nCurrent maintainer:\tBruce Momjian ([email protected])\n\nThe most recent version of this document can be viewed at\nthe PostgreSQL WWW site, http://www.postgreSQL.org.\n\nTHE CHANGES FOR 6.4.1 APPEAR AT THE END OF THIS DOCUMENT\n\nA dash(-) marks changes to be in the next release.\n\nDevelopers who have claimed items are:\n-------------------------------------\n\t* Billy is Billy G. Allie <[email protected]>\n\t* Brook is Brook Milligan <[email protected]>\n\t* Bruce is Bruce Momjian<[email protected]>\n\t* Bryan is Bryan Henderson<[email protected]>\n\t* D'Arcy is D'Arcy J.M. Cain <[email protected]>\n\t* Dan is Dan McGuirk <[email protected]>\n\t* Darren is Darren King <[email protected]>\n\t* David is David Hartwig <[email protected]>\n\t* Edmund is Edmund Mergl <[email protected]>\n\t* Goran is Goran Thyni <[email protected]>\n\t* Henry is Henry B. Hotz <[email protected]>\n\t* Jan is Jan Wieck <[email protected]>\n\t* Jun is Jun Kuwamura <[email protected]>\n\t* Maarten is Maarten Boekhold <[email protected]>\n \t* Marc is Marc Fournier <[email protected]>\n \t* Martin is Martin S. Utesch <[email protected]>\n\t* Massimo Dal Zotto <[email protected]>\n\t* Michael is Michael Meskes <[email protected]>\n\t* Oleg is Oleg Bartunov <[email protected]>\n\t* Paul is Paul M. Aoki <[email protected]>\n\t* Peter is Peter T Mount <[email protected]>\n\t* Phil is Phil Thompson <[email protected]>\n\t* Ryan is Ryan Kirkpatrick <[email protected]>\n\t* Soo-Ho Ok <[email protected]>\n\t* Stefan Simkovics <[email protected]>\n\t* Sven is Sven Verdoolaege <[email protected]>\n\t* Tatsuo is Tatsuo Ishii <[email protected]>\n\t* Tom is Tom Lane <[email protected]>\n\t* Thomas is Thomas Lockhart <[email protected]>\n\t* TomH is Tom I Helbekkmo <[email protected]>\n\t* Vadim is \"Vadim B. Mikheev\" <[email protected]>\n\nRELIABILITY\n-----------\n* Overhaul mdmgr/smgr to fix double unlinking and double opens, cleanup\n* Overhaul bufmgr/lockmgr/transaction manager\n* Remove EXTEND?\n* Can lo_export()/lo_import() read/write anywhere, causing a security problem?\n* Tables that start with xinv confused to be large objects\n* Two and three dimmensional arrays display improperly, missing {}\n* GROUP BY in INSERT INTO table SELECT * FROM table2 fails\n* Prevent auto-table reference, like SELECT table.col WHERE col = 3 (?)\n* SELECT * FROM table WHERE int4_column = '1' fails\n* SELECT a[1] FROM test fails, it needs test.a[1]\n* UPDATE table SET table.value = 3 fails\n* User who can create databases can modify pg_database table\n* optimizer memory exhaustion with many OR's\n* elog() does not free all its memory(Jan)\n* views on subselects fail\n* disallow inherited columns with the same name as new columns\n* recover or force failure when disk space is exhausted\n* default char() value not to full length crashes server on some OS's\n* allow UPDATE using aggregate to affect all rows, not just one\n* computations in views fail:\n\tcreate view test as select usesysid * usesysid from pg_shadow;\n* views containing aggregates sometimes fail(Jan)\n* ALTER TABLE ADD COLUMN does not honor DEFAULT, add CONSTRAINT\n* SELECT DISTINCT i FROM dtest ORDER BY j generates strange output\n* fix memory leak in aborted transactions\n* array index references without table name cause problems\n* aggregates on array indexes crash backend\n* subqueries containing HAVING return incorrect results\n* DEFAULT handles single quotes in value by requiring too many quotes\n* make CURSOR valid even after you hit end of cursor\n* views with spaces in view name fail when referenced\n* plpgsql does not handle quoted mixed-case identifiers\n* do not allow bpchar column creation without length\n\nENHANCEMENTS\n------------\n* Replace table-level locking with row or page-level locking(Vadim)\n* Transaction log, so re-do log can be on a separate disk\n* Allow transaction commits with rollback with no-fsync performance\n* More access control over who can create tables and access the database\n* Add full ANSI SQL capabilities\n\t* add OUTER joins, left and right (Thomas)\n\t* add INTERSECTS, SUBTRACTS(Stephan)\n\t* add temporary tables\n\t* add sql3 recursive unions\n\t* add the concept of dataspaces\n\t* add DECIMAL, NUMERIC, DOUBLE PRECISION, BIT, BIT VARYING\n \t* NCHAR (as distinguished from ordinary varchar),\n\t* DOMAIN capability\n* Allow compression of large fields or a compressed field type\n* Fix the rules system(Jan,Soo-Ho)\n\t* add CONSTRAINT\n* Full set of text operations and functions\n\t* word searches, concat,max() on text, char\n* Large objects\n\t* Fix large object mapping scheme, own reltype(Peter)\n\t* Allow large text type to use large objects(Peter)\n\t* not to stuff everything as files in a single directory\n\t* Fix large object memory leaks\n\t* delete orphaned large objects\n* Better interface for adding to pg_group\n* Make MONEY/DECIMAL have a defined precision\n* Fix tables >2G, or report error when 2G size reached\n\t(fix lseek()/off_t, mdextend()/RELSEG_SIZE)\n* Add REGEX internationalization\n* allow row re-use without vacuum, maybe?(Vadim)\n* Add word index for text fields, maybe with trigrams, i.e.:\n * ' (cat | dog) & ! fox ' meaning text has cat aor dog, but not fox\n* Populate backend status area and write program to dump status data\n* Add ALTER TABLE DROP/ALTER COLUMN feature\n* Allow INSERT INTO ... SELECT to convert column types\n* Add syslog functionality(Marc)\n* Add STDDEV/VARIANCE() function for standard deviation computation/variance\n* add UNIQUE capability to non-btree indexes\n* make number of backends a config parameter, storage/sinvaladt.h:MaxBackendId\n* certain indexes will not shrink, i.e. oid indexes with many inserts\n* make NULL's come out at the beginning or end depending on the ORDER BY direction\n* change the library/backend interface to use network byte order\n* Restore unused oid's on backend exit if no one else has gotten oids\n* remove non-standard types from the system, and make them loadable\n* have UPDATE/DELETE clean out indexes\n* allow WHERE restriction on ctid\n* allow pg_descriptions when creating types, tables, columns, and functions\n* Fix compile and security of Kerberos/GSSAPI code\n* Allow psql to print nulls as distinct from \"\"(?)\n* Allow variable casts with BETWEEN 'today'::asbtime AND 'today'::abstime\n* Allow INSERT INTO ... SELECT ... FROM view to work\n* Make VACUUM on database not lock pg_class\n* Make VACUUM ANALYZE only use a readlock\n* Allow cursors to be DECLAREd/OPENed/CLOSEed outside transactions\n* Allow installation data block size and max tuple size configuration(Darren)\n* Allow views on a UNION\n* Allow DISTINCT on view\n* Allow views of aggregate columns\n* Allow variable block sizes(Darren)\n* System tables are now more update-able from SQL(Jan)\n* New pg_shadow file, pg_user is now a view of pg_shadow(Jan)\n* Allow flag to control COPY input/output of NULLs\n* Allow CLUSTER on all tables at once, and improve CLUSTER\n* Change all references of Postgres to PostgreSQL, including binary names\n* Add ELOG_TIMESTAMPS to elog()(?)\n* Change LOCK tablename to LOCK TABLE tablename(?)\n* Allow max tuple length to be changed(Darren)\n* Have psql with no database name not connect to username as default(?)\n* Allow subqueries in target list\n* Granting permissions to another user looses permissions for the owner\n* Allow queries across multiple databases\n* Add replication of distributed databases\n* Allow table destruction/alter to be rolled back\n* Add pg_attribute.atttypmod/Resdom->restypmod to PGresult structure\n* Generate error on CREATE OPERATOR of ~~, ~ and and ~*\n* Allow constraint NULL just as we honor NOT NULL\n* Add version number in startup banners for psql and postmaster\n* Restructure storing of GRANT permission information to allow +-=\n* allow psql \\copy to allow delimiters\n* allow international error message support and add error codes\n* allow usernames with dashes(GRANT fails)\n* add a function to return the last inserted oid, for use in psql scripts\n* allow creation of functional indexes to use default types\n* put sort files, large objects in their on directory\n* CREATE VIEW myview (name) AS SELECT lname FROM wages fails\n* do autocommit so always in a transaction block\n* add SIMILAR TO to allow character classes, 'pg_[a-c]%'\n* multi-verion concurrency control(Vadim)\n* improve reporting of syntax errors by showing location of error in query\n* allow chaining of pages to allow >8k tuples\n* no min/max for oid type\n* remove un-needed conversion functions\n* redesign the function call interface to handle NULLs better(Jan)\n* permissions on indexes - prevent them?\n* allow multiple generic operators in expressions without the use of parentheses\n* document/trigger/rule so changes to pg_shadow create pg_pwd\n* improve group handling\n* generate postmaster pid file and remove flock/fcntl lock code\n* improve PRIMARY KEY handling\n* add ability to specifiy location of lock/socket files\n* psql \\d on index with char()/varchar() fields shows improper length\n* disallow LOCK outside a transaction, change message to LOCK instead of DELETE\n* Fix roundoff problems in \"cash\" datatype\n* fix any sprintf() overruns\n* add portable vsnprintf()\n* auto-destroy sequence on SERIAL removal\n* CREATE TABLE inside aborted transaction causes stray table file\n* allow user to define char1 column\n\nPERFORMANCE\n-----------\n* Use indexes in ORDER BY for restrictive data sets, min(), max()\n* Allow LIMIT ability on single-table queries that have no ORDER BY or \n\ta matching index\n* Pull requested data directly from indexes, bypassing heap data\n* Prevent psort() usage when query already using index matching ORDER BY\n* Optimizing disjunctive queries\n* Fix bushy-plans\n* Prevent fsync in SELECT-only queries\n* Cache most recent query plan(s?)\n* Shared catalog cache, reduce lseek()'s by caching table size in shared area\n* Allow compression of log and meta data\n* Add FILLFACTOR to index creation\n* update pg_statistic table to remove operator column\n* make index creation use psort code, because it is now faster(Vadim)\n* remove fork()/exec() of backend and make it just fork()\n* Add base table name to \\d index\n* Allow char() not to use variable-sized header to reduce disk size\n* Do async I/O to do better read-ahead of data\n* Fix optmizer problem with self-table joins\n* Fix memory exhaustion when using many OR's\n* Use spin locks only on multi-CPU systems, yield CPU instead\n* Get faster regex() code from Henry Spencer <[email protected]>\n\twhen it is available\n* use mmap() rather than SYSV shared memory(?)\n* use index to restrict rows returned by multi-key index when used with\n\tnon-consecutive keys or OR clauses, so fewer heap accesses\n* use index with constants on functions\n\nDOCUMENTATION\n-------------\n* Update usermanual source(many)\n* added features used in grammer but not in docs, like :: and CAST\n* Add keyword list to documentation, already in /tools\n* Add 'man pgsql' to show all manual page names\n* Add use of 'const' for varibles in source tree\n\n\n=============================================================================\n\nCHANGES IN THE 6.4.1 RELEASE\n----------------------------\nAdd pg_dump -N flag to force double quotes around identifiers. This is\n\tthe default\nFix for NOT in where clause causing crash(Bruce)\nEXPLAIN VERBOSE coredump fix(Vadim)\nFix shared-library problems on Linux\nFix test for table existance to allow mixed-case and whitespace in\n\tthe table name\nFix a couple of pg_dump bugs\nFix for creating tables with constraints when table name is mixed-case(Billy)\nConfigure matches template/.similar entries better(Tom)\nChange builtin function names from SPI_* to spi_*\nOR WHERE clause fix(Vadim)\nFixes for mixed-case table names(Billy)\ncontrib/linux/postgres.init.csh/sh fix\nlibpq memory overrun fix\nSunOS fixes(Tom)\nChange exp() behavior to generate error on underflow\npg_dump fixes for memory leak, inheritance constraints, layout change\nupdate pgaccess to 0.93\nFix prototype for 64-bit platforms\nMulti-byte fixes(Tatsuo)\nNew ecpg man page\nFix memory overruns(Tatsuo)\nFix for lo_import() crash(Bruce)\nBetter search for install program(Tom)\nTimezone fixes(Tom)\nHPUX fixes(Tom)\nUse implicit type coersion for matching DEFAULT values\nAdd routines to help with single-byte (internal) character type(Thomas)\nCompilation of libpq for Win32 fixes(Magnus)\nUpgrade to PyGreSQL 2.2(D'Arcy)",
"msg_date": "Fri, 18 Dec 1998 00:33:43 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Upgrades for 6.4.1"
},
{
"msg_contents": "> * Two and three dimmensional arrays display improperly, missing {}\n\nMy simple test case shows as many brackets as I expect. Can someone\nremember what this was about? Or at least take the extra \"m\" out of\n\"dimensional\"? :)\n\n> * SELECT * FROM table WHERE int4_column = '1' fails\n\npostgres=> select * from x where i = '1';\ni\n-\n1\n(1 row)\n\nWorks now, no? The new type coersion stuff...\n\n> * default char() value not to full length crashes server on some OS's\n\npostgres=> create table x (c char(10) default 'abc');\nCREATE\npostgres=> insert into x default values;\nINSERT 901452 1\npostgres=> select * from x;\nc\n---\nabc\n(1 row)\n\nI distinctly recall fixing this, or watching someone else do it...\n\n> * SELECT DISTINCT i FROM dtest ORDER BY j generates strange output\n\nIn my simple test case, it orders by j, then only shows i. Is that\nstrange? What did it used to do? The current behavior makes sense, if it\nshould be allowed at all...\n\n> * views with spaces in view name fail when referenced\n\nStill a problem.\n\n> * plpgsql does not handle quoted mixed-case identifiers\n\nDitto.\n\n> * do not allow bpchar column creation without length\n\n? char is now the same as char(1), which should solve this, if that's\nwhat it means...\n\n> ENHANCEMENTS\n> ------------\n> * Add full ANSI SQL capabilities\n> * add OUTER joins, left and right (Thomas)\n> * add INTERSECTS, SUBTRACTS(Stephan)\n> * add temporary tables\n> * add sql3 recursive unions\n> * add the concept of dataspaces\n> * add DECIMAL, NUMERIC, DOUBLE PRECISION, BIT, BIT VARYING\n\nWe've got DOUBLE PRECISION, DECIMAL, and NUMERIC (the latter two are\nbrain-damaged though)\n\n> * add CONSTRAINT\n\nAlready here, from Vadim.\n\n> * Full set of text operations and functions\n> * word searches, concat,max() on text, char\n\nNot sure about \"word searches\", others are done. v6.4.1 will require a\ndump/reload to find min/max for string types. In fact, can someone test\nto make sure it is ok that I added these to the pg_aggregate system\ntable? And a couple of support functions to pg_proc? Since they weren't\navailable before, and since everything else still works, I thought it\nmight be OK to include them. But I can rip them out of the system tables\nif that would be best for v6.4.1.\n\n> * Add word index for text fields, maybe with trigrams, i.e.:\n> * ' (cat | dog) & ! fox ' meaning text has cat aor dog, but not fox\n\nHow is this different from the \"word searches\" above?\n\n> * Allow INSERT INTO ... SELECT to convert column types\n\nProbably does. The new type coersion stuff :)\n\n> * remove non-standard types from the system, and make them loadable\n\nThis made the ToDo list, but I'm not sure we want to do it. Native types\nstill have a few advantages over loadable types, and our types are one\nof our strongest features...\n\n> * Allow variable casts with BETWEEN 'today'::asbtime AND 'today'::abstime\n\npostgres=> select 'yes' where 'now'::datetime \n between 'today'::datetime and 'tomorrow'::datetime;\n?column?\n--------\nyes\n(1 row)\n\nWhat did this item mean??\n\n> * Change all references of Postgres to PostgreSQL, including binary \n> names\n\nI've standardized all Postgres docs to use \"Postgres\" within the body,\nand \"PostgreSQL\" for titles and in the introduction.\n\n> * Change LOCK tablename to LOCK TABLE tablename(?)\n\nBoth syntaxes are now legal and accepted.\n\n> * Generate error on CREATE OPERATOR of ~~, ~ and and ~*\n\n? Because gram.y mucks around with them?\n\n> * Allow constraint NULL just as we honor NOT NULL\n\nFundamental yacc problem with this as I recall. Gives rise to\nshift/reduce problems since it is ambiguous with other uses of \"NULL\" in\nthe same area.\n\n> * do autocommit so always in a transaction block\n\nThis is already the case, right? Every statement has an implicit\nbegin/end around it, unless it is within an explicit begin/end. Or does\nthis mean something else?\n\n> * no min/max for oid type\n\nFixed. New type coersion stuff.\n\n> * remove un-needed conversion functions\n\n? An explicit conversion function is always/usually faster than one\nwhich requires multiple steps.\n\n> DOCUMENTATION\n> -------------\n> * Add keyword list to documentation, already in /tools\n\nDone. In the User's Guide chapter on \"Syntax\".\n\n> CHANGES IN THE 6.4.1 RELEASE\n> ----------------------------\n> Add pg_dump -N flag to force double quotes around identifiers. This is\n> the default\n\n(Thomas) in case you have room :)\n\n> Fix test for table existance to allow mixed-case and whitespace in\n> the table name\n\nDitto. At least when combined with Billy's fixes to make a complete set.\n\n> Fix for creating tables with constraints when table name is mixed-case(Billy)\n> Fixes for mixed-case table names(Billy)\n\nThese all sure sound similar...\n\n> contrib/linux/postgres.init.csh/sh fix\n\n(Thomas)\n\n> Change exp() behavior to generate error on underflow\n\n(Tom), right?\n\n> Timezone fixes(Tom)\n\n? Not sure we see an improvement yet, if this is the date->datetime\nconversion problem reported by Oleg.\n\n> Use implicit type coersion for matching DEFAULT values\n\n(Thomas) but wasn't this mentioned above somewhere? Maybe not...\n\n - Tom\n",
"msg_date": "Fri, 18 Dec 1998 07:08:32 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "> > * Two and three dimmensional arrays display improperly, missing {}\n> \n> My simple test case shows as many brackets as I expect. Can someone\n> remember what this was about? Or at least take the extra \"m\" out of\n> \"dimensional\"? :)\n\nBug report attached. It is an old one.\n\n> \n> > * SELECT * FROM table WHERE int4_column = '1' fails\n> \n> postgres=> select * from x where i = '1';\n> i\n> -\n> 1\n> (1 row)\n> \n> Works now, no? The new type coersion stuff...\n\nItem removed.\n\n> \n> > * default char() value not to full length crashes server on some OS's\n> \n> postgres=> create table x (c char(10) default 'abc');\n> CREATE\n> postgres=> insert into x default values;\n> INSERT 901452 1\n> postgres=> select * from x;\n> c\n> ---\n> abc\n> (1 row)\n> \n> I distinctly recall fixing this, or watching someone else do it...\n\nRemoved.\n\n> \n> > * SELECT DISTINCT i FROM dtest ORDER BY j generates strange output\n> \n> In my simple test case, it orders by j, then only shows i. Is that\n> strange? What did it used to do? The current behavior makes sense, if it\n> should be allowed at all...\n\nRemoved.\n\n> \n> > * views with spaces in view name fail when referenced\n> \n> Still a problem.\n\nOk.\n\n> \n> > * plpgsql does not handle quoted mixed-case identifiers\n> \n> Ditto.\n\nOk.\n\n> \n> > * do not allow bpchar column creation without length\n> \n> ? char is now the same as char(1), which should solve this, if that's\n> what it means...\n\nTry:\n\tcreate table x(y bpchar);\n\nand try pg_dumping it.\n\n> \n> > ENHANCEMENTS\n> > ------------\n> > * Add full ANSI SQL capabilities\n> > * add OUTER joins, left and right (Thomas)\n> > * add INTERSECTS, SUBTRACTS(Stephan)\n> > * add temporary tables\n> > * add sql3 recursive unions\n> > * add the concept of dataspaces\n> > * add DECIMAL, NUMERIC, DOUBLE PRECISION, BIT, BIT VARYING\n> \n> We've got DOUBLE PRECISION, DECIMAL, and NUMERIC (the latter two are\n> brain-damaged though)\n\nRemoved.\n\n> \n> > * add CONSTRAINT\n> \n> Already here, from Vadim.\n\nRemoved.\n\n> \n> > * Full set of text operations and functions\n> > * word searches, concat,max() on text, char\n> \n> Not sure about \"word searches\", others are done. v6.4.1 will require a\n> dump/reload to find min/max for string types. In fact, can someone test\n> to make sure it is ok that I added these to the pg_aggregate system\n> table? And a couple of support functions to pg_proc? Since they weren't\n> available before, and since everything else still works, I thought it\n> might be OK to include them. But I can rip them out of the system tables\n> if that would be best for v6.4.1.\n\nRemoved.\n\n> \n> > * Add word index for text fields, maybe with trigrams, i.e.:\n> > * ' (cat | dog) & ! fox ' meaning text has cat aor dog, but not fox\n> \n> How is this different from the \"word searches\" above?\n\nRemoved.\n\n> \n> > * Allow INSERT INTO ... SELECT to convert column types\n> \n> Probably does. The new type coersion stuff :)\n\nRemoved.\n\n> \n> > * remove non-standard types from the system, and make them loadable\n> \n> This made the ToDo list, but I'm not sure we want to do it. Native types\n> still have a few advantages over loadable types, and our types are one\n> of our strongest features...\n\nRemoved.\n\n> \n> > * Allow variable casts with BETWEEN 'today'::asbtime AND 'today'::abstime\n> \n> postgres=> select 'yes' where 'now'::datetime \n> between 'today'::datetime and 'tomorrow'::datetime;\n> ?column?\n> --------\n> yes\n> (1 row)\n> \n> What did this item mean??\n\nSome problem with OR gram.y processing. Fixed, I guess.\n\n> \n> > * Change all references of Postgres to PostgreSQL, including binary \n> > names\n> \n> I've standardized all Postgres docs to use \"Postgres\" within the body,\n> and \"PostgreSQL\" for titles and in the introduction.\n\nRemoved.\n\n> \n> > * Change LOCK tablename to LOCK TABLE tablename(?)\n> \n> Both syntaxes are now legal and accepted.\n\nRemoved.\n\n> \n> > * Generate error on CREATE OPERATOR of ~~, ~ and and ~*\n> \n> ? Because gram.y mucks around with them?\n\nOk.\n\n> \n> > * Allow constraint NULL just as we honor NOT NULL\n> \n> Fundamental yacc problem with this as I recall. Gives rise to\n> shift/reduce problems since it is ambiguous with other uses of \"NULL\" in\n> the same area.\n\nOk.\n\n> \n> > * do autocommit so always in a transaction block\n> \n> This is already the case, right? Every statement has an implicit\n> begin/end around it, unless it is within an explicit begin/end. Or does\n> this mean something else?\n\nNot the same. Withouth autocommit on, you start in a transaction, and\nhave to issue a commit to commit the changes, and you are back in a new\ntransaction. Ingres SQL does this.\n\n> \n> > * no min/max for oid type\n> \n> Fixed. New type coersion stuff.\n\nRemoved.\n\n> \n> > * remove un-needed conversion functions\n> \n> ? An explicit conversion function is always/usually faster than one\n> which requires multiple steps.\n\nAdded word 'where appropriate'. We just have tons of them. Would be\nnice to pair them down a little.\n\n> \n> > DOCUMENTATION\n> > -------------\n> > * Add keyword list to documentation, already in /tools\n> \n> Done. In the User's Guide chapter on \"Syntax\".\n\nRemoved.\n\n> \n> > CHANGES IN THE 6.4.1 RELEASE\n> > ----------------------------\n> > Add pg_dump -N flag to force double quotes around identifiers. This is\n> > the default\n> \n> (Thomas) in case you have room :)\n\n\nOk.\n\n> \n> > Fix test for table existance to allow mixed-case and whitespace in\n> > the table name\n> \n> Ditto. At least when combined with Billy's fixes to make a complete set.\n\nOk. I loose the committers name in the script I wrote, and I am not\nsure who did it. We also commit other's patches, so even knowing the\ncommitters name sometimes doesn't help. I know my fixes, however. :-)\n\n> \n> > Fix for creating tables with constraints when table name is mixed-case(Billy)\n> > Fixes for mixed-case table names(Billy)\n> \n> These all sure sound similar...\n\nRemoved.\n\n> \n> > contrib/linux/postgres.init.csh/sh fix\n> \n> (Thomas)\n\nAdded\n\n> \n> > Change exp() behavior to generate error on underflow\n> \n> (Tom), right?\nOK.\n\n> \n> > Timezone fixes(Tom)\n> \n> ? Not sure we see an improvement yet, if this is the date->datetime\n> conversion problem reported by Oleg.\n\nWe'll see.\n\n> \n> > Use implicit type coersion for matching DEFAULT values\n> \n> (Thomas) but wasn't this mentioned above somewhere? Maybe not...\n\nNot mentioned. Updated RELEASE tree and web page with new TODO copy.\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",
"msg_date": "Fri, 18 Dec 1998 02:30:42 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "> Bruce Momjian wrote:\n> > >\n> > > > ENHANCEMENTS\n> > > > ------------\n> > > > * Add full ANSI SQL capabilities\n> > > > * add OUTER joins, left and right (Thomas)\n> > > > * add INTERSECTS, SUBTRACTS(Stephan)\n> > > > * add temporary tables\n> > > > * add sql3 recursive unions\n> > > > * add the concept of dataspaces\n> > > > * add DECIMAL, NUMERIC, DOUBLE PRECISION, BIT, BIT VARYING\n> > >\n> > > We've got DOUBLE PRECISION, DECIMAL, and NUMERIC (the latter two are\n> > > brain-damaged though)\n> > \n> > Removed.\n> \n> If it's brain-damaged should'nt it be justr Replaced instead of Removed \n\nWe just need to add better precision to the type.\n\n> \n> > \n> > >\n> > > > * add CONSTRAINT\n> > >\n> > > Already here, from Vadim.\n> > Removed.\n> > \n> \n> AFAIK, CONSTRAINTS ar currently available only at table create time ?\n> \n> So once you have (or have.nt) them, you must stay with it ;(\n> \n> creating/dropping constraints should be made to work in ALTER TABLE as\n> well\n\nAlready on list:\n\n\t* ALTER TABLE ADD COLUMN does not honor DEFAULT, add CONSTRAINT\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, 18 Dec 1998 11:07:17 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "Bruce Momjian wrote:\n> >\n> > > ENHANCEMENTS\n> > > ------------\n> > > * Add full ANSI SQL capabilities\n> > > * add OUTER joins, left and right (Thomas)\n> > > * add INTERSECTS, SUBTRACTS(Stephan)\n> > > * add temporary tables\n> > > * add sql3 recursive unions\n> > > * add the concept of dataspaces\n> > > * add DECIMAL, NUMERIC, DOUBLE PRECISION, BIT, BIT VARYING\n> >\n> > We've got DOUBLE PRECISION, DECIMAL, and NUMERIC (the latter two are\n> > brain-damaged though)\n> \n> Removed.\n\nIf it's brain-damaged should'nt it be justr Replaced instead of Removed \n\n> \n> >\n> > > * add CONSTRAINT\n> >\n> > Already here, from Vadim.\n> Removed.\n> \n\nAFAIK, CONSTRAINTS ar currently available only at table create time ?\n\nSo once you have (or have.nt) them, you must stay with it ;(\n\ncreating/dropping constraints should be made to work in ALTER TABLE as\nwell\n\n\n------------------\nHannu\n",
"msg_date": "Fri, 18 Dec 1998 18:08:33 +0200",
"msg_from": "Hannu Krosing <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": " Would people please review the attached TODO list, and tell me what TODO\n items can be removed, and if the attached 6.4.1 changes are correct.\n\nIt looks from your list like a couple of patches I sent in have not\nmade it in for one reason or another, though I haven't checked the\nsource yet but never saw any acknowledgement.\n\nHere are descriptions:\n\n- the first patch is just to preven listing the perl warning in the\n make output unless it is actually emitted by the make. this may\n prevent new users from being confused by the warning in their output\n when in fact is is just echoing part of a command not actually run.\n\n- the second patch (to 2 files) just enables building/installing\n pgaccess if TCL and TK are available. a Makefile is created to do\n this, but you may wish to change the heading information in it since\n I just copied another Makefile to use as a template.\n\nI hope these make it into 6.4.1.\n\nCheers,\nBrook\n\n===========================================================================\n$NetBSD$\n\n--- interfaces/Makefile.orig\tTue Oct 27 14:27:36 1998\n+++ interfaces/Makefile\tFri Nov 6 23:12:08 1998\n@@ -48,11 +48,11 @@\n perl5/Makefile: perl5/Makefile.PL\n \tcd perl5 && perl Makefile.PL\n \n-install-perl5:\n+install-perl5: perl5/Makefile\n \t$(MAKE) -C perl5 clean\n \tcd perl5 && POSTGRES_HOME=\"$(POSTGRESDIR)\" perl Makefile.PL\n \t$(MAKE) -C perl5 all\n-\tif [ -w `sed -n -e 's/^ *SITELIBEXP *= *//p' perl5/Makefile` ]; then \\\n+\t@if [ -w `sed -n -e 's/^ *SITELIBEXP *= *//p' perl5/Makefile` ]; then \\\n \t\t$(MAKE) $(MFLAGS) -C perl5 install; \\\n \t\trm -f perl5/Makefile; \\\n \telse \\\n===========================================================================\n$NetBSD$\n\n--- bin/Makefile.orig\tSat Jul 25 22:31:08 1998\n+++ bin/Makefile\tMon Nov 9 08:35:00 1998\n@@ -25,6 +25,9 @@\n # TCL/TK programs\n #\n ifeq ($(USE_TCL), true)\n+ifeq ($(USE_TK), true)\n+\tDIRS += pgaccess\n+endif\n \tDIRS += pgtclsh\n endif\n \n$NetBSD$\n\n--- bin/pgaccess/Makefile.orig\tFri Nov 6 23:43:38 1998\n+++ bin/pgaccess/Makefile\tFri Nov 6 23:46:26 1998\n@@ -0,0 +1,21 @@\n+#-------------------------------------------------------------------------\n+#\n+# Makefile\n+# Makefile for pgaccess\n+#\n+# Copyright (c) 1994, Regents of the University of California\n+#\n+# IDENTIFICATION\n+# $Header: /usr/local/cvsroot/pgsql/src/interfaces/libpgtcl/Makefile.in,v 1.33 1998/10/19 00:00:41 tgl Exp $\n+#\n+#-------------------------------------------------------------------------\n+\n+SRCDIR= ../..\n+include $(SRCDIR)/Makefile.global\n+\n+.PHONY: all clean\n+all clean: \n+\n+.PHONY: install\n+install: pgaccess.tcl\n+\t$(INSTALL) $(INSTL_EXE_OPTS) pgaccess.tcl $(BINDIR)/pgaccess\n===========================================================================\n",
"msg_date": "Fri, 18 Dec 1998 10:40:28 -0700 (MST)",
"msg_from": "Brook Milligan <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "Applied.\n\nBoth of them somehow were not applied. Not sure how I missed them.\n\nSorry and thanks.\n\n> Would people please review the attached TODO list, and tell me what TODO\n> items can be removed, and if the attached 6.4.1 changes are correct.\n> \n> It looks from your list like a couple of patches I sent in have not\n> made it in for one reason or another, though I haven't checked the\n> source yet but never saw any acknowledgement.\n> \n> Here are descriptions:\n> \n> - the first patch is just to preven listing the perl warning in the\n> make output unless it is actually emitted by the make. this may\n> prevent new users from being confused by the warning in their output\n> when in fact is is just echoing part of a command not actually run.\n> \n> - the second patch (to 2 files) just enables building/installing\n> pgaccess if TCL and TK are available. a Makefile is created to do\n> this, but you may wish to change the heading information in it since\n> I just copied another Makefile to use as a template.\n> \n> I hope these make it into 6.4.1.\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, 18 Dec 1998 12:55:00 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "Constantin, should you add this Makefile to the pgaccess distribution?\n\nI am not sure. It is in the PostgreSQL source tree.\n\n> \n> --- bin/pgaccess/Makefile.orig\tFri Nov 6 23:43:38 1998\n> +++ bin/pgaccess/Makefile\tFri Nov 6 23:46:26 1998\n> @@ -0,0 +1,21 @@\n> +#-------------------------------------------------------------------------\n> +#\n> +# Makefile\n> +# Makefile for pgaccess\n> +#\n> +# Copyright (c) 1994, Regents of the University of California\n> +#\n> +# IDENTIFICATION\n> +# $Header: /usr/local/cvsroot/pgsql/src/interfaces/libpgtcl/Makefile.in,v 1.33 1998/10/19 00:00:41 tgl Exp $\n> +#\n> +#-------------------------------------------------------------------------\n> +\n> +SRCDIR= ../..\n> +include $(SRCDIR)/Makefile.global\n> +\n> +.PHONY: all clean\n> +all clean: \n> +\n> +.PHONY: install\n> +install: pgaccess.tcl\n> +\t$(INSTALL) $(INSTL_EXE_OPTS) pgaccess.tcl $(BINDIR)/pgaccess\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": "Fri, 18 Dec 1998 12:59:42 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "Bruce Momjian wrote:\n> \n> Constantin, should you add this Makefile to the pgaccess distribution?\n> \n> I am not sure. It is in the PostgreSQL source tree.\n\nYes. I will install pgaccess.tcl as executable pgaccess and it seems ok.\nI will add it to my usual tar.gz\n\n-- \nConstantin Teodorescu\nFLEX Consulting Braila, ROMANIA\n",
"msg_date": "Fri, 18 Dec 1998 20:40:04 +0200",
"msg_from": "Constantin Teodorescu <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "> > > > * add DECIMAL, NUMERIC, DOUBLE PRECISION, BIT, BIT VARYING\n> > >\n> > > We've got DOUBLE PRECISION, DECIMAL, and NUMERIC (the latter two are\n> > > brain-damaged though)\n> >\n> > Removed.\n>\n> If it's brain-damaged should'nt it be justr Replaced instead of Removed\n\n NUMERIC and DECIMAL are identical, but should be different\n from INTEGER (what they are in Postgres for now).\n\n All databases share the definition\n\n NUMERIC [(precision [, scale] )]\n\n And NUMERIC (NUMBER in Oracle :-) is defined as a datatype\n that uses exact representation of arbitrary precise numbers.\n But everyone has different legal value ranges and defaults\n for it. The ranges for the precision (number of total\n digits) varies from 38 (Oracle) to over 1000 sometimes.\n\n I'll hack around a little on it to see what's possible for\n us.\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, 18 Dec 1998 19:50:45 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "Since I need to learn PosgreSQL, I'd like to help \nwith documentation. It's the best way to learn... \n\nTo start, can't seem to get the doc/src/sgml \nto compile. What version of jade do I need \nand what version of the DocBook DTD. Is there \na link where I can download?\n\nAfter I am able to re-build the documentation,\ncould I get access to the CVS server so I can\nupdate documentation?\n\nMy long term goal is to put documentation\nin the database itself, and to provide tools\nto import/export SGML documents to PostgreSQL\nand to crate a \"document server\" that\nis accessable from the web.\n\nBest,\n\nClark\n",
"msg_date": "Sat, 19 Dec 1998 00:41:41 +0000",
"msg_from": "Clark Evans <[email protected]>",
"msg_from_op": false,
"msg_subject": "Help with Documentation?"
},
{
"msg_contents": "I wrote:\n\n> NUMERIC and DECIMAL are identical, but should be different\n> from INTEGER (what they are in Postgres for now).\n>\n> All databases share the definition\n>\n> NUMERIC [(precision [, scale] )]\n>\n> [...]\n>\n> I'll hack around a little on it to see what's possible for\n> us.\n\n Easy - and the type coersion stuff helps alot!\n\n Up to now (1.5 hours hacking) I have a NUMERIC type that\n handles '+' completely, including overflow checks and\n rounding.\n\n Subtract will be trivial, because the core funcitons already\n exist for the add. Multiply and divide will take some time\n and then there are all the comparision operators, an operator\n class and all the type conversion (int<->numeric,\n float<->numeric ...).\n\n Another tricky part I expect when telling the parser that a\n literal NUMERIC must not be enclosed into single quotes.\n\n The only ugly thing is, that I needed to put the precision\n AND the scale together into atttypmod (I limited both to 99\n for now and put them as prec<<8 | scale into). So pg_dump\n will need attention later.\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": "Sat, 19 Dec 1998 02:01:16 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "> The only ugly thing is, that I needed to put the precision\n> AND the scale together into atttypmod (I limited both to 99\n> for now and put them as prec<<8 | scale into). So pg_dump\n> will need attention later.\n\nAtttypmod was expanded to 32 bits for exactly that use. I recommend\n16-bits for each option.\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, 18 Dec 1998 20:39:10 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "Bruce Momjian wrote:\n> \n> --ELM913959223-29577-0_\n> Content-Type: text/plain; charset=US-ASCII\n> Content-Transfer-Encoding: 7bit\n> \n> I have made the needed changes for 6.4.1.\n> \n> Would people please review the attached TODO list, and tell me what TODO\n> items can be removed, and if the attached 6.4.1 changes are correct.\n\nHi Bruce,\n\nDid my patches to psql to correct handle mixed case table names make it into 6.4.1. I can re-send the patches if needed.\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": "Fri, 18 Dec 1998 21:31:49 -0500",
"msg_from": "\"Billy G. Allie\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1 "
},
{
"msg_contents": "> Bruce Momjian wrote:\n> > \n> > --ELM913959223-29577-0_\n> > Content-Type: text/plain; charset=US-ASCII\n> > Content-Transfer-Encoding: 7bit\n> > \n> > I have made the needed changes for 6.4.1.\n> > \n> > Would people please review the attached TODO list, and tell me what TODO\n> > items can be removed, and if the attached 6.4.1 changes are correct.\n> \n> Hi Bruce,\n> \n> Did my patches to psql to correct handle mixed case table names make it into 6.4.1. I can re-send the patches if needed.\n\nYes, I believe I applied those patches. The list of changes includes\nseveral mixed-case fixes.\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, 18 Dec 1998 21:34:04 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "> Since I need to learn PosgreSQL, I'd like to help\n> with documentation.\n\nGreat. Welcome.\n\n> To start, can't seem to get the doc/src/sgml\n> to compile. What version of jade do I need\n> and what version of the DocBook DTD. Is there\n> a link where I can download?\n\nLook in Appendix A of the \"integrated\" postgres docs (postgres.html). It\ndescribes two methods of installation; from sources and from RPM (for\nlinux). A few days ago we re-discovered how to install and run on\nFreeBSD using available \"ports\". Linux and FreeBSD are easy; from\nsources is more difficult. A detailed description for FreeBSD is in the\nworks, but the info can be posted now I would think if that is your\nplatform.\n\nI'm planning on implementing \"automatic builds\" on the current docs tree\non postgresql.org so that fresh working docs are available routinely.\nAll the pieces are in place now, just need to put the framework in\nplace.\n\nAs an aside, it is possible to contribute docs without having the\nability to build the output. For the Postgres project, it doesn't matter\nmuch, but imho the jade/docbook toolset is so powerful and useful that\nit would be a good opportunity for you to learn a tool for other\nprojects too.\n\n> After I am able to re-build the documentation,\n> could I get access to the CVS server so I can\n> update documentation?\n\nAt the moment, a relatively limited number of contributors have direct\n*write* access to the CVS server. Typically, one gets cvs access after\nhammering the current committers hard enough with patches that they get\ntired of handling them. Everyone has CVSup and anon-CVS access to read\nthe server.\n\nAs the de facto current docs coordinator, I'll make sure contributions\nare used, and would welcome any suggestions (especially backed with\nwork! :) on document structure and style.\n\nbtw, no document contributions or updates have ever been rejected, lost,\nor unused.\n\n> My long term goal is to put documentation\n> in the database itself, and to provide tools\n> to import/export SGML documents to PostgreSQL\n> and to crate a \"document server\" that\n> is accessable from the web.\n\nSounds like a neat project.\n\n - Tom\n",
"msg_date": "Sat, 19 Dec 1998 02:52:07 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Help with Documentation?"
},
{
"msg_contents": "> Easy - and the type coersion stuff helps alot!\n> Up to now (1.5 hours hacking) I have a NUMERIC type that\n> handles '+' completely, including overflow checks and\n> rounding.\n\nNeat. I was poking around waiting on a freeware extended-precision\nnumerical package, but hadn't found anything with a BSD-style license.\n\n> Another tricky part I expect when telling the parser that a\n> literal NUMERIC must not be enclosed into single quotes.\n\nThis is probably the worst part, since you would hate to take the hit\nrepresenting everything as extended precision even if the actual range\nis int4/float8. Perhaps we can read as those, but if we get a failure\nthen escalate to your extended precision type(s). The automatic type\nconversion stuff should convert later if necessary, so it might be\ntransparent and relatively fast.\n\n> The only ugly thing is, that I needed to put the precision\n> AND the scale together into atttypmod (I limited both to 99\n> for now and put them as prec<<8 | scale into). So pg_dump\n> will need attention later.\n\nAs Bruce points out, that was anticipated (but I agree it's ugly). Shift\nover 16 bits if you want...\n\nCongrats!\n\n - Tom\n",
"msg_date": "Sat, 19 Dec 1998 02:58:32 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "Is this running anywhere?\n\nClark\n",
"msg_date": "Sat, 19 Dec 1998 04:31:48 +0000",
"msg_from": "Clark Evans <[email protected]>",
"msg_from_op": false,
"msg_subject": "CVS Web?"
},
{
"msg_contents": "\nJust installed the cgi for it at http://www.postgresql.org/cgi/cvsweb ...\nI leave it to our esteemed webmaster to make it look good :)\n\nOn Sat, 19 Dec 1998, Clark Evans wrote:\n\n> Is this running anywhere?\n> \n> Clark\n> \n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Sat, 19 Dec 1998 01:29:53 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] CVS Web?"
},
{
"msg_contents": "On Sat, 19 Dec 1998, Thomas G. Lockhart wrote:\n\n> Date: Sat, 19 Dec 1998 02:58:32 +0000\n> From: \"Thomas G. Lockhart\" <[email protected]>\n> To: Jan Wieck <[email protected]>\n> Cc: [email protected], [email protected], [email protected]\n> Subject: Re: [HACKERS] Upgrades for 6.4.1\n> \n> > Easy - and the type coersion stuff helps alot!\n> > Up to now (1.5 hours hacking) I have a NUMERIC type that\n> > handles '+' completely, including overflow checks and\n> > rounding.\n> \n> Neat. I was poking around waiting on a freeware extended-precision\n> numerical package, but hadn't found anything with a BSD-style license.\n\nDid you try search SAL ( Sci. Application for Linux) ?\nI just did search for 'precision' and found abot 12 items. Don't know\nabout license bit some of them like\n<a href=\"http://www.hut.fi/~mtommila/apfloat/\">apfloat - A C++ High Performance \nArbitrary Precision Arithmetic Package</a> are freeware.\n>From web page:\nApfloat is a high performance arbitrary precision package. That means you can do calculations involving\nmillions of digits with it. It uses Number Theoretic Transforms . It's simple to use. It's fast. It's freeware.\n\n\n\n> \n> > Another tricky part I expect when telling the parser that a\n> > literal NUMERIC must not be enclosed into single quotes.\n> \n> This is probably the worst part, since you would hate to take the hit\n> representing everything as extended precision even if the actual range\n> is int4/float8. Perhaps we can read as those, but if we get a failure\n> then escalate to your extended precision type(s). The automatic type\n> conversion stuff should convert later if necessary, so it might be\n> transparent and relatively fast.\n> \n> > The only ugly thing is, that I needed to put the precision\n> > AND the scale together into atttypmod (I limited both to 99\n> > for now and put them as prec<<8 | scale into). So pg_dump\n> > will need attention later.\n> \n> As Bruce points out, that was anticipated (but I agree it's ugly). Shift\n> over 16 bits if you want...\n> \n> Congrats!\n> \n> - Tom\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": "Sat, 19 Dec 1998 13:10:11 +0300 (MSK)",
"msg_from": "Oleg Bartunov <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "\"Thomas G. Lockhart\" <[email protected]> writes:\n>> * SELECT DISTINCT i FROM dtest ORDER BY j generates strange output\n\n> In my simple test case, it orders by j, then only shows i. Is that\n> strange?\n\nThe thing that is \"strange\" is that you get nonunique values of i,\nwhich is definitely a bit unexpected for \"SELECT DISTINCT\":\n\nplay=> SELECT * FROM dtest;\ni| j\n-+--\n1|10\n1|11\n1|12\n2|20\n3|30\n(5 rows)\n\nplay=> SELECT DISTINCT i FROM dtest ORDER BY j;\ni\n-\n1\n1\n1\n2\n3\n(5 rows)\n\nThe reason that this is happening is that the \"distinct\" filter is\nactually being run on i,j not just i (see \"distinct + order by\" thread\nin the hackers archives around 8-Nov-98).\n\nI don't know whether the SQL standard defines how this combination of\nfeatures ought to work ... but our current behavior seems fairly\nsurprising...\n\n\n>> * Allow constraint NULL just as we honor NOT NULL\n\n> Fundamental yacc problem with this as I recall. Gives rise to\n> shift/reduce problems since it is ambiguous with other uses of \"NULL\" in\n> the same area.\n\nMore to the point, what possible use would a column constrained to NULL\nbe? Might as well just not have it in the table...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 19 Dec 1998 12:42:49 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1 "
},
{
"msg_contents": "Hope you don't mind my feedback on this one. If so, \nI'll stick to the general list.\n\n> play=> SELECT DISTINCT i FROM dtest ORDER BY j;\n\n> I don't know whether the SQL standard defines how this combination of\n> features ought to work ... but our current behavior seems fairly\n> surprising...\n\nIt suprised me. Perhaps the query should be re-written using \na from \"select clause\" (which is an instant view) on it's way \ndown the pipe, like:\n\nSELECT DISTINCT i FROM (SELECT i FROM dest ORDER BY j);\n\n\n> More to the point, what possible use would a column constrained to NULL\n> be? Might as well just not have it in the table...\n\nBackward compatibility with a client appliation (which has the \ncolumn hard coded in a select statement), only you want to make \nsure that it isn't used... I've actually done this in real life. \nThe client was deployed to thousands of users, and was being phased\nout, yet we had to support some reporting it did. \n*shudder* Don't remind me.\n",
"msg_date": "Sat, 19 Dec 1998 20:31:19 +0000",
"msg_from": "Clark Evans <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "> \"Thomas G. Lockhart\" <[email protected]> writes:\n> >> * SELECT DISTINCT i FROM dtest ORDER BY j generates strange output\n> \n> > In my simple test case, it orders by j, then only shows i. Is that\n> > strange?\n> \n> The thing that is \"strange\" is that you get nonunique values of i,\n> which is definitely a bit unexpected for \"SELECT DISTINCT\":\n> I don't know whether the SQL standard defines how this combination of\n> features ought to work ... but our current behavior seems fairly\n> surprising...\n\nRe-added to TODO list.\n\n> \n> \n> >> * Allow constraint NULL just as we honor NOT NULL\n> \n> > Fundamental yacc problem with this as I recall. Gives rise to\n> > shift/reduce problems since it is ambiguous with other uses of \"NULL\" in\n> > the same area.\n> \n> More to the point, what possible use would a column constrained to NULL\n> be? Might as well just not have it in the table...\n\nIt just says the column _may_ accept nulls. It is the default anyway.\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, 19 Dec 1998 18:37:59 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": "> > >> * SELECT DISTINCT i FROM dtest ORDER BY j generates strange output\n> > \n> > > In my simple test case, it orders by j, then only shows i. Is that\n> > > strange?\n> > \n> > The thing that is \"strange\" is that you get nonunique values of i,\n> > which is definitely a bit unexpected for \"SELECT DISTINCT\":\n> > I don't know whether the SQL standard defines how this combination of\n> > features ought to work ... but our current behavior seems fairly\n> > surprising...\n> \n> Re-added to TODO list.\n\nI don't know if I would necessarily call it \"strange\", more like\n\"understandable\" if you consider the fact that the ORDER BY is being\ndone first on the data set rather than the distinct.\n\nIt would seem to me that the distinct should apply first though.\nI would expect the ORDER BY clause to order whatever tuples are\nreturned by the SELECT, and that would imply doing DISTINCT first.\n\nJust my two cents...\ndarrenk\n",
"msg_date": "Sat, 19 Dec 1998 22:41:56 -0500",
"msg_from": "\"Stupor Genius\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "RE: [HACKERS] Upgrades for 6.4.1"
},
{
"msg_contents": ">From P121 \"A Guide to the SQL Standard, C.J. Date, 1997\",\n\nS is defined as:\n\n> SNO SNAME STATUS CITY\n> --- ----- ------ ------\n> 0S1 SMITH 000020 LONDON\n> 0S2 JONES 000010 PARIS\n> 0S3 BLAKE 000030 PARIS\n> 0S4 CLARK 000020 LONDON\n> 0S5 ADAMS 000030 ATHENS\n\nand SP is defined as:\n\n> SNO PNO QTY\n> --- --- ---\n> 0S1 0P1 300\n> 0S1 0P2 200\n> 0S1\n> 0S2\n> 0S2 etc...\n> 0S3\n> 0S4\n> 0S5\n \nThe book says:\n\n> Note that each order-item must identify a column \n> of T itself, not just a column of some table from \n> which T is derived. Thus, for example, the following\n> is ***ILLEGAL***:\n>\n> DELCARE Z CURSOR FOR\n> SELECT S.SNO\n> FROM S\n> ORDER BY CITY\n> -- *** ILLEGAL *** !!!\n>\n\nThe book is rather unclear as to what \"T\" means... although\nI assume that it means the relation T (SNO), and not the\ntable S (SNO,SNAME,STATUS,CITY). Thus, if you want CITY \nto go in the order by, then you must include it in the \nselect list. \n\nBut wait! Oracle allows the above query! From what I \nunderstand though, the database engine implicitly includes\nthe CITY in the internal processing, the information \nis merely discarded after the order by and not returned.\n\nThus, it is a very useful, pratical short hand for:\n\nSELECT SNO\n FROM\n ( SELECT S.SNO, S.CITY\n FROM S\n ORDER BY CITY\n )\n\nI suggest that it might be useful to consider it in this \nmanner...\n\n\nSo, assuming that this is a \"good\" interpretation of\nthe above item. Let's consider how a DISTINCT fits in.\nClearly, it has no use in the inner query, where the\nordering occurs.\n\nThus, the distinct occurs on the outside, like:\n\nSELECT DISTINCT i\n FROM\n ( SELECT i, j\n FROM dtest\n ORDER BY i\n )\n\nNice, hunh? Well not exactly. Consider this data:\n\ni j\n- -\nB 3\nA 1\nA 5\n\nThus, after the inner query, we have the following:\n\ni j\n- -\nA 1\nB 3\nA 5\n\n\nNow we take the the DISTINCT i right?\n\nWhat is the anwer? The book talks about this in\nthe very next paragraph:\n\n> Finally, if no ORDER BY clause is specified, the rows\n> of T will hve an implementation-dependent ordering. \n> Likewise, if an ORDER BY clause is specified but does not\n> define a \"total\" ordering, then the relative order of\n> rows within T that have the same value for the order-item(s)\n> will again be implemetnation-dependent. For example:\n>\n> DELCARE W CURSOR FOR\n> SELECT SP.SNO, SP.PNO, SP.QTY\n> FROM SP\n> ORDER BY SNO\n>\n> Here the relevant order of SP rows with the same PNO value\n> will be inplementation-dependent\n> \n\nSince a \"total\" ordering was not provided in this \ncase, and assuming that the \"shorthand\" above\nis allowed, the answer is implementation-dependent.\n\ni\n-\nA\nB\n\nor\n\ni\n-\nB\nA\n\n\nHope this helps. \n\n:) Clark\n",
"msg_date": "Sun, 20 Dec 1998 04:49:10 +0000",
"msg_from": "Clark Evans <[email protected]>",
"msg_from_op": false,
"msg_subject": "SELECT DISTINCT i FROM dtest ORDER BY j"
},
{
"msg_contents": "Clark Evans <[email protected]> writes:\n>> From P121 \"A Guide to the SQL Standard, C.J. Date, 1997\":\n>> Note that each order-item must identify a column \n>> of T itself, not just a column of some table from \n>> which T is derived. Thus, for example, the following\n>> is ***ILLEGAL***:\n>> \n>> DELCARE Z CURSOR FOR\n>> SELECT S.SNO\n>> FROM S\n>> ORDER BY CITY\n>> -- *** ILLEGAL *** !!!\n\nHow interesting. I believe that in fact Postgres used to refuse\nsuch queries, and that the ability to order by a field not present\nin the result is new in 6.4. Maybe now we are starting to find out\nwhy the SQL spec forbids it ;-)\n\n> But wait! Oracle allows the above query! From what I \n> understand though, the database engine implicitly includes\n> the CITY in the internal processing, the information \n> is merely discarded after the order by and not returned.\n\nRight, that's how Postgres does it too.\n\n\nMeanwhile darrenk wrote:\n> It would seem to me that the distinct should apply first though.\n> I would expect the ORDER BY clause to order whatever tuples are\n> returned by the SELECT, and that would imply doing DISTINCT first.\n\nThe trouble is that if you have several tuples with the same i and\ndifferent j, doing the DISTINCT first implies throwing away all but\none of those tuples. Which one do you keep? It matters because\nsome of those tuples might sort differently than others.\n\nAs far as I can see, this combination of features is not well-defined on\nits surface. You have to make some additional assumptions (about which\nof the possible j values is kept for sorting) in order to define a\nunique result.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 20 Dec 1998 12:53:30 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] SELECT DISTINCT i FROM dtest ORDER BY j "
},
{
"msg_contents": "> The trouble is that if you have several tuples with the same i and\n> different j, doing the DISTINCT first implies throwing away all but\n> one of those tuples. Which one do you keep? It matters because\n> some of those tuples might sort differently than others.\n> \n> As far as I can see, this combination of features is not well-defined on\n> its surface. You have to make some additional assumptions (about which\n> of the possible j values is kept for sorting) in order to define a\n> unique result.\n\nThat is an interesting argument. It suggests it should not be a TODO\nitem. I can see cases where you would want to do the DISTINCT before\nthe ORDER BY, so the current behavior seems correct.\n\nRemoving it from TODO list.\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, 20 Dec 1998 14:03:45 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] SELECT DISTINCT i FROM dtest ORDER BY j"
},
{
"msg_contents": ">\n> > The only ugly thing is, that I needed to put the precision\n> > AND the scale together into atttypmod (I limited both to 99\n> > for now and put them as prec<<8 | scale into). So pg_dump\n> > will need attention later.\n>\n> Atttypmod was expanded to 32 bits for exactly that use. I recommend\n> 16-bits for each option.\n\n I tried that first and it failed. Now that you said that,\n I've found in nodes/parsenodes.h that the typmod field of\n TypeName is still int16.\n\n I'll change that to int32 after having checked that it\n doesn't break anything.\n\n I also read the comment from Oleg about the other numeric\n package out. Haven't taken a look at it, but I think for\n Postgres I would have to build a complete new set of\n functions from scratch anyway. The reasons for that are:\n\n 1. I want to store the digits of the NUMERIC type as nibbles\n in the tuple to save space (so it will become a real\n packed format). My first hack used one byte per digit and\n I think it's wasted space.\n\n 2. I want to use the VARSIZE, one combined int16 holding the\n sign and the number of digits before the decimal point\n and one int16 holding the digits to DISPLAY after the\n decimal point as metadata describing the precision and\n calc-/display scale.\n\n None of the packages I've seen up to now has a separate\n DISPLAY scale in the number itself. They all use some\n global scale variable.\n\n 3. The packed internal format will need to be\n unpacked/packed for speed on complex operations (divide,\n multiply, log etc.). While add and subtract should be\n possible directly on the packed format.\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, 21 Dec 1998 14:00:34 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Upgrades for 6.4.1"
}
] |
[
{
"msg_contents": "Hello!\n\nOn Thu, 17 Dec 1998, Tim Williams wrote:\n> When I run the following set of statements in postgres v6.4 (built on\n> Debian Linux):\n> \n> create table tmp (v1 date, v2 datetime);\n> insert into tmp values ('06-01-1999', '06-01-1999');\n> select date_part('month', v1) as m1, date_part('month', v2) as m2 from\n> tmp;\n> \n> I see the following results:\n> \n> m1|m2\n> --+--\n> 5| 6\n> (1 row)\n> \n> Why is date_part giving different results for date and date_time? Why\n> is it wrong for 'date'? Anybody seen this before?\n\n I tried this and it works pretty good for me on Solaris.\n It looks like an error in postgres on glibc2-based linux platform. It is\ndiscussing now on pgsql-hackers list. The solution is not ready yet. I am\nworking on this with a good deal of help from other people...\n\nOleg.\n---- \n Oleg Broytmann http://members.tripod.com/~phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n",
"msg_date": "Fri, 18 Dec 1998 11:41:34 +0300 (MSK)",
"msg_from": "Oleg Broytmann <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [GENERAL] date_part bug?"
},
{
"msg_contents": "Oleg Broytmann wrote:\n >> create table tmp (v1 date, v2 datetime);\n >> insert into tmp values ('06-01-1999', '06-01-1999');\n >> select date_part('month', v1) as m1, date_part('month', v2) as m2 from\n >> tmp;\n >> \n >> I see the following results:\n >> \n >> m1|m2\n >> --+--\n >> 5| 6\n >> (1 row)\n...\n > I tried this and it works pretty good for me on Solaris.\n > It looks like an error in postgres on glibc2-based linux platform. It is\n >discussing now on pgsql-hackers list. The solution is not ready yet. I am\n >working on this with a good deal of help from other people...\n\nNot all glibc2: I run glibc2 Debian Linux and do not see this problem.\nI was wondering if it was libc5 that was giving trouble...\n\n-- \nOliver Elphick [email protected]\nIsle of Wight http://www.lfix.co.uk/oliver\n PGP key from public servers; key ID 32B8FAA1\n ========================================\n \"In the beginning was the Word, and the Word was with \n God, and the Word was God. The same was in the \n beginning with God. All things were made by him; and \n without him was not any thing made that was made.\" \n John 1:1-3 \n\n\n",
"msg_date": "Fri, 18 Dec 1998 09:48:07 +0000",
"msg_from": "\"Oliver Elphick\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [GENERAL] date_part bug? "
},
{
"msg_contents": "> Not all glibc2: I run glibc2 Debian Linux and do not see this problem.\n> I was wondering if it was libc5 that was giving trouble...\n\nNot on my libc5-only machine (where I developed the code). We'll need to\nget a reproducible case to be able to track it down. I'm guessing that\nwe are seeing float->int rounding problems, though I don't know why this\ntest query should show different results for the two columns.\n\nWhat glibc2, compiler, and optimization level are you using? If you are\nusing anything above -O2 try backing down to that; if you are already\nthere then try -O0 and tell us what changes.\n\n - Tom\n\npostgres=> create table tmp (v1 date, v2 datetime);\nCREATE\npostgres=> insert into tmp values ('06-01-1999', '06-01-1999');\nINSERT 901482 1\npostgres=> select date_part('month', v1) as m1, date_part('month', v2)\nas m2 from tmp;\nm1|m2\n--+--\n 6| 6\n(1 row)\n",
"msg_date": "Fri, 18 Dec 1998 15:18:32 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: [GENERAL] date_part bug?"
}
] |
[
{
"msg_contents": "Hi,\n\n just committed some fixes to get back the old \"<>\" for NULL\n strings on nodeToString().\n\n Marc's recent changes to appendStringInfo() handling broke\n it. The old version converted a NULL string into a \"<>\".\n Handing a NULL for %s to vsnprintf() is implementation\n dependent (some put nothing, some put (null) and some might\n simply crash - but none puts \"<>\").\n\n There is a new macro stringStringInfo(s) in include/lib.h\n which checks for NULL and substitutes by \"<>\". All string\n arguments to appendStringInfo() are now embedded into that\n (also in commands/explain.c).\n\n Regression tests for select_views, rules and plpgsql are O.K.\n 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\n",
"msg_date": "Fri, 18 Dec 1998 16:07:43 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": true,
"msg_subject": "Fixed outfuncs"
}
] |
[
{
"msg_contents": "I get this error:\n\ngcc2 -I../../../include -I../../../backend -I/usr/local/include/readline -O2 -\nm486 -pipe -g -Wall -I../.. -c nabstime.c -o nabstime.o\nnabstime.c: In function `GetCurrentAbsoluteTime':\nnabstime.c:78: `tzn' undeclared (first use this function)\nnabstime.c:78: (Each undeclared identifier is reported only once\nnabstime.c:78: for each function it appears in.)\ngmake[3]: *** [nabstime.o] Error 1\n\nMarc plans to package up the release on Monday night.\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, 18 Dec 1998 14:57:02 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "6.4.1 fails to compile"
},
{
"msg_contents": "> nabstime.c: In function `GetCurrentAbsoluteTime':\n> nabstime.c:78: `tzn' undeclared (first use this function)\n> Marc plans to package up the release on Monday night.\n\nI thought he or I fixed it. Will check...\n\n - Tom\n",
"msg_date": "Sat, 19 Dec 1998 03:04:37 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.1 fails to compile"
},
{
"msg_contents": "> > nabstime.c: In function `GetCurrentAbsoluteTime':\n> > nabstime.c:78: `tzn' undeclared (first use this function)\n> > Marc plans to package up the release on Monday night.\n\nFixed. The fix had been only applied to the main tree. Sorry.\n\n - Tom\n",
"msg_date": "Sat, 19 Dec 1998 03:23:58 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.1 fails to compile"
},
{
"msg_contents": "On Sat, 19 Dec 1998, Thomas G. Lockhart wrote:\n\n> > > nabstime.c: In function `GetCurrentAbsoluteTime':\n> > > nabstime.c:78: `tzn' undeclared (first use this function)\n> > > Marc plans to package up the release on Monday night.\n> \n> Fixed. The fix had been only applied to the main tree. Sorry.\n\n\tMy fauilt too...I thought the bug was only in the main\ntree...wonder how that bug sneaked into v6.4's tree...eh Tom? :)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Sat, 19 Dec 1998 00:13:03 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.1 fails to compile"
},
{
"msg_contents": "> > > > nabstime.c: In function `GetCurrentAbsoluteTime':\n> > > > nabstime.c:78: `tzn' undeclared (first use this function)\n> > > > Marc plans to package up the release on Monday night.\n> > Fixed. The fix had been only applied to the main tree. Sorry.\n> My fault too...I thought the bug was only in the main\n> tree...wonder how that bug sneaked into v6.4's tree...eh Tom? :)\n\nBy popular acclaim :)\n\nIt was discussed on the hacker's list before being applied. Caught ya'\nnapping, eh? Maybe we don't have much adult supervision after all? :)\n\n - Tom\n",
"msg_date": "Sat, 19 Dec 1998 04:29:51 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.1 fails to compile"
}
] |
[
{
"msg_contents": "Jan,\n\nThanks for this, I'm happy now :-)\n\nRegression tests all pass except for a couple of\n\"non functional update\" messages in misc.\n\nGreat job,\nKeith.\n\n\[email protected] (Jan Wieck)\n> \n> Hi,\n> \n> just committed some fixes to get back the old \"<>\" for NULL\n> strings on nodeToString().\n> \n\n\n",
"msg_date": "Sat, 19 Dec 1998 00:32:49 +0000 (GMT)",
"msg_from": "Keith Parks <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Fixed outfuncs"
}
] |
[
{
"msg_contents": "I haven't tried elsewhere first because this really isn't a question\nper-se. It is a request. All software packages have bugs. Known bugs\nshould be made public!\n\nIs it possible to make a BUGS-list (not a mailing list... just a text\nfile or html page retrievable from web space) and stick in a very\nobvious/prominent place on www.postgresql.org?\n\nI know many people who would like to dabble, and might help out with\nbugs. I also know people (namely myself) who are stuck using 6.3.x\nbecause they don't know whether bugs (hearing there are many) in 6.4\nwill affect them. Obviously this list would help programmers submit\npatches. It would also help implementers know whether a version is safe\nfor testing and then deployment.\n\nThanks for your consideration. Again, just a suggestion, I know this is\n\"Open Source\" software and therefore no one has to do squat.\n\nTrever Adams\n",
"msg_date": "Sat, 19 Dec 1998 05:51:12 -0600",
"msg_from": "Trever Adams <[email protected]>",
"msg_from_op": true,
"msg_subject": "BUGS list"
}
] |
[
{
"msg_contents": "Hi all\n\nshell_connection=> create view orders_v as\nshell_connection-> select\nshell_connection-> o.qty as oqty,\nshell_connection-> odate,\nshell_connection-> sdate,\n\n(odate and sdate are in orders only)\n\nshell_connection-> o.price as oprice,\nshell_connection-> c.*,\nshell_connection-> s.*\nshell_connection-> from\nshell_connection-> orders o,\nshell_connection-> cust c,\nshell_connection-> shells s\nshell_connection-> where\nshell_connection-> o.accno = c.accno and\nshell_connection-> o.stockno = s.stockno;\nERROR: DefineQueryRewrite: rule plan string too big.\nshell_connection=>\n\nIs this a bug? or am I doing some thing wrong?\nAnd if it is a bug, is it fixed in 6.4.1?\n\nWhat's a 'rule plan string'? the part I typed? if so, how long can it be?\nI also tried typing it all as one (wraped) line, but same thing.\n\nThanks and have a great day\nTerry Mackintosh <[email protected]> http://www.terrym.com\nsysadmin/owner Please! No MIME encoded or HTML mail, unless needed.\n\nProudly powered by R H Linux 4.2, Apache 1.3, PHP 3, PostgreSQL 6.4\n-------------------------------------------------------------------\nSuccess Is A Choice ... book by Rick Patino, get it, read it!\n\n",
"msg_date": "Sat, 19 Dec 1998 13:30:10 -0500 (EST)",
"msg_from": "Terry Mackintosh <[email protected]>",
"msg_from_op": true,
"msg_subject": "Is this a bug? or am I doing some thing wrong?"
},
{
"msg_contents": "Terry Mackintosh wrote:\n\n> [...]\n> shell_connection-> o.accno = c.accno and\n> shell_connection-> o.stockno = s.stockno;\n> ERROR: DefineQueryRewrite: rule plan string too big.\n> shell_connection=>\n>\n> Is this a bug? or am I doing some thing wrong?\n> And if it is a bug, is it fixed in 6.4.1?\n\n It's not a real bug, it's a side effect from the fact that a\n tuple cannot span multiple disk blocks.\n\n>\n> What's a 'rule plan string'? the part I typed? if so, how long can it be?\n> I also tried typing it all as one (wraped) line, but same thing.\n\n First it's wrong. The system stores a string representation\n of the rules querytree in pg_rewrite (not the rules plan).\n Anything the rule system deals with are querytrees. Those\n ones coming from the parser for the actual query and those\n ones generated for the rule actions and qualifications. A\n querytree consists of cascaded node structures in memory. The\n function nodeToString() builds a string which is stored in\n pg_rewrite and stringToNode() is used to rebuild the in\n memory structures from that.\n\n These strings are very explanative and thus very long\n compared against the query they represent. They must fit into\n one text type attribute and from the limitation above, the\n resulting pg_rewrite tuple must fit into one (8k by default)\n disk block.\n\n While fixing the rule system for v6.4 I thought about\n implementing continuation rows for rewrite rules. But there\n are at least ideas out how tuples could eventually span disk\n blocks in the future, so I left that problem untouched\n waiting for this general solution.\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, 21 Dec 1998 11:42:38 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Is this a bug? or am I doing some thing wrong?"
}
] |
[
{
"msg_contents": "Hello!\n\n Continuing getting bug reports...\n This guy see the same error on FreeBSD 2.2.5, pgsql 6.4. He reports\n6.3.2 had no the problem...\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---------- Forwarded message ----------\nDate: Sat, 19 Dec 1998 18:39:49 +1200\nFrom: Vladimir Dobrokhotov <[email protected]>\nTo: [email protected]\nSubject: PostgreSQL 6.4 like bug(?)\n\nDate: Mon, 14 Dec 1998 16:37:29 +0300 (MSK)\nFrom: Oleg Broytmann <[email protected]>\nSubject: Re: [GENERAL] Re: locale in postgresql\n> test=> select * from test;\n> name\n> ----\n> abc\n> bca\n> cab\n> (3 rows)\n>\n> test=> select * from test where name like 'a%';\n> name\n> ----\n> (0 rows)]\n>\n\nО©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫, О©╫ О©╫О©╫О©╫О©╫ О©╫О©╫ О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫\nО©╫ О©╫ О©╫О©╫ О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫ О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫.\nО©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ FreeBSD 225 О©╫ postgrteSQL 6.4.\n\nО©╫ О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫ О©╫ О©╫ О©╫О©╫О©╫О©╫О©╫\nО©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫\nО©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫ 6.4 - О©╫ 6.3.2 О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫ О©╫О©╫О©╫О©╫.\n\nО©╫О©╫О©╫ bug О©╫О©╫О©╫ О©╫ О©╫О©╫О©╫-О©╫О©╫ О©╫О©╫ О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫?\n(О©╫ О©╫О©╫О©╫ О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫ locale?).\n\nО©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫,\nО©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫,\nО©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫.\[email protected],\[email protected]\n\n\n",
"msg_date": "Mon, 21 Dec 1998 13:24:42 +0300 (MSK)",
"msg_from": "Oleg Broytmann <[email protected]>",
"msg_from_op": true,
"msg_subject": "PostgreSQL 6.4 like bug(?)"
}
] |
[
{
"msg_contents": "I made an error in the script that builds the tar file for PyGreSQL.\nCan someone with access please move the following files from the\n.../src/interfaces/python directory down one level to\n.../src/interfaces/python/tutorial please?\n\nadvanced.py\nbasics.py\nfunc.py\npgtools.py\nsyscat.py\n\nThanks.\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, 21 Dec 1998 11:11:25 -0500 (EST)",
"msg_from": "\"D'Arcy\" \"J.M.\" Cain <[email protected]>",
"msg_from_op": true,
"msg_subject": "Python interface"
}
] |
[
{
"msg_contents": "Hello!\n\n Continuing getting bug reports...\n This guy see the same error on FreeBSD 2.2.5, pgsql 6.4. He reports\n6.3.2 didn't have the problem...\n\nOleg.\n---- \n Oleg Broytmann http://members.tripod.com/~phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n---------- Forwarded message ----------\nDate: Sat, 19 Dec 1998 18:39:49 +1200\nFrom: Vladimir Dobrokhotov <[email protected]>\nTo: [email protected]\nSubject: PostgreSQL 6.4 like bug(?)\n\nDate: Mon, 14 Dec 1998 16:37:29 +0300 (MSK)\nFrom: Oleg Broytmann <[email protected]>\nSubject: Re: [GENERAL] Re: locale in postgresql\n> test=> select * from test;\n> name\n> ----\n> abc\n> bca\n> cab\n> (3 rows)\n>\n> test=> select * from test where name like 'a%';\n> name\n> ----\n> (0 rows)]\n>\n\nО©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫, О©╫ О©╫О©╫О©╫О©╫ О©╫О©╫ О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫\nО©╫ О©╫ О©╫О©╫ О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫ О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫.\nО©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ FreeBSD 225 О©╫ postgrteSQL 6.4.\n\nО©╫ О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫ О©╫ О©╫ О©╫О©╫О©╫О©╫О©╫\nО©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫\nО©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫ 6.4 - О©╫ 6.3.2 О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫ О©╫О©╫О©╫О©╫.\n\nО©╫О©╫О©╫ bug О©╫О©╫О©╫ О©╫ О©╫О©╫О©╫-О©╫О©╫ О©╫О©╫ О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫?\n(О©╫ О©╫О©╫О©╫ О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫ locale?).\n\nО©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫,\nО©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫ О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫,\nО©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫О©╫.\[email protected],\[email protected]\n\n\n\n",
"msg_date": "Mon, 21 Dec 1998 19:39:43 +0300 (MSK)",
"msg_from": "Oleg Broytmann <[email protected]>",
"msg_from_op": true,
"msg_subject": "PostgreSQL 6.4 like bug(?)"
},
{
"msg_contents": "> Continuing getting bug reports...\n> This guy see the same error on FreeBSD 2.2.5, pgsql 6.4. He reports\n> 6.3.2 didn't have the problem...\n\nI cannot reproduce this. Is the database compiled with MultiByte\nsupport? Mine is not...\n\n - Tom\n",
"msg_date": "Tue, 22 Dec 1998 06:33:13 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] PostgreSQL 6.4 like bug(?)"
},
{
"msg_contents": "Hi!\n\nOn Tue, 22 Dec 1998, Thomas G. Lockhart wrote:\n> > This guy see the same error on FreeBSD 2.2.5, pgsql 6.4. He reports\n> > 6.3.2 didn't have the problem...\n> \n> I cannot reproduce this. Is the database compiled with MultiByte\n> support? Mine is not...\n\n I don't think so. Cyrillic locale is not multibyte.\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": "Tue, 22 Dec 1998 11:20:21 +0300 (MSK)",
"msg_from": "Oleg Broytmann <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] PostgreSQL 6.4 like bug(?)"
}
] |
[
{
"msg_contents": "\nThings have quieted down considerably this past weekend, so is it safe to\nassume that ppl are happy where the v6.4.1 tree is sitting right now?\n\nPending any outcry against, I will create the tar-ball of it tonight after\nI get home...~5hrs from now or so...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Mon, 21 Dec 1998 13:44:59 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": true,
"msg_subject": "v6.4.1 - to be released..."
}
] |
[
{
"msg_contents": "Hi,\n\n sometimes it's good not to spend too much efford implementing\n the final solution first. So for the NUMERIC.\n\n First I wonder why the can_coerce... stuff is #if'd out of\n parse_relation.c? For the NUMERIC type the\n numeric(num,typmod) must be called if someone does an\n\n INSERT INTO ... SELECT * FROM ...\n\n But it isn't. It is only called when there are calculations\n done on the columns. I also checked that for BPCHAR type and\n it simply throws an ERROR if the target's length doesn't\n match.\n\n This might be easy to fix, but the other problem I have is a\n bit more difficult.\n\n When binary operators (add, sub, mul, div) are called, the\n required precision of the result isn't known. And the coerce\n function numeric(num,typmod) will only be called for the\n final result. Now have the following situation:\n\n CREATE TABLE t1 (id int4, annual_val numeric(20,4));\n CREATE TABLE t2 (id int4, monthly_val numeric(24,8));\n\n INSERT INTO t2 SELECT id, annual_val / '12' FROM t1;\n\n A multiplication would have a maximum number of digits that\n can appear after the decimal point. It is the sum of number\n of digits present in the two operators. But not so for a\n division.\n\n If we want to implement NUMERIC with a real high precision\n (maybe 4000 or more digits), there would currently be no\n other chance than to do the division with the full ever\n possible precision and then throw away most of the digits\n when the result is assigned to the target column. Wasted\n efford and more important MUCH WASTED CPU.\n\n I can think of something like this:\n\n On add/subtract the results precision after the decimal point\n is the higher of the two operands.\n\n On multiply the results precision after the decimal point is\n the sum of the precisions of the two operands.\n\n On divide the results precision after the decimal point is\n like for mult or the double of the higher precision of the\n two operands. Any other suggestions?\n\n On the other hand it is possible to do it as\n\n INSERT INTO t2 SELECT id, ROUND(annual_val,8) / '12' FROM t1;\n\n How do other databases handle this problem. How is the\n precision of a numeric result defined?\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, 22 Dec 1998 09:58:35 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": true,
"msg_subject": "Problems on NUMERIC"
},
{
"msg_contents": "> First I wonder why the can_coerce... stuff is #if'd out of\n> parse_relation.c?\n\nOh! That looks like my style of #if FALSE, but I can't recall why it is\nthat way. Will look at it. Does it work to just substitute an #if TRUE?\nPerhaps I had it disabled during debugging, but...\n\n> How do other databases handle this problem. How is the\n> precision of a numeric result defined?\n\nI've enclosed some snippets from my SQL92 2nd Draft Standard doc. It\ngives you a lot of latitude :)\n\n - Tom\n\n Syntax Rules\n\n 1) If the data type of both operands of a dyadic arithmetic opera-\n tor is exact numeric, then the data type of the result is exact\n numeric, with precision and scale determined as follows:\n\n a) Let S1 and S2 be the scale of the first and second operands\n respectively.\n\n b) The precision of the result of addition and subtraction is\n implementation-defined, and the scale is the maximum of S1\n and S2.\n\n c) The precision of the result of multiplication is implementation-\n defined, and the scale is S1 + S2.\n\n d) The precision and scale of the result of division is\n implementation-defined.\n\n<snip large amounts>\n\n Whenever an exact or approximate numeric value is assigned to a\n data item or parameter representing an exact numeric value, an\n approximation of its value that preserves leading significant dig-\n its after rounding or truncating is represented in the data type\n of the target. The value is converted to have the precision and\n scale of the target. The choice of whether to truncate or round is\n implementation-defined.\n\n An approximation obtained by truncation of a numerical value N\n for an <exact numeric type> T is a value V representable in T such\n that N is not closer to zero than the numerical value of V and such\n that the absolute value of the difference between N and the numer-\n ical value of V is less than the absolute value of the difference\n between two successive numerical values representable in T.\n\n An approximation obtained by rounding of a numerical value N for\n an <exact numeric type> T is a value V representable in T such\n that the absolute value of the difference between N and the nu-\n merical value of V is not greater than half the absolute value\n of the difference between two successive numerical values repre-\n sentable in T. If there are more than one such values V, then it is\n implementation-defined which one is taken.\n\n All numerical values between the smallest and the largest value,\n inclusive, representable in a given exact numeric type have an\n approximation obtained by rounding or truncation for that type; it\n is implementation-defined which other numerical values have such\n approximations.\n",
"msg_date": "Tue, 22 Dec 1998 14:59:18 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Problems on NUMERIC"
},
{
"msg_contents": "> > First I wonder why the can_coerce... stuff is #if'd out of\n> > parse_relation.c?\n\nIt looks like the routine where the code appears cannot return a\nmodified expression, so I just placed this code in there as a marker.\nShould be possible to add some code to get this feature working.\n\n - Tom\n",
"msg_date": "Tue, 22 Dec 1998 15:15:57 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Problems on NUMERIC"
},
{
"msg_contents": ">\n> > First I wonder why the can_coerce... stuff is #if'd out of\n> > parse_relation.c?\n>\n> Oh! That looks like my style of #if FALSE, but I can't recall why it is\n> that way. Will look at it. Does it work to just substitute an #if TRUE?\n> Perhaps I had it disabled during debugging, but...\n\n Hmmm - elog(ERROR, \"Type %s(%d) can be coerced to... looks\n like debugging code for me. Maybe you wanted to elog(DEBUG...\n ?\n\n>\n> > How do other databases handle this problem. How is the\n> > precision of a numeric result defined?\n>\n> I've enclosed some snippets from my SQL92 2nd Draft Standard doc. It\n> gives you a lot of latitude :)\n>\n\n Thanks! That helps alot!\n\n>\n> d) The precision and scale of the result of division is\n> implementation-defined.\n\n I love those definitions :-)\n\n So I'll make the display scale of a division\n\n min( max(S1, S2), SLIMIT)\n\n and the internal result scale\n\n min( RMINIMUM, max(R1, R2) + 2, RLIMIT)\n\n where S1 and S2 are the display scales of the two operands,\n R1 and R2 are the internal present scales and SLIMIT, RLIMIT\n is the implementation-defined maximum allowed scale at all\n (what about 4000 for SLIMIT ?). The RMINIMUM is 8 to have\n anything at least computed internal with 8 digits after the\n decimal point (because the defaults for NUMERIC are precision\n 30 scale 6).\n\n If then the result is assigned to another tuples attribute,\n numeric(num,typmod) will be called and do the rounding with\n the scale defined in typmod. If numeric_out(num) is called\n for it, it will be output rounded to the above display scale.\n\n With 'round(att1, 500) / att2' someone can then get the\n result with 500 digits scale.\n\n This way it is flexible enough but not to much wasted\n computing is done.\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, 22 Dec 1998 16:53:46 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Problems on NUMERIC"
},
{
"msg_contents": "> First I wonder why the can_coerce... stuff is #if'd out of\n> parse_relation.c? For the NUMERIC type the\n> numeric(num,typmod) must be called if someone does an\n> \n> INSERT INTO ... SELECT * FROM ...\n> \n> But it isn't. It is only called when there are calculations\n> done on the columns. I also checked that for BPCHAR type and\n> it simply throws an ERROR if the target's length doesn't\n> match.\n\nSorry, I'm having trouble thinking of a case which does not behave\nproperly with the existing types. I've tried inserting varchar(10)\ncolumns into a varchar(1) column, I've tried inserting int columns into\nfloat columns, etc etc. How are you getting handleTargetColname() /\ncheckTargetTypes() called where it is rejecting things?\n\nIt may be that splitting that attribute field into two pieces for\nNUMERIC is opening a can of worms, since there are specific assumptions\nabout what that field means throughout the code :(\n\nMaybe we should think about how to isolate the type-specific\ninterpretation of that attribute field into a type-specific handler\nroutine? Ooh, that sounds like a pain...\n\n - Tom\n",
"msg_date": "Tue, 22 Dec 1998 16:16:23 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Problems on NUMERIC"
},
{
"msg_contents": ">\n> > First I wonder why the can_coerce... stuff is #if'd out of\n> > parse_relation.c? For the NUMERIC type the\n> > numeric(num,typmod) must be called if someone does an\n> >\n> > INSERT INTO ... SELECT * FROM ...\n> >\n> > But it isn't. It is only called when there are calculations\n> > done on the columns. I also checked that for BPCHAR type and\n> > it simply throws an ERROR if the target's length doesn't\n> > match.\n>\n> Sorry, I'm having trouble thinking of a case which does not behave\n> properly with the existing types. I've tried inserting varchar(10)\n> columns into a varchar(1) column, I've tried inserting int columns into\n> float columns, etc etc. How are you getting handleTargetColname() /\n> checkTargetTypes() called where it is rejecting things?\n\n pgsql=> create table t1 (a char(10));\n CREATE\n pgsql=> create table t2 (a char(4));\n CREATE\n pgsql=> insert into t2 select * from t1;\n ERROR: Length of a is not equal to the length of target column a\n pgsql=>\n\n>\n> It may be that splitting that attribute field into two pieces for\n> NUMERIC is opening a can of worms, since there are specific assumptions\n> about what that field means throughout the code :(\n\n It doesn't produce any problems so far, only that the\n function numeric(num,typmod) isn't called when doing a plain\n INSERT ... SELECT. It is only called when comparisions where\n performed in the SELECT clause of the INSERT on the numeric\n attributes. But I need that call to force the rounding and\n range check at INSERT time. Otherwise, the values in the\n target table will output later with the scale of their\n original source table, and that's wrong. Also it would be\n possible to insert 1000.0 into a numeric(5,2) attribute, and\n that shouldn't be.\n\n Maybe I have to hook for NUMERIC there in parse_relation too.\n Up to now I'm compiling the whole thing as loadable module.\n I'll check it that's possible when moving it to the builtins.\n\n But in general I think if there is a function with the same\n name as a type, that take this type plus another int4\n argument, this must be a range checker/padder/truncator or\n the like and it should be called before values are assigned\n to attributes.\n\n>\n> Maybe we should think about how to isolate the type-specific\n> interpretation of that attribute field into a type-specific handler\n> routine? Ooh, that sounds like a pain...\n\n Noooooooooo\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, 22 Dec 1998 17:47:19 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Problems on NUMERIC"
},
{
"msg_contents": "> > How are you getting handleTargetColname() /\n> > checkTargetTypes() called where it is rejecting things?\n\nOK, I'm not sure why the behavior is different if I explicitly specify\nthe columns (which of course I had for testing):\n\npostgres=> insert into t2 select a from t1;\nINSERT 0 0\npostgres=> insert into t2 select * from t1;\nERROR: Length of 'a' is not equal to the length of target column 'a'\n\n> It doesn't produce any problems so far, only that the\n> function numeric(num,typmod) isn't called when doing a plain\n> INSERT ... SELECT.\n\nHmm. Even when you explicitly specify the columns as I did in my example\nabove? I should be able to get the wildcard example to work sometime\nbefore v6.5, and I *think* that the explicit cases should do what you\nwant. As a loadable module, your data type will only match itself for\ntype coersion, but that's what you want for now. When it is built in,\nthen you will be able to specify that it is higher or lower in a\nheirarchy with, for example, int4 and float8.\n\n> Maybe I have to hook for NUMERIC there in parse_relation too.\n> Up to now I'm compiling the whole thing as loadable module.\n> I'll check it that's possible when moving it to the builtins.\n\nUnless you can't find a test case which does work for you, don't bother\nlooking at it; I'll pick it up some time soon.\n\n> But in general I think if there is a function with the same\n> name as a type, that take this type plus another int4\n> argument, this must be a range checker/padder/truncator or\n> the like and it should be called before values are assigned\n> to attributes.\n\nThat's how it should work afaik, at least for variable-length types. Not\nall types are checked for this conversion function...\n\n - Tom\n",
"msg_date": "Wed, 23 Dec 1998 01:54:24 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Problems on NUMERIC"
}
] |
[
{
"msg_contents": "I have no idea whether this will go through to patches as I've changed my\nemail address. Is patches run as a private mailing list too? \n\nAnyway, here's the latest ecpg update. It contains all changes newer than my\nlatest CVSUP (this morning).\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!",
"msg_date": "Tue, 22 Dec 1998 14:20:10 +0100",
"msg_from": "Michael Meskes <[email protected]>",
"msg_from_op": true,
"msg_subject": "ecpg patches"
},
{
"msg_contents": "\napplied, and yes...\n\nOn Tue, 22 Dec 1998, Michael Meskes wrote:\n\n> I have no idea whether this will go through to patches as I've changed my\n> email address. Is patches run as a private mailing list too? \n> \n> Anyway, here's the latest ecpg update. It contains all changes newer than my\n> latest CVSUP (this morning).\n> \n> Michael\n> -- \n> Michael Meskes | Go SF 49ers!\n> Th.-Heuss-Str. 61, D-41812 Erkelenz | Go Rhein Fire!\n> Tel.: (+49) 2431/72651 | Use Debian GNU/Linux!\n> Email: [email protected] | Use PostgreSQL!\n> \n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Tue, 22 Dec 1998 14:50:58 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] ecpg patches"
}
] |
[
{
"msg_contents": "Hallo,\n\n I'm trying postgres 6.4.1 and I have problem with select command:\n\n> create table t( text_column text );\n> select sum( text_column ) from t;\nThis command makes no sense, but causes crash in backend:\n\npqReadData() -- backend closed the channel unexpectedly.\n This 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\nSee backtrace:\n\n(gdb) bt\n#0 agg_select_candidate (typeid=25, candidates=0x81e7838) at\nparse_func.c:230\n#1 0x80bd5be in ParseFuncOrColumn (pstate=0x81ded30,\n funcname=0x81deb40 \"sum\", fargs=0x81deba0, curr_resno=0x81ded30,\n precedence=1) at parse_func.c:391\n#2 0x80bc7b9 in transformExpr (pstate=0x81ded30, expr=0x81debc0,\nprecedence=1)\n at parse_expr.c:267\n#3 0x80c12fa in MakeTargetEntryComplex (pstate=0x81ded30, res=0x81debe0)\n at parse_target.c:363\n#4 0x80c1754 in transformTargetList (pstate=0x81ded30,\ntargetlist=0x81dec08)\n at parse_target.c:569\n#5 0x80b45f1 in transformSelectStmt (pstate=0x81ded30, stmt=0x81deca0)\n at analyze.c:912\n#6 0x80b382d in transformStmt (pstate=0x81ded30, parseTree=0x81deca0)\n at analyze.c:200\n... more ....\n\nProblem is in function 'agg_select_candidate' itself, simple debugging\nsession follows: (comment starts with '!')\n\nReading symbols from /lib/libnsl.so.1...done.\nReading symbols from /lib/libdl.so.2...done.\nReading symbols from /lib/libm.so.6...done.\nReading symbols from /usr/lib/libreadline.so.3...done.\nReading symbols from /usr/lib/libhistory.so.3...done.\nReading symbols from /lib/libtermcap.so.2...done.\nReading symbols from /usr/lib/libncurses.so.4...done.\nReading symbols from /lib/libc.so.6...done.\nReading symbols from /lib/ld-linux.so.2...done.\n0x401356f4 in __read ()\n(gdb) break agg_select_candidate Quit\n(gdb) break agg_select_candidate\nBreakpoint 1 at 0x80bd2f9: file parse_func.c, line 198.\n(gdb) cont\nContinuing.\n\nBreakpoint 1, agg_select_candidate (typeid=25, candidates=0x81e7838)\n at parse_func.c:198\n198 category = TypeCategory(typeid);\n(gdb) next\n199 ncandidates = 0;\n(gdb)\n200 last_candidate = NULL;\n\n!!!!! setting to NULL (0)\n\n(gdb)\n201 for (current_candidate = candidates;\n(gdb) print last_candidate\n$1 = (struct _CandidateList *) 0x0\n(gdb) next\n202 current_candidate != NULL;\n(gdb)\n205 current_typeid = current_candidate->args[0];\n(gdb)\n206 current_category = TypeCategory(current_typeid);\n(gdb)\n208 if ((current_category == category)\n\n!!!!! condidion isn't true\n\n(gdb) print current_category\n$2 = TIMESPAN_TYPE\n(gdb) print category\n$3 = STRING_TYPE\n(gdb) next\n230 last_candidate->next = NULL;\n\n!!!! NULL is dereferenced ... this will crash backend\n\n(gdb) next\n\nProgram received signal SIGSEGV, Segmentation fault.\nagg_select_candidate (typeid=25, candidates=0x81e7838) at parse_func.c:230\n Machine is x86, compiled with egcs-1.1.1.\n\n\t\t\tthanks for any fix and happy Xmas,\n\t\t\t\t\t David\n\n-- \n* David Sauer, student of Czech Technical University\n* electronic mail: [email protected] (mime compatible)\n\n\n",
"msg_date": "23 Dec 1998 01:05:23 +0100",
"msg_from": "David Sauer <[email protected]>",
"msg_from_op": true,
"msg_subject": "(stupid) bug in agg_select_candidate"
},
{
"msg_contents": "> I'm trying postgres 6.4.1 and I have problem with select command:\n> > create table t( text_column text );\n> > select sum( text_column ) from t;\n> This command makes no sense, but causes crash in backend:\n\nThanks for finding this. It was in new code for automatic type matching\nand coersion on aggregate functions to help implement aggregates for the\nstring types.\n\nI've enclosed a simple patch which fixes the problem, and which at the\nsame time makes a nicer warning message as a result:\n\npostgres=> select sum(c) from cc;\nERROR: Unable to select an aggregate function sum(bpchar)\npostgres=> select min(c) from cc;\nmin\n----------\nabc\n(1 row)\n\nWill commit to the cvs tree sometime soon; let me know if you have any\ncontinuing problems. Good luck.\n\n - Tom",
"msg_date": "Wed, 23 Dec 1998 07:00:27 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] (stupid) bug in agg_select_candidate"
},
{
"msg_contents": ">>>> \"Thomas\" == Thomas G Lockhart <[email protected]> writes:\n\n >> I'm trying postgres 6.4.1 and I have problem with select command:\n >> > create table t( text_column text );\n >> > select sum( text_column ) from t;\n >> This command makes no sense, but causes crash in backend:\n\n Thomas> Thanks for finding this. It was in new code for automatic type matching\n Thomas> and coersion on aggregate functions to help implement aggregates for the\n Thomas> string types.\n\n Thomas> I've enclosed a simple patch which fixes the problem, and which at the\n Thomas> same time makes a nicer warning message as a result:\n\n Thomas> postgres=> select sum(c) from cc;\n Thomas> ERROR: Unable to select an aggregate function sum(bpchar)\n Thomas> postgres=> select min(c) from cc;\n Thomas> min\n Thomas> ----------\n Thomas> abc\n Thomas> (1 row)\n\n Thomas> Will commit to the cvs tree sometime soon; let me know if you have any\n Thomas> continuing problems. Good luck.\n\nThanks, works fine.\n\t\t\tHappy Xmas,\n\t\t\t\tDavid \n\n-- \n* David Sauer, student of Czech Technical University\n* electronic mail: [email protected] (mime compatible)\n",
"msg_date": "23 Dec 1998 12:49:13 +0100",
"msg_from": "David Sauer <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] (stupid) bug in agg_select_candidate"
}
] |
[
{
"msg_contents": "\nMy database (pgsql 6.4.1, linux i386, compiled by egcs-1.1.1) does weird things:\ndavid=> create table t( i int primary key );\nNOTICE: CREATE TABLE/PRIMARY KEY will create implicit index t_pkey for table t\nCREATE\ndavid=> drop table t;\nDROP\n\n*** table t doesn't exist\n\ndavid=> create table t( i int8 primary key );\nNOTICE: CREATE TABLE/PRIMARY KEY will create implicit index t_pkey for table t\nERROR: Can't find a default operator class for type 20.\n\n*** exist table t or not ?\n\ndavid=> create table t( i int primary key );\nNOTICE: CREATE TABLE/PRIMARY KEY will create implicit index t_pkey for table t\nERROR: t relation already exists\n\n*** yes!\n\ndavid=> drop table t;\nERROR: Relation t Does Not Exist!\n\n*** no!\n\nIf I restart pgsql, I can create table t (with int as primary key, not\nint8).\n\n\t\t\tHappy Xmas,\n\t\t\t\t\tDavid\n\n--\n* David Sauer, student of Czech Technical University\n* electronic mail: [email protected] (mime compatible)\n",
"msg_date": "23 Dec 1998 14:26:06 +0100",
"msg_from": "David Sauer <[email protected]>",
"msg_from_op": true,
"msg_subject": "bug(?) if int8 as primary key"
},
{
"msg_contents": "On 23 Dec 1998, David Sauer wrote:\n\n> Date: 23 Dec 1998 14:26:06 +0100\n> From: David Sauer <[email protected]>\n> To: [email protected]\n> Subject: [HACKERS] bug(?) if int8 as primary key\n> \n> \n> My database (pgsql 6.4.1, linux i386, compiled by egcs-1.1.1) does weird things:\n> david=> create table t( i int primary key );\n> NOTICE: CREATE TABLE/PRIMARY KEY will create implicit index t_pkey for table t\n> CREATE\n> david=> drop table t;\n> DROP\n> \n> *** table t doesn't exist\n\nJust tried on my linux box with pgsql 6.4 and this works ok\n\n\n> \n> david=> create table t( i int8 primary key );\n> NOTICE: CREATE TABLE/PRIMARY KEY will create implicit index t_pkey for table t\n> ERROR: Can't find a default operator class for type 20.\n> \n> *** exist table t or not ?\n> \n> david=> create table t( i int primary key );\n> NOTICE: CREATE TABLE/PRIMARY KEY will create implicit index t_pkey for table t\n> ERROR: t relation already exists\n> \n> *** yes!\n> \n> david=> drop table t;\n> ERROR: Relation t Does Not Exist!\n> \n> *** no!\n> \n> If I restart pgsql, I can create table t (with int as primary key, not\n> int8).\n> \n\nYes, this problem exists also for me in 6.4\n\n\tOleg\n\n\n> \t\t\tHappy Xmas,\n> \t\t\t\t\tDavid\n> \n> --\n> * David Sauer, student of Czech Technical University\n> * electronic mail: [email protected] (mime compatible)\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, 23 Dec 1998 17:10:16 +0300 (MSK)",
"msg_from": "Oleg Bartunov <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] bug(?) if int8 as primary key"
},
{
"msg_contents": "> david=> create table t( i int8 primary key );\n> NOTICE: CREATE TABLE/PRIMARY KEY will create implicit index t_pkey \n> for table t\n> ERROR: Can't find a default operator class for type 20.\n\nint8 is a new data type, and I did not create the functions and catalog\nentries required to implement a b-tree index yet. Since int8 relies on\ncompiler support for 8 byte integers, I wasn't sure how many platforms\nwould actually be able to compile the type. I think we know now that\nmost platforms do support the data type, so we should have more complete\nsupport for it in the next release.\n\n> If I restart pgsql, I can create table t (with int as primary key, not\n> int8).\n\nYes, that's funny. It apparently does not quite clean up after itself.\nThe good news is that you only have to restart your session, not the\npostmaster :)\n\n - Tom\n",
"msg_date": "Wed, 23 Dec 1998 14:56:15 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] bug(?) if int8 as primary key"
}
] |
[
{
"msg_contents": "> I'm trying to \"convert\" the sample db found in \"The Practical SQL \n> Handbook\" Bowman, et al. to postgres. When trying to create the \n> following table I get \"parser: parse error at or near 'null'\" \n> create table authors (\n> zip char(5) null\n> );\n> The table creation works fine twith the NOT NULL but won't work with \n> the NULL. NULL is supported isn't it?\n\nThe NULL constraint syntax is *not* supported, since it results in\nshift/reduce conflicts in our yacc parser. This is because the token is\nambiguous with other uses of NULL in the same area, at least as far as\nyacc is concerned.\n\nHowever, the default behavior for all columns is to allow NULL values,\nso it is a noise word which can be omitted without ill effect.\n\n - Tom\n",
"msg_date": "Wed, 23 Dec 1998 15:38:14 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: NULL & NOT NULL "
},
{
"msg_contents": "> > > I'm trying to \"convert\" the sample db found in \"The Practical SQL\n> > > Handbook\" Bowman, et al. to postgres. When trying to create the\n> > > following table I get \"parser: parse error at or near 'null'\"\n> > > create table authors (\n> > > zip char(5) null\n> > > );\n> > > The table creation works fine with the NOT NULL but won't work \n> > > with the NULL. NULL is supported isn't it?\n> > The NULL constraint syntax is *not* supported, since it results in\n> > shift/reduce conflicts in our yacc parser. This is because the token \n> > is ambiguous with other uses of NULL in the same area, at least as \n> > far as yacc is concerned.\n\nSheesh. After that long song and dance about why we can't implement\nthis, it turns out that it works fine. We had been trying to implement a\nslightly different syntax, \"WITH NULL\", which conflicted with the\nSQL92-defined data type declaration \"TIMESTAMP WITH TIME ZONE\".\n\nThe \"Practical SQL Handbook\"-compatible form will be available in the\nnext full release of Postgres. Thanks.\n\n - Tom\n",
"msg_date": "Thu, 24 Dec 1998 14:05:59 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: NULL & NOT NULL"
},
{
"msg_contents": "\"Thomas G. Lockhart\" <[email protected]> writes:\n>>>>>> create table authors (\n>>>>>> zip char(5) null\n>>>>>> );\n\n> Sheesh. After that long song and dance about why we can't implement\n> this, it turns out that it works fine. We had been trying to implement a\n> slightly different syntax, \"WITH NULL\", which conflicted with the\n> SQL92-defined data type declaration \"TIMESTAMP WITH TIME ZONE\".\n\n> The \"Practical SQL Handbook\"-compatible form will be available in the\n> next full release of Postgres. Thanks.\n\nNow that we have the syntax problem straightened out: I'm still confused\nabout the semantics. Does a \"NULL\" constraint say that the field\n*must* be null, or only that it *can* be null (in which case NULL is\njust a noise word, since that's the default condition)? I had assumed\nthe former, but Bruce seemed to think the latter...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Thu, 24 Dec 1998 09:33:18 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: NULL & NOT NULL "
},
{
"msg_contents": "> \"Thomas G. Lockhart\" <[email protected]> writes:\n> >>>>>> create table authors (\n> >>>>>> zip char(5) null\n> >>>>>> );\n> \n> > Sheesh. After that long song and dance about why we can't implement\n> > this, it turns out that it works fine. We had been trying to implement a\n> > slightly different syntax, \"WITH NULL\", which conflicted with the\n> > SQL92-defined data type declaration \"TIMESTAMP WITH TIME ZONE\".\n> \n> > The \"Practical SQL Handbook\"-compatible form will be available in the\n> > next full release of Postgres. Thanks.\n> \n> Now that we have the syntax problem straightened out: I'm still confused\n> about the semantics. Does a \"NULL\" constraint say that the field\n> *must* be null, or only that it *can* be null (in which case NULL is\n> just a noise word, since that's the default condition)? I had assumed\n> the former, but Bruce seemed to think the latter...\n\n\nCan be null. Noise word. At least that is what I rememeber Thomas\nsaying, and because it was noise, we removed it. In fact, it doesn't\nlook like the standard accepts it, but there is no reason we can't.\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, 24 Dec 1998 10:47:27 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: NULL & NOT NULL"
},
{
"msg_contents": "> > > The \"Practical SQL Handbook\"-compatible form will be available in \n> > > the next full release of Postgres. Thanks.\n> > Now that we have the syntax problem straightened out: I'm still \n> > confused about the semantics.\n> Can be null. Noise word. At least that is what I rememeber Thomas\n> saying, and because it was noise, we removed it. In fact, it doesn't\n> look like the standard accepts it, but there is no reason we can't.\n\nYeah, it's noise. And not in SQL92. I've flagged places in gram.y which\nare syntax extensions included just to be compatible with specific\nproducts so that if we ever run into parser conflicts with them we can\nyank them out. The M$ usage for \"where var = NULL\" rather than the SQL\nstandard usage \"where var is null\" is another example of this.\n\nI don't own \"The Practical SQL Handbook\", but it is funny that it\ncontains examples which are not part of the SQL standard (this is the\nsecond one as I recall; can't remember the first one though).\n\n - Tom\n",
"msg_date": "Mon, 28 Dec 1998 15:50:57 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Re: NULL & NOT NULL"
},
{
"msg_contents": "\nOn 28-Dec-98 Thomas G. Lockhart wrote:\n\n> I don't own \"The Practical SQL Handbook\", but it is funny that it\n> contains examples which are not part of the SQL standard (this is the\n> second one as I recall; can't remember the first one though).\n\nI've got the book if ya need something looked up or whatever.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Searchable Campground Listings http://www.camping-usa.com\n \"There is no outfit less entitled to lecture me about bloat\n than the federal government\" -- Tony Snow\n==========================================================================\n\n\n",
"msg_date": "Mon, 28 Dec 1998 11:19:15 -0500 (EST)",
"msg_from": "Vince Vielhaber <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: NULL & NOT NULL"
},
{
"msg_contents": "Bruce Momjian wrote:\n> \n> > \"Thomas G. Lockhart\" <[email protected]> writes:\n> > >>>>>> create table authors (\n> > >>>>>> zip char(5) null\n> > >>>>>> );\n> >\n> > > Sheesh. After that long song and dance about why we can't implement\n> > > this, it turns out that it works fine. We had been trying to implement a\n> > > slightly different syntax, \"WITH NULL\", which conflicted with the\n> > > SQL92-defined data type declaration \"TIMESTAMP WITH TIME ZONE\".\n> >\n> > > The \"Practical SQL Handbook\"-compatible form will be available in the\n> > > next full release of Postgres. Thanks.\n> >\n> > Now that we have the syntax problem straightened out: I'm still confused\n> > about the semantics. Does a \"NULL\" constraint say that the field\n> > *must* be null, or only that it *can* be null (in which case NULL is\n> > just a noise word, since that's the default condition)? I had assumed\n> > the former, but Bruce seemed to think the latter...\n> \n> Can be null. Noise word. At least that is what I rememeber Thomas\n> saying, and because it was noise, we removed it. In fact, it doesn't\n> look like the standard accepts it, but there is no reason we can't.\n\n\nThis NULL clause is not part of constraints it is a default option and\nwe already support it,\nthere's nothing like: \nCREATE TABLE table1 (field1 type NULL) in SQL92.\n\nbut the following is SQL92 and it works on PostgreSQL:\n \nprova=> CREATE TABLE table1 (field1 INTEGER DEFAULT NULL);\nCREATE\n\n\nSQL92 syntax:\n\n 11.5 <default clause>\n\n Function\n\n Specify the default for a column or domain.\n\n Format\n\n <default clause> ::=\n DEFAULT <default option>\n\n <default option> ::=\n <literal>\n | <datetime value function>\n | USER\n | CURRENT_USER\n | SESSION_USER\n | SYSTEM_USER\n | NULL\n\n-Jose'-\n",
"msg_date": "Tue, 29 Dec 1998 15:43:13 +0100",
"msg_from": "\"Jose' Soares\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: NULL & NOT NULL"
},
{
"msg_contents": "> This NULL clause is not part of constraints it is a default option and\n> we already support it,\n> prova=> CREATE TABLE table1 (field1 INTEGER DEFAULT NULL);\n> CREATE\n\nThat is certainly a clearer way of specifying it. Should we forget about\nthe other syntax?\n\n - Tom\n",
"msg_date": "Tue, 29 Dec 1998 15:44:29 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Re: NULL & NOT NULL"
},
{
"msg_contents": "Thomas G. Lockhart wrote:\n> \n> > This NULL clause is not part of constraints it is a default option and\n> > we already support it,\n> > prova=> CREATE TABLE table1 (field1 INTEGER DEFAULT NULL);\n> > CREATE\n> \n> That is certainly a clearer way of specifying it. Should we forget about\n> the other syntax?\n> \n> - Tom\n\nImho yes. The syntax CREATE TABLE table (field type NULL) has no sense\nthe NULL keyword may be used on a DEFAULT clause (if you want to specify\na default value)\nor on a column constraint (if you want to avoid data integrity\nviolation).\n\n1) Column Constraint definition:\n [ CONSTRAINT name ] NOT NULL\n \n2) Default clause:\n DEFAULT NULL\n\n-Jose'-\n",
"msg_date": "Wed, 30 Dec 1998 16:14:29 +0100",
"msg_from": "\"Jose' Soares\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: NULL & NOT NULL"
}
] |
[
{
"msg_contents": "Can we *really* fix up the docs list, or kill it? People ask what list\nthey should be posting on for docs questions, and I'm not sure what\nanswer I should give them. Without it being archived, it means that\nthere is no searchable list of answers to questions, so I've been\nrecommending using the hackers list instead.\n\nHeck, on the web page it isn't even listed anymore afaik.\n\nbtw, the web pages are looking great. A couple of small things:\n\nThe \n Info Central->Searchable Mailing List:PostgreSQL.org\npage may have some formatting problems in that all the available lists\nare written across the top of my browser as one line.\n\nAlso, the mhonarc thread listings are showing a line\n BODY BGCOLOR=\"#FFFDEC\">\nat the top of each page.\n\nI'm running Netscape-4.07, which is fairly recent...\n\nTIA\n\n - Tom\n",
"msg_date": "Wed, 23 Dec 1998 16:14:58 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Mhonarc for the docs list"
},
{
"msg_contents": "Hi!\n\nOn Wed, 23 Dec 1998, Thomas G. Lockhart wrote:\n> Also, the mhonarc thread listings are showing a line\n> BODY BGCOLOR=\"#FFFDEC\">\n> at the top of each page.\n> \n> I'm running Netscape-4.07, which is fairly recent...\n\n It don't depends on browser - generated HTML just left '<'.\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": "Wed, 23 Dec 1998 19:20:23 +0300 (MSK)",
"msg_from": "Oleg Broytmann <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Mhonarc for the docs list"
},
{
"msg_contents": "\nOn 23-Dec-98 Thomas G. Lockhart wrote:\n> Can we *really* fix up the docs list, or kill it? People ask what list\n> they should be posting on for docs questions, and I'm not sure what\n> answer I should give them. Without it being archived, it means that\n> there is no searchable list of answers to questions, so I've been\n> recommending using the hackers list instead.\n> \n> Heck, on the web page it isn't even listed anymore afaik.\n\nSure. Now that I subscribe you mention killing it! :)\n\nAnd no it's not listed under mailing lists.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Searchable Campground Listings http://www.camping-usa.com\n \"There is no outfit less entitled to lecture me about bloat\n than the federal government\" -- Tony Snow\n==========================================================================\n\n\n",
"msg_date": "Wed, 23 Dec 1998 11:30:11 -0500 (EST)",
"msg_from": "Vince Vielhaber <[email protected]>",
"msg_from_op": false,
"msg_subject": "RE: [HACKERS] Mhonarc for the docs list"
},
{
"msg_contents": "\n\"Thomas G. Lockhart\" <[email protected]> writes:\n\n> Can we *really* fix up the docs list, or kill it? People ask what list\n> they should be posting on for docs questions, and I'm not sure what\n> answer I should give them. Without it being archived, it means that\n> there is no searchable list of answers to questions, so I've been\n> recommending using the hackers list instead.\n\nI have a pretty good feel for which HTML files need revision - need\nMarc to fix or turn me loose on several points of breakage in the\narchiving scripts. [nudges Marc, who has understandably been busy with\nother things lately]\n\n> Heck, on the web page it isn't even listed anymore afaik.\n> \n> btw, the web pages are looking great. A couple of small things:\n> \n> The \n> Info Central->Searchable Mailing List:PostgreSQL.org\n> page may have some formatting problems in that all the available lists\n> are written across the top of my browser as one line.\n\nGak. Will look at it ASAP.\n\n> Also, the mhonarc thread listings are showing a line\n> BODY BGCOLOR=\"#FFFDEC\">\n> at the top of each page.\n\nKnown breakage in archiving scripts. See above.\n\n> I'm running Netscape-4.07, which is fairly recent...\n\nPlan is to change the look of the site gradually to make it hang\ntogether better and put most frequently accessed info toward the top\nof the site tree. I've been phasing out .shtml and phasing in CSS.\n\nThanks for the feedback on the web site.\n",
"msg_date": "23 Dec 1998 10:39:16 -0600",
"msg_from": "Hal Snyder <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Mhonarc for the docs list"
},
{
"msg_contents": "On 23 Dec 1998, Hal Snyder wrote:\n\n> \n> \"Thomas G. Lockhart\" <[email protected]> writes:\n> \n> > Can we *really* fix up the docs list, or kill it? People ask what list\n> > they should be posting on for docs questions, and I'm not sure what\n> > answer I should give them. Without it being archived, it means that\n> > there is no searchable list of answers to questions, so I've been\n> > recommending using the hackers list instead.\n> \n> I have a pretty good feel for which HTML files need revision - need\n> Marc to fix or turn me loose on several points of breakage in the\n> archiving scripts. [nudges Marc, who has understandably been busy with\n> other things lately]\n\n\tJust an FYI...I'm not home right now...at folks spending\nXmas...get back on the 27th, have the 28th off, so will be spending time\ngetting the 'missing lists' mhonarc'd and indexed...\n\n> Plan is to change the look of the site gradually to make it hang\n> together better and put most frequently accessed info toward the top\n> of the site tree. I've been phasing out .shtml and phasing in CSS.\n\n\tAny further word from/on that Italian company that wanted to do a\nrebuild for us? Last I recall, you were going to send them a site map to\nwork off of? *raised eyebrow*\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Thu, 24 Dec 1998 16:00:10 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Mhonarc for the docs list"
},
{
"msg_contents": "On 23 Dec 1998, Hal Snyder wrote:\n\n> Plan is to change the look of the site gradually to make it hang\n> together better and put most frequently accessed info toward the top\n> of the site tree. I've been phasing out .shtml and phasing in CSS.\n\nThat would be good for me. I have spare web space that I can provide a UK\nmirror, except I can't handle .shtml - CSS is possible however.\n\nPS: My details in the developers page seems to have reverted back to an\nold email address (fortunately they still work).\n\nThe correct email address should be [email protected]\n\n-- \n Peter T Mount [email protected]\n Main Homepage: http://www.retep.org.uk\nPostgreSQL JDBC Faq: http://www.retep.org.uk/postgres\n Java PDF Generator: http://www.retep.org.uk/pdf\n\n",
"msg_date": "Mon, 28 Dec 1998 16:40:12 +0000 (GMT)",
"msg_from": "Peter T Mount <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Mhonarc for the docs list"
}
] |
[
{
"msg_contents": "Hello\n\n\tIn 6.4.1, PG_SUBVERSION is still \"0\" in version.h.\n\nRegards\n\n-- \nASCII CORPORATION\nTechnical Center\nSHIOZAKI Takehiko\n<[email protected]>\n",
"msg_date": "Thu, 24 Dec 1998 10:16:37 +0900 (JST)",
"msg_from": "SHIOZAKI Takehiko <[email protected]>",
"msg_from_op": true,
"msg_subject": "PG_SUBVERSION"
},
{
"msg_contents": "> Hello\n> \n> \tIn 6.4.1, PG_SUBVERSION is still \"0\" in version.h.\n> \n> Regards\n\nYes, someone pointed this out today, and I have added the SUBVERSION\nmention to my checklist. Perhaps we can repackage 6.4.1 when Marc\nreturns. It only affect select pg_version().\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, 23 Dec 1998 20:59:20 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] PG_SUBVERSION"
}
] |
[
{
"msg_contents": "HISTORY file coming with 6.4.1 seems old one. I tried\nhttp://www.postgresql.org/ to get latest HISTORY but it miserably\nfailed.\n\nAlso, I noticed that the name of 6.4.1 source directory is \"pgsql\" not\npostgresql-v6.4.1 or postgresql-6.4.1. I'm afraid this confuses users.\n--\nTatsuo Ishii\[email protected]\n",
"msg_date": "Thu, 24 Dec 1998 10:42:25 +0900",
"msg_from": "Tatsuo Ishii <[email protected]>",
"msg_from_op": true,
"msg_subject": "6.4.1 HISTORY etc."
},
{
"msg_contents": "> HISTORY file coming with 6.4.1 seems old one. I tried\n> http://www.postgresql.org/ to get latest HISTORY but it miserably\n> failed.\n\nsystem was down today. It is up now. I think it is up-to-date.\n\n> \n> Also, I noticed that the name of 6.4.1 source directory is \"pgsql\" not\n> postgresql-v6.4.1 or postgresql-6.4.1. I'm afraid this confuses users.\n> --\n> Tatsuo Ishii\n> [email protected]\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": "Wed, 23 Dec 1998 20:59:59 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.1 HISTORY etc."
},
{
"msg_contents": ">> HISTORY file coming with 6.4.1 seems old one. I tried\n>> http://www.postgresql.org/ to get latest HISTORY but it miserably\n>> failed.\n>\n>system was down today. It is up now. I think it is up-to-date.\n\nHumm... I cannot find any description regarding changes between 6.4\nand 6.4.1 in the HISTORY file?\n\n\nPostgreSQL 6.4 Mon Oct 26 11:40:42 EST 1998\n=========================================================\n\nA dump/restore is required for those wishing to migrate data from\nprevious releases of PostgreSQL. You may also use the new pg_upgrade\nutility to avoid a dump/reload.\n\nBug Fixes\n---------\n:\n:\n--\nTatsuo Ishii\[email protected]\n",
"msg_date": "Thu, 24 Dec 1998 11:08:40 +0900",
"msg_from": "Tatsuo Ishii <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] 6.4.1 HISTORY etc. "
},
{
"msg_contents": "> >> HISTORY file coming with 6.4.1 seems old one. I tried\n> >> http://www.postgresql.org/ to get latest HISTORY but it miserably\n> >> failed.\n> >\n> >system was down today. It is up now. I think it is up-to-date.\n> \n> Humm... I cannot find any description regarding changes between 6.4\n> and 6.4.1 in the HISTORY file?\n> \n\nHow can that be. I see 6.4.1 at the top of the HISTORY file in my\nRELEASE tree. Now, my current tree does not have that, because we are\nnot using that tree for a release right now, though I guess I should\nupdate it so it stays current for 6.5.\n\ncvs log shows for HISTORY shows:\n\n\trevision 1.44.2.3\n\tdate: 1998/12/18 16:51:39; author: momjian; state: Exp; lines: +5 -6\n\tUpgrade for 6.4.1.\n\nI wonder if Marc packaged the CURRENT by accident. I can only guess\nwhat has happened.\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, 24 Dec 1998 00:20:15 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.1 HISTORY etc."
},
{
"msg_contents": "On Thu, 24 Dec 1998, Bruce Momjian wrote:\n\n> > >> HISTORY file coming with 6.4.1 seems old one. I tried\n> > >> http://www.postgresql.org/ to get latest HISTORY but it miserably\n> > >> failed.\n> > >\n> > >system was down today. It is up now. I think it is up-to-date.\n> > \n> > Humm... I cannot find any description regarding changes between 6.4\n> > and 6.4.1 in the HISTORY file?\n> > \n> \n> How can that be. I see 6.4.1 at the top of the HISTORY file in my\n> RELEASE tree. Now, my current tree does not have that, because we are\n> not using that tree for a release right now, though I guess I should\n> update it so it stays current for 6.5.\n> \n> cvs log shows for HISTORY shows:\n> \n> \trevision 1.44.2.3\n> \tdate: 1998/12/18 16:51:39; author: momjian; state: Exp; lines: +5 -6\n> \tUpgrade for 6.4.1.\n> \n> I wonder if Marc packaged the CURRENT by accident. I can only guess\n> what has happened.\n\n\t*sigh* You are right...I futzed it :( Am re-packaging it from\nscratch now, will send an email out when fixed :(\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Thu, 24 Dec 1998 16:03:27 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.1 HISTORY etc."
},
{
"msg_contents": "> > I wonder if Marc packaged the CURRENT by accident. I can only guess\n> > what has happened.\n> \n> \t*sigh* You are right...I futzed it :( Am re-packaging it from\n> scratch now, will send an email out when fixed :(\n\nI have fixed some snprintf problems since Monday, so it is good we are\nrepackaging. What do we call it? 6.4.2?\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, 24 Dec 1998 16:11:00 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.1 HISTORY etc."
},
{
"msg_contents": "On Thu, 24 Dec 1998, Bruce Momjian wrote:\n\n> > > I wonder if Marc packaged the CURRENT by accident. I can only guess\n> > > what has happened.\n> > \n> > \t*sigh* You are right...I futzed it :( Am re-packaging it from\n> > scratch now, will send an email out when fixed :(\n> \n> I have fixed some snprintf problems since Monday, so it is good we are\n> repackaging. What do we call it? 6.4.2?\n\n\tI just left it as 6.4.1...saves Hal having to through and fixing\n*all* the links again :)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Thu, 24 Dec 1998 21:12:25 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.1 HISTORY etc."
},
{
"msg_contents": "> On Thu, 24 Dec 1998, Bruce Momjian wrote:\n> \n> > > > I wonder if Marc packaged the CURRENT by accident. I can only guess\n> > > > what has happened.\n> > > \n> > > \t*sigh* You are right...I futzed it :( Am re-packaging it from\n> > > scratch now, will send an email out when fixed :(\n> > \n> > I have fixed some snprintf problems since Monday, so it is good we are\n> > repackaging. What do we call it? 6.4.2?\n> \n> \tI just left it as 6.4.1...saves Hal having to through and fixing\n> *all* the links again :)\n\nI agree. How many people downloaded the bad version. Can you find 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": "Thu, 24 Dec 1998 21:13:33 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.1 HISTORY etc."
}
] |
[
{
"msg_contents": "\nDue to an error on my part, I mistakenly packaged up the -CURRENT source\ntree as being v6.4.1, vs the actual v6.4.1 source tree...\n\nThis has *just* been fixed as of two minutes ago...\n\nFor those running production environments, *please* download this new\npackage and rebuild :(\n\nSorry for the inconvience that this is going to cause most ppl...\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n\n\n",
"msg_date": "Thu, 24 Dec 1998 16:13:28 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": true,
"msg_subject": "[HACKERS] [URGENT] v6.4.1 was a mis-package..."
},
{
"msg_contents": "On Thu, 24 Dec 1998, The Hermit Hacker wrote:\n\n> \n> Due to an error on my part, I mistakenly packaged up the -CURRENT source\n> tree as being v6.4.1, vs the actual v6.4.1 source tree...\n> \n> This has *just* been fixed as of two minutes ago...\n> \n> For those running production environments, *please* download this new\n> package and rebuild :(\n\nAre there any big flaws in it which could result in data loss? I ran some\nupdates just yesterday with the what-I-thought-would-be 6.4.1. \n\n Regards,\n\n Sascha Schumann |\n Consultant | finger [email protected]\n | for PGP public key\n\n",
"msg_date": "Thu, 24 Dec 1998 21:36:20 +0100 (CET)",
"msg_from": "Sascha Schumann <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] [URGENT] v6.4.1 was a mis-package..."
},
{
"msg_contents": "> On Thu, 24 Dec 1998, The Hermit Hacker wrote:\n> \n> > \n> > Due to an error on my part, I mistakenly packaged up the -CURRENT source\n> > tree as being v6.4.1, vs the actual v6.4.1 source tree...\n> > \n> > This has *just* been fixed as of two minutes ago...\n> > \n> > For those running production environments, *please* download this new\n> > package and rebuild :(\n> \n> Are there any big flaws in it which could result in data loss? I ran some\n> updates just yesterday with the what-I-thought-would-be 6.4.1. \n\nGood question. We have some new stuff in 6.5 tree, and none of it\ntested. The only thing that I think could be a problem is Vadim's\nchanges to the structure of the row headers. This could majorly break\nthings internally.\n\nThe big problem is that the data rows were created with a different\nformat than the backend expected. At least that is my conclusion based\non what Vadim had done.\n\nA dump/restore should may fix any breakage, but you may have to restore\nand re-run the updates.\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, 24 Dec 1998 21:12:44 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] [URGENT] v6.4.1 was a mis-package..."
}
] |
[
{
"msg_contents": "Here are the list of changes for 6.4.1.\n\nIf your 6.4.1 pgsql/HISTORY file does not contain this at the top, you\nhave a faulty 6.4.1 distribution. I recommend you go to\nftp.postgresql.org or www.postgresql.org to get a new one.\n\n\n---------------------------------------------------------------------------\n\nPostgreSQL 6.4.1 Fri Dec 18 00:29:11 EST 1998\n=========================================================\n\nA dump/restore is NOT required for those running 6.4. Those running\npre-6.4 require a dump/restore, or for 6.3.* users, the use of the\npg_upgrade utility.\n\nChanges\n-------\nAdd pg_dump -N flag to force double quotes around identifiers. This is\n\tthe default(Thomas)\nFix for NOT in where clause causing crash(Bruce)\nEXPLAIN VERBOSE coredump fix(Vadim)\nFix shared-library problems on Linux\nFix test for table existance to allow mixed-case and whitespace in\n\tthe table name(Thomas)\nFix a couple of pg_dump bugs\nConfigure matches template/.similar entries better(Tom)\nChange builtin function names from SPI_* to spi_*\nOR WHERE clause fix(Vadim)\nFixes for mixed-case table names(Billy)\ncontrib/linux/postgres.init.csh/sh fix(Thomas)\nlibpq memory overrun fix\nSunOS fixes(Tom)\nChange exp() behavior to generate error on underflow(Thomas)\npg_dump fixes for memory leak, inheritance constraints, layout change\nupdate pgaccess to 0.93\nFix prototype for 64-bit platforms\nMulti-byte fixes(Tatsuo)\nNew ecpg man page\nFix memory overruns(Tatsuo)\nFix for lo_import() crash(Bruce)\nBetter search for install program(Tom)\nTimezone fixes(Tom)\nHPUX fixes(Tom)\nUse implicit type coersion for matching DEFAULT values(Thomas)\nAdd routines to help with single-byte (internal) character type(Thomas)\nCompilation of libpq for Win32 fixes(Magnus)\nUpgrade to PyGreSQL 2.2(D'Arcy)\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, 24 Dec 1998 21:32:25 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Release of 6.4.1"
},
{
"msg_contents": "On Thu, 24 Dec 1998, Bruce Momjian wrote:\n\n> Here are the list of changes for 6.4.1.\n> \n> Fix prototype for 64-bit platforms\n\nIf this is the prototyping problem in backend/port/snprintf.c that I\ncontributed, it did not help the problem of compiling 6.4.1 on Digital\nUnix 4.0d using either gcc or cc. Here is my environment:\n\n gcc 2.8.1\n gmake GNU Make version 3.77\n Digital cc 4.25\n\nTo set up cc as the compiler, I used the following command:\n\n ./configure --with-CC=cc --without-CXX\n\nThis failed, so I did the following:\n\n gmake distclean\n ./configure\n\nThis also failed, but for a different cause.\n\nI am attaching transcripts of the gmake session for each of the compilers. \nCan anyone provide a suggestion to get 6.4.1 to compile?\n\nThanks,\nCarl\n\nCarl G. Riches\nSoftware Engineer\nDepartment of Mathematics\nBox 354350\t\t\tvoice: 206-543-5082 or 206-616-3636\nUniversity of Washington\tfax: 206-543-0397\nSeattle, WA 98195-4350\t\tinternet: [email protected]\n\n",
"msg_date": "Mon, 28 Dec 1998 17:05:50 -0800 (PST)",
"msg_from": "Carl Riches <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [GENERAL] Release of 6.4.1"
}
] |
[
{
"msg_contents": "The web page, under NEWS says about 6.4.1:\n\n\tTue Dec 22, 1998 - PostgreSQL v6.4.1 Released. This release is a\n\tmaintenance release. There have been no features added to it, only fixes\n\tagainst v6.4. If you are having problems, please download and install\n\tv6.4.1 before reporting any bugs. If you aren't having any problems,\n\tthere is no reason to download/install this version. \n\nI recommend everyone upgrading to 6.4.1, even if they have not seen\nproblems. I think they will see problems at some point, and the upgrade\ndoes not require a lot of work. The big fixes are:\n\n\tEXPLAIN VERBOSE coredump fix(Vadim)\n\tFix shared-library problems on Linux\n\tFix a couple of pg_dump bugs\n\tOR WHERE clause fix(Vadim)\n\tFix for lo_import() crash(Bruce)\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, 24 Dec 1998 22:03:41 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "6.4.1 and web page"
}
] |
[
{
"msg_contents": "Hi,\n\nit would have been a much cleaner solution to increment\nthe version number after realizing, that a wrong version\nhas been distributed.\nTo distribute another software with the same version\nnumber - even if the wrong version was available only for \na very short time - is definitely a no no.\n\nEdmund\n-- \nEdmund Mergl mailto:[email protected]\nIm Haldenhau 9 http://www.bawue.de/~mergl\n70565 Stuttgart fon: +49 711 747503\nGermany\n",
"msg_date": "Fri, 25 Dec 1998 09:35:13 +0100",
"msg_from": "Edmund Mergl <[email protected]>",
"msg_from_op": true,
"msg_subject": "postgresql-6.4.1.tar.gz"
},
{
"msg_contents": "[Charset iso-8859-2 unsupported, filtering to ASCII...]\n> Hi,\n> \n> it would have been a much cleaner solution to increment\n> the version number after realizing, that a wrong version\n> has been distributed.\n> To distribute another software with the same version\n> number - even if the wrong version was available only for \n> a very short time - is definitely a no no.\n\nIt was Marc's call.\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, 25 Dec 1998 06:22:01 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] postgresql-6.4.1.tar.gz"
},
{
"msg_contents": "> > it would have been a much cleaner solution to increment\n> > the version number after realizing, that a wrong version\n> > has been distributed.\n> > To distribute another software with the same version\n> > number - even if the wrong version was available only for\n> > a very short time - is definitely a no no.\n> It was Marc's call.\n\nBut still not too late to fix. I think Edmund has a very valid point. We\nshould be releasing v6.4.2 to eliminate any ambiguity in which version\npeople are running.\n\nIt is much easier to say: \"If you are running v6.4.1, then kill it and\ninstall v6.4.2.\"\n\nThis is important imho. Sorry I was out of town and not contributing to\nthe discussion earlier...\n\n - Tom\n",
"msg_date": "Mon, 28 Dec 1998 16:01:31 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] postgresql-6.4.1.tar.gz"
},
{
"msg_contents": "Thomas G. Lockhart wrote:\n> \n> It is much easier to say: \"If you are running v6.4.1, then kill it and\n> install v6.4.2.\"\n\nBTW , 6.4.1 didn't solved the -lcrypt problem in Makefile for libpgtcl\n:-(\n\nI had to add it manually. I was testing that on two different computers,\nwith RedHat 5.1 and 5.2 versions, still the same problem.\n\nAs I recall, there was someone at a moment that claimed that the problem\nwas solved, he send me a patch and I tested it and it worked. Was that\npatch included at that time ?\n\n-- \nConstantin Teodorescu\nFLEX Consulting Braila, ROMANIA\n",
"msg_date": "Mon, 28 Dec 1998 18:13:58 +0200",
"msg_from": "Constantin Teodorescu <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] postgresql-6.4.1.tar.gz"
}
] |
[
{
"msg_contents": "Long awaited port for NetBSD/m68k was finally done by Mr. Mutsuki\nNakajima. Since he is not subscribing the mailing list, I'm posting\nhis patches by his request. According to him, he has successfully\ncompiled and passed the regression test on Mac SE/30 running\nNetBSD/m68k. Also, another person has reported that with the patches\nPostgreSQL is working on NetBSD/sun3 too.\n--\nTatsuo Ishii\[email protected]\n---------------------------- cut here ----------------------------\ndiff -crP postgresql-v6.4.org/src/include/utils/memutils.h postgresql-v6.4/src/include/utils/memutils.h\n*** postgresql-v6.4.org/src/include/utils/memutils.h\tMon Sep 7 14:35:48 1998\n--- postgresql-v6.4/src/include/utils/memutils.h\tSat Dec 26 05:51:30 1998\n***************\n*** 58,70 ****\n #define SHORTALIGN(LEN)\\\n \t\t(((long)(LEN) + (sizeof (short) - 1)) & ~(sizeof (short) - 1))\n \n #define INTALIGN(LEN)\\\n \t\t(((long)(LEN) + (sizeof (int) - 1)) & ~(sizeof (int) -1))\n \n /*\n *\t\tLONGALIGN(LEN)\t- length (or address) aligned for longs\n */\n! #if defined(sun) && ! defined(sparc)\n #define LONGALIGN(LEN)\tSHORTALIGN(LEN)\n #elif defined (__alpha)\n \n--- 58,74 ----\n #define SHORTALIGN(LEN)\\\n \t\t(((long)(LEN) + (sizeof (short) - 1)) & ~(sizeof (short) - 1))\n \n+ #if defined(m68k)\n+ #define INTALIGN(LEN)\t SHORTALIGN(LEN)\n+ #else\n #define INTALIGN(LEN)\\\n \t\t(((long)(LEN) + (sizeof (int) - 1)) & ~(sizeof (int) -1))\n+ #endif\n \n /*\n *\t\tLONGALIGN(LEN)\t- length (or address) aligned for longs\n */\n! #if (defined(sun) && ! defined(sparc)) || defined(m68k)\n #define LONGALIGN(LEN)\tSHORTALIGN(LEN)\n #elif defined (__alpha)\n \n***************\n*** 81,87 ****\n \t\t(((long)(LEN) + (sizeof (long) - 1)) & ~(sizeof (long) -1))\n #endif\n \n! #if ! defined(sco)\n #define DOUBLEALIGN(LEN)\\\n \t\t(((long)(LEN) + (sizeof (double) - 1)) & ~(sizeof (double) -1))\n \n--- 85,94 ----\n \t\t(((long)(LEN) + (sizeof (long) - 1)) & ~(sizeof (long) -1))\n #endif\n \n! #if defined(m68k)\n! #define DOUBLEALIGN(LEN) SHORTALIGN(LEN)\n! #define MAXALIGN(LEN)\t SHORTALIGN(LEN)\n! #elif ! defined(sco)\n #define DOUBLEALIGN(LEN)\\\n \t\t(((long)(LEN) + (sizeof (double) - 1)) & ~(sizeof (double) -1))\n \ndiff -crP postgresql-v6.4.org/src/backend/storage/buffer/s_lock.c postgresql-v6.4/src/backend/storage/buffer/s_lock.c\n*** postgresql-v6.4.org/src/backend/storage/buffer/s_lock.c\tSat Sep 19 02:18:39 1998\n--- postgresql-v6.4/src/backend/storage/buffer/s_lock.c\tSat Dec 26 06:26:24 1998\n***************\n*** 117,122 ****\n--- 117,141 ----\n \n #endif\t /* PPC */\n \n+ #if defined(__m68k__)\n+ static void\n+ tas_dummy()\t/* really means: extern int tas(slock_t **lock); */\n+ {\n+ \t__asm__(\"\t\t\\n\\\n+ .global\t\t_tas\t\t\\n\\\n+ _tas:\t\t\t\t\\n\\\n+ \tmovel sp@(0x4),a0\t\\n\\\n+ \ttas a0@\t\t\t\\n\\\n+ \tbeq _success\t\t\\n\\\n+ \tmoveq #-128,d0\t\\n\\\n+ \trts\t\t\t\\n\\\n+ _success:\t\t\t\\n\\\n+ \tmoveq #0,d0\t\t\\n\\\n+ \trts\t\t\t\\n\\\n+ \t\");\n+ }\n+ \n+ #endif\t /* __m68k__ */\n \n \n #else\t\t\t\t\t\t\t/* defined(__GNUC__) */\n\n",
"msg_date": "Sun, 27 Dec 1998 00:11:06 +0900",
"msg_from": "Tatsuo Ishii <[email protected]>",
"msg_from_op": true,
"msg_subject": "NetBSD/m68k port"
},
{
"msg_contents": "Patch applied.\n\n> Long awaited port for NetBSD/m68k was finally done by Mr. Mutsuki\n> Nakajima. Since he is not subscribing the mailing list, I'm posting\n> his patches by his request. According to him, he has successfully\n> compiled and passed the regression test on Mac SE/30 running\n> NetBSD/m68k. Also, another person has reported that with the patches\n> PostgreSQL is working on NetBSD/sun3 too.\n> --\n> Tatsuo Ishii\n> [email protected]\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": "Sat, 26 Dec 1998 13:15:21 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [PATCHES] NetBSD/m68k port"
},
{
"msg_contents": "This sounds like an item worthy of a newsflash on the web site. I'll\nadd one if you like. What's the approved way of getting the\npatch/release in question? How do we describe present status? Beta?\n\nBruce Momjian <[email protected]> writes:\n\n> Patch applied.\n> \n> > Long awaited port for NetBSD/m68k was finally done by Mr. Mutsuki\n> > Nakajima. Since he is not subscribing the mailing list, I'm posting\n> > his patches by his request. According to him, he has successfully\n> > compiled and passed the regression test on Mac SE/30 running\n> > NetBSD/m68k. Also, another person has reported that with the patches\n> > PostgreSQL is working on NetBSD/sun3 too.\n> > --\n> > Tatsuo Ishii\n> > [email protected]\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---\nHal / [email protected]\n",
"msg_date": "26 Dec 1998 13:59:19 -0600",
"msg_from": "Hal Snyder <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: [PATCHES] NetBSD/m68k port"
},
{
"msg_contents": "> This sounds like an item worthy of a newsflash on the web site. I'll\n> add one if you like. What's the approved way of getting the\n> patch/release in question? How do we describe present status? Beta?\n> \n> Bruce Momjian <[email protected]> writes:\n> \n> > Patch applied.\n> > \n> > > Long awaited port for NetBSD/m68k was finally done by Mr. Mutsuki\n> > > Nakajima. Since he is not subscribing the mailing list, I'm posting\n> > > his patches by his request. According to him, he has successfully\n> > > compiled and passed the regression test on Mac SE/30 running\n> > > NetBSD/m68k. Also, another person has reported that with the patches\n> > > PostgreSQL is working on NetBSD/sun3 too.\n\nWe would have to have it packaged, and put in the patches directory on\nthe ftp site. Then we can just tell them to look in the patches\ndirectory for 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": "Sat, 26 Dec 1998 15:10:13 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: [PATCHES] NetBSD/m68k port"
}
] |
[
{
"msg_contents": "\nhas anyone noticed that the backend still dies when you do a \"\\d\" from\npsql ? This is with 6.4.1 on bsdi 3.1 \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": "Sat, 26 Dec 1998 20:11:38 +0200 (EET)",
"msg_from": "Angelos Karageorgiou <[email protected]>",
"msg_from_op": true,
"msg_subject": "psql \\d does not work"
},
{
"msg_contents": "> \n> has anyone noticed that the backend still dies when you do a \"\\d\" from\n> psql ? This is with 6.4.1 on bsdi 3.1 \n> \n\nJust checked it on bsdi 4.0, and it works fine. We have also had no\nother problem reports like yours.\n\n\ttest=> create table test (x int);\n\tCREATE\n\ttest=> \\d\n\t\n\tDatabase = test\n\t +------------------+----------------------------------+----------+\n\t | Owner | Relation | Type |\n\t +------------------+----------------------------------+----------+\n\t | postgres | test | table |\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": "Sat, 26 Dec 1998 14:11:58 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] psql \\d does not work"
},
{
"msg_contents": "On Sat, 26 Dec 1998, Angelos Karageorgiou wrote:\n\n> has anyone noticed that the backend still dies when you do a \"\\d\" from\n> psql ? This is with 6.4.1 on bsdi 3.1 \n> \n\nHate to respond to myself, but Bruce noted that other people's \\d on the\nback end did not crash.\n\nI had compiled postgres with shlicc2 which is the shared lib version of gcc2 \non bsdi. That version was crashing the back end. I recompiled with the\ndefault gcc2 and it works magnificently \n\nAny clues ?\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": "Sun, 27 Dec 1998 12:38:24 +0200 (EET)",
"msg_from": "Angelos Karageorgiou <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] psql \\d does not work"
},
{
"msg_contents": "> On Sat, 26 Dec 1998, Angelos Karageorgiou wrote:\n> \n> > has anyone noticed that the backend still dies when you do a \"\\d\" from\n> > psql ? This is with 6.4.1 on bsdi 3.1 \n> > \n> \n> Hate to respond to myself, but Bruce noted that other people's \\d on the\n> back end did not crash.\n> \n> I had compiled postgres with shlicc2 which is the shared lib version of gcc2 \n> on bsdi. That version was crashing the back end. I recompiled with the\n> default gcc2 and it works magnificently \n> \n> Any clues ?\n\nStrange. I thought I used shlicc on 6.3 and bsdi 3.1, but I am not\nsure.\n\nCan you send a backtrace of the crash through gdb? It will be tricky\nbecause you have to get the query from psql.c that is \\d, start up gdb\non a backend, and run the query. The backtrace may show that if you\nsent it to me.\n\nDo all queries fail, or just \\d? Strange. I don't know what it is\nabout shlicc that would cause a failure.\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, 27 Dec 1998 13:10:51 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] psql \\d does not work"
},
{
"msg_contents": "On Sun, 27 Dec 1998, Bruce Momjian wrote:\n\n\n> Strange. I thought I used shlicc on 6.3 and bsdi 3.1, but I am not\n> sure.\n>\n\nWell you are right, after a lot of investigating I found out that it was \nneither shlicc2's fault nor postgres's what had happened is that I was using \n6.4.beta2 and have moved things about a bit. As a result the pg_shadow\ntable existed but the pg_user did NOT.\n\nSo when psql was doing a select from pg_user the back end would crash, not\ncrash really but disconnect.\n\nI redid the initdb and reloaded all my postgress databases and everything is\nfine.\n\nBTW the reason I was using a beta is that 6.3.2 and locales did not work\ntogether all to well. Now I am happy\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, 28 Dec 1998 10:27:55 +0200 (EET)",
"msg_from": "Angelos Karageorgiou <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] psql \\d does not work"
}
] |
[
{
"msg_contents": "\n Hello,\n\n a quick question (I couldn't contact postgresql.org, so I didn't check\nthe FAQ, but I doubt it would be there.)\n\n When I create a table with field \"nnn float8\" and total row size, say\n140 bytes, and then I insert a row with NULL in this field, will my record\ntake up those 8 bytes in physical page (say 140 bytes) or will it take up\nless (like 140-8=132), i.e. will it be shortened by those 8 bytes, since\nthe value is NULL and no information is stored in that field ? ( I *hate*\nrun-on sentences :)\n\n In general, what are the savings/overheads associated in PG with\nNULL/not NULL fields ? I.e. if will I save the space if I define the field\nas varchar(8) instead of float8 ?\n\n Thanx,\n\n Jan\n\n -- Gospel of Jesus is the saving power of God for all who believe --\n ## To some, nothing is impossible. ##\n http://Vicherek.Waterloo.on.ca/\n\n",
"msg_date": "Sat, 26 Dec 1998 22:36:53 -0500 (EST)",
"msg_from": "Jan Vicherek <[email protected]>",
"msg_from_op": true,
"msg_subject": "space usage of NULL fields"
},
{
"msg_contents": "> \n> Hello,\n> \n> a quick question (I couldn't contact postgresql.org, so I didn't check\n> the FAQ, but I doubt it would be there.)\n> \n> When I create a table with field \"nnn float8\" and total row size, say\n> 140 bytes, and then I insert a row with NULL in this field, will my record\n> take up those 8 bytes in physical page (say 140 bytes) or will it take up\n> less (like 140-8=132), i.e. will it be shortened by those 8 bytes, since\n> the value is NULL and no information is stored in that field ? ( I *hate*\n> run-on sentences :)\n\nShorter. Nulls are not stored in the row. There is a bitmask on every\nrow that shows the number of NULL fields.\n\n> \n> In general, what are the savings/overheads associated in PG with\n> NULL/not NULL fields ? I.e. if will I save the space if I define the field\n> as varchar(8) instead of float8 ?\n\nno. float8 is shorter because varchar has 4-byte overhead. The faq is\nin docs/FAQ 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": "Sun, 27 Dec 1998 11:00:54 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] space usage of NULL fields"
}
] |
[
{
"msg_contents": "\n\nI have been using postgres for about 4 months now, and I have to admit\nit is an awesome product , congrats to you all people but I do have a \nlingering question\n\n<pest mode>\nin pg_hba.conf when my netmask is 255.255.255.0 like below the authentication\nmechanism works fine\n\nhost databank x.y.z.12 255.255.255.0 ident sameuser \n\n\nbut if I change my netmask to 255.255.255.255 like below it does not work\nany more. Am I doing something wrong or did I not understand thing correctly ?\nI think it would be better security-wise to have 255.255.255.255 netmasks for\nindividual hosts. \nhost databank x.y.z.12 255.255.255.255 ident sameuser \n</pest mode>\n\nBTW I am working in Greece and postgres with the Greek locale on BSDI3.1 works\nfabulously ! Again congrats to the people who plugged in the locale code.\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": "Sun, 27 Dec 1998 12:49:02 +0200 (EET)",
"msg_from": "Angelos Karageorgiou <[email protected]>",
"msg_from_op": true,
"msg_subject": "pg_hba.conf"
},
{
"msg_contents": "Angelos Karageorgiou <[email protected]> writes:\n> in pg_hba.conf when my netmask is 255.255.255.0 like below the authentication\n> mechanism works fine\n> host databank x.y.z.12 255.255.255.0 ident sameuser \n> but if I change my netmask to 255.255.255.255 like below it does not work\n> any more. Am I doing something wrong or did I not understand thing correctly?\n> host databank x.y.z.12 255.255.255.255 ident sameuser \n\nI use 255.255.255.255 netmask and it works as expected for me...\ndo you still see this problem now that you have fixed your libc and\ndatabase setup problems?\n\nAlso, what do you mean by \"it does not work\"? Does it fail to let in\nnew connections from the intended host, or improperly let in connections\nfrom other hosts?\n\n\t\t\tregards, tom lane\n\nPS: I assume you've double-checked that the intended host really is\nx.y.z.12, and not x.y.z.something-else...\n",
"msg_date": "Mon, 28 Dec 1998 10:55:42 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] pg_hba.conf "
}
] |
[
{
"msg_contents": "Hi all\n\nThe following is a snippet from the expected trigger.out file, the only\ndifferance between this and what I got is that the following 'NOTICE:'\nlines are missing in mine.\nIs this a problem?\nThis was the only failure.\n\nQUERY: delete from pkeys where pkey1 = 30 and pkey2 = '3';\nNOTICE: check_pkeys_fkey_cascade: 1 tuple(s) of fkeys are deleted\nERROR: check_fkeys2_fkey_restrict: tuple referenced in fkeys\nQUERY: delete from pkeys where pkey1 = 40 and pkey2 = '4';\nNOTICE: check_pkeys_fkey_cascade: 1 tuple(s) of fkeys are deleted\nNOTICE: check_pkeys_fkey_cascade: 1 tuple(s) of fkeys2 are deleted\nQUERY: update pkeys set pkey1 = 7, pkey2 = '70' where pkey1 = 50 and pkey2\n= '5'NOTICE: check_pkeys_fkey_cascade: 1 tuple(s) of fkeys are deleted\nERROR: check_fkeys2_fkey_restrict: tuple referenced in fkeys\nQUERY: update pkeys set pkey1 = 7, pkey2 = '70' where pkey1 = 10 and pkey2\n= '1'NOTICE: check_pkeys_fkey_cascade: 1 tuple(s) of fkeys are deleted\nNOTICE: check_pkeys_fkey_cascade: 1 tuple(s) of fkeys2 are deleted\n\nThanks, and have a great day\nTerry Mackintosh <[email protected]> http://www.terrym.com\nsysadmin/owner Please! No MIME encoded or HTML mail, unless needed.\n\nProudly powered by R H Linux 4.2, Apache 1.3, PHP 3, PostgreSQL 6.4.1\n---------------------------------------------------------------------\nSuccess Is A Choice ... book by Rick Patino, get it, read it!\n\n",
"msg_date": "Sun, 27 Dec 1998 12:35:02 -0500 (EST)",
"msg_from": "Terry Mackintosh <[email protected]>",
"msg_from_op": true,
"msg_subject": "regression test: triggers fial:"
}
] |
[
{
"msg_contents": "\n Hello,\n\n I'm looking for a int8-like type or equivalent substitute.\n\n float8 is the first one that comes to mind, but a question arises :\n isn't there a chance that if I store float8 values be integers, it may\non some occasions return to me a decimal ? I.e. I would store something\nlike\n 2814749767106569\nand on retrieval it would return something like\n 2814749767106568.999999999998 instead of 2814749767106569 ?\n\n If there is such danger, what other suitable 8 byte data type (other\nthan char(8)) would there be in PG to give me the ablility to store\nintegers > 2^32 ?\n\n Or is it quite fast, easy, simple and performance-effective to add my\nown 'int8' data type ?\n\n Thanx,\n\n Jan\n\n\n -- Gospel of Jesus is the saving power of God for all who believe --\n ## To some, nothing is impossible. ##\n http://Vicherek.Waterloo.on.ca/\n\n",
"msg_date": "Sun, 27 Dec 1998 13:29:23 -0500 (EST)",
"msg_from": "Jan Vicherek <[email protected]>",
"msg_from_op": true,
"msg_subject": "int8 type substitute ?"
},
{
"msg_contents": "\n/me goes red in face. 10 seconds after I sent this off I found\ncontrib/int8. That would answer one of my questions -- I.e. it would be\nreasonably simple to add my own data type. If somebody feels like\nanswering the other questions too, please go ahead.\n\n Jan\n\nOn Sun, 27 Dec 1998, Jan Vicherek wrote:\n\n> \n> Hello,\n> \n> I'm looking for a int8-like type or equivalent substitute.\n> \n> float8 is the first one that comes to mind, but a question arises :\n> isn't there a chance that if I store float8 values be integers, it may\n> on some occasions return to me a decimal ? I.e. I would store something\n> like\n> 2814749767106569\n> and on retrieval it would return something like\n> 2814749767106568.999999999998 instead of 2814749767106569 ?\n> \n> If there is such danger, what other suitable 8 byte data type (other\n> than char(8)) would there be in PG to give me the ablility to store\n> integers > 2^32 ?\n> \n> Or is it quite fast, easy, simple and performance-effective to add my\n> own 'int8' data type ?\n> \n> Thanx,\n> \n> Jan\n> \n> \n> -- Gospel of Jesus is the saving power of God for all who believe --\n> ## To some, nothing is impossible. ##\n> http://Vicherek.Waterloo.on.ca/\n> \n> \n\n -- Gospel of Jesus is the saving power of God for all who believe --\n ## To some, nothing is impossible. ##\n http://Vicherek.Waterloo.on.ca/\n\n",
"msg_date": "Sun, 27 Dec 1998 13:35:35 -0500 (EST)",
"msg_from": "Jan Vicherek <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: int8 type substitute ?"
},
{
"msg_contents": "> \n> /me goes red in face. 10 seconds after I sent this off I found\n> contrib/int8. That would answer one of my questions -- I.e. it would be\n> reasonably simple to add my own data type. If somebody feels like\n> answering the other questions too, please go ahead.\n> \n> Jan\n\n6.4.* already has int8 support included, if your compiler supports it. \nWhen you run configure, it checks to see if 'long long' is int8, and if\nso, enables int8 data types.\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, 27 Dec 1998 14:17:15 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: int8 type substitute ?"
}
] |
[
{
"msg_contents": "Marc,\n\nI just noticed that my mirror of www.postgresql.org\nos totally broken ! The problem is in your recent changes -\nmany documents used /style1.css which of course doesn't exists in\nmy document root. I notice other mirrors also broken on that reason.\nAlso main page doesn't have a referene to my mirror in Russia\n(both ftp and web sites, please note we're in Europe still :-)\n\nDoes it means you changed mirror policy ?\n\nI had to reconfigure my server to get mirror works properly - \nnow it has URL http://www.sai.msu.su:8000/\n\n\tRegards,\n\t\n\t\tOleg\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\n",
"msg_date": "Mon, 28 Dec 1998 00:20:57 +0300 (MSK)",
"msg_from": "Oleg Bartunov <[email protected]>",
"msg_from_op": true,
"msg_subject": "Web site mirroring trouble"
}
] |
[
{
"msg_contents": "\t>> Shouldn't we change default transaction mode now?\n\t>> And use option/SET TRANSACTION MODE to switch to\n\t>> un-chained mode if one like it?\n\t>\n\t>No comments on this?\n\t>I would like to make BEGIN implicit...\n\t>Objections?\n\n\tYes, I object. I think the default behavior should be the one that \n\tis sufficient in most cases. \n\n\tAs to the isolation level: \n\tInformix and DB/2 default to read committed mode. In most cases this\nis \n\tsufficient.\n\tTherefore most clients don't need the extra set transaction\nisolation statement.\n\n\tAs to transaction blocks: \n\tActually most clients never use commit work, and remember\n\tthat postgresql will never free it's memory before commit. This will\nhog up \n\tmemory for otherwise lightweight clients.\n\n\tI also think that the begin work; commit work; statement block is\n\tsomewhat part of the postgresql philosophy. We do not stand alone\n\twith this approach (Informix).\n\tActually the other DB's start the transaction with the first\nmodifying \n\tstatement, in my opinion a rather weird approach. \n\tI have seen programs that do a dummy update, just to simulate a \n\tbegin work, what a waste.\n\tBut this alone will not work, because a select statement, that needs\n\n\ta sort would have started the transaction earlier (Since it created\nan implicit\n\ttemp table).\n\tI could go on, but you see I don't really like this ...... \n\n\tIf you want to change the behavior, I think we will need a\nchangeable default \n\tat database level.\n\n\tAndreas\n\n",
"msg_date": "Mon, 28 Dec 1998 10:30:01 +0100",
"msg_from": "Zeugswetter Andreas IZ5 <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] MVCC works in serialized mode!"
},
{
"msg_contents": "\n> -----Original Message-----\n> From: [email protected]\n> [mailto:[email protected]]On Behalf Of Zeugswetter\n> Andreas IZ5\n> Sent: Monday, December 28, 1998 6:30 PM\n> To: [email protected]\n> Cc: '[email protected]'\n> Subject: Re: [HACKERS] MVCC works in serialized mode!\n> \n> \n> \t>> Shouldn't we change default transaction mode now?\n> \t>> And use option/SET TRANSACTION MODE to switch to\n> \t>> un-chained mode if one like it?\n> \t>\n> \t>No comments on this?\n> \t>I would like to make BEGIN implicit...\n> \t>Objections?\n> \n> \tYes, I object. I think the default behavior should be the one that \n> \tis sufficient in most cases. \n>\n\nI have a question about new default(isolation level / transaction mode).\n\nHow do we upgrade existent programs,if default is different \nfrom current ?\nI think the isolation level of current PostgreSQL is SERIALIZABLE \nand the transaction mode is un-chained..\n\nAs to the isolation level,even SERIALIZABLE isolaton level can't guarantee \nthe integrity of current level.\nSo we must change existent programs anyway ?\n\nThanks.\n\nHiroshi Inoue\[email protected]\n \n> \tAs to the isolation level: \n> \tInformix and DB/2 default to read committed mode. In most cases this\n> is \n> \tsufficient.\n> \tTherefore most clients don't need the extra set transaction\n> isolation statement.\n> \n> \tAs to transaction blocks: \n> \tActually most clients never use commit work, and remember\n> \tthat postgresql will never free it's memory before commit. This will\n> hog up \n> \tmemory for otherwise lightweight clients.\n> \n> \tI also think that the begin work; commit work; statement block is\n> \tsomewhat part of the postgresql philosophy. We do not stand alone\n> \twith this approach (Informix).\n> \tActually the other DB's start the transaction with the first\n> modifying \n> \tstatement, in my opinion a rather weird approach. \n> \tI have seen programs that do a dummy update, just to simulate a \n> \tbegin work, what a waste.\n> \tBut this alone will not work, because a select statement, that needs\n> \n> \ta sort would have started the transaction earlier (Since it created\n> an implicit\n> \ttemp table).\n> \tI could go on, but you see I don't really like this ...... \n> \n> \tIf you want to change the behavior, I think we will need a\n> changeable default \n> \tat database level.\n> \n> \tAndreas\n> \n> \n> \n> \n> \n",
"msg_date": "Wed, 6 Jan 1999 09:30:01 +0900",
"msg_from": "\"Hiroshi Inoue\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "RE: [HACKERS] MVCC works in serialized mode!"
},
{
"msg_contents": "Hiroshi Inoue wrote:\n> \n> I have a question about new default(isolation level / transaction mode).\n> \n> How do we upgrade existent programs,if default is different\n> from current ?\n> I think the isolation level of current PostgreSQL is SERIALIZABLE\n> and the transaction mode is un-chained..\n> \n> As to the isolation level,even SERIALIZABLE isolaton level can't guarantee\n> the integrity of current level.\n> So we must change existent programs anyway ?\n\nYou're right!\nThe word from Oracle: (-:))\n\n\"Because Oracle does not use read locks, even in serializable \ntransactions, data read by one transaction can be overwritten \nby another. Transactions that perform database consistency\nchecks at the application level should not assume that the data \nthey read will not change during the execution of the transaction \n(even though such changes are not visible to the transaction). \nDatabase inconsistencies can result unless such application-level \nconsistency checks are coded with this in mind, even when using \nserializable transactions.\n...\nAlthough Oracle serializable mode is compatible with SQL92 and \noffers many benefits as compared with read-locking implementations, \nit does not provide semantics identical to such systems. \nApplication designers must take into account the fact that reads \nin Oracle do not block writes as they do in other systems. \nTransactions that check for database consistency at the application \nlevel may require coding techniques such as the use of\nSELECT FOR UPDATE. This issue should be considered when \n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\napplications using serializable mode are ported to Oracle \n^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\nfrom other environments.\"\n^^^^^^^^^^^^^^^^^^^^^^^\n\n1. All our applications run in serializable mode now.\n2. \"Environment\" is changed from locking to\n multi-versioning.\n\nThis has to be explained in release notes.\n\nShould we implement ability to run backend in mode compatible\nwith old versions (it seems easy to do - just use AccessExclusive \nLock for UPDATE/INSERT/DELETE in Executor and don't release\nAccessShare Lock in heap_endscan) ?\n\nVadim\n",
"msg_date": "Sun, 10 Jan 1999 23:26:27 +0700",
"msg_from": "Vadim Mikheev <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] MVCC works in serialized mode!"
},
{
"msg_contents": "Zeugswetter Andreas IZ5 wrote:\n> \n> >> Shouldn't we change default transaction mode now?\n> >> And use option/SET TRANSACTION MODE to switch to\n> >> un-chained mode if one like it?\n> >\n> >No comments on this?\n> >I would like to make BEGIN implicit...\n> >Objections?\n> \n> Yes, I object. I think the default behavior should be the one that\n> is sufficient in most cases.\n> \n> As to the isolation level:\n> Informix and DB/2 default to read committed mode. In most cases this\n> is\n> sufficient.\n> Therefore most clients don't need the extra set transaction\n> isolation statement.\n\nAnd this is the same as in Oracle & SyBase.\nI don't object - currently READ COMMITTED is not\nsupported by DELETE/UPDATE code and so default is\nSERIALIZABLE, to be changed latter.\n\n> \n> As to transaction blocks:\n> Actually most clients never use commit work, and remember\n\nIn Oracle, COMMIT is implicit if DB connection is closed\nin \"normal\" way - by using PQfinish in our case.\n\n> that postgresql will never free it's memory before commit. This will\n> hog up\n> memory for otherwise lightweight clients.\n\nThis is the bug, but I believe that it's fixed for\nmost cases ~1year ago by this\n\n#ifdef TBL_FREE_CMD_MEMORY\n EndPortalAllocMode();\n StartPortalAllocMode(DefaultAllocMode, 0);\n#endif\n\nin xact.c\n\n> I also think that the begin work; commit work; statement block is\n> somewhat part of the postgresql philosophy. We do not stand alone\n> with this approach (Informix).\n\nFirst, BEGIN/END is used in standard for grouping queries\n(in PL etc), not for transaction control, and so currently\nusing of BEGIN/END is ambiguous.\nSecond, Informix is using locking => chained transaction\nmode is way to lock-escalation.\n\n> Actually the other DB's start the transaction with the first\n> modifying\n> statement, in my opinion a rather weird approach.\n> I have seen programs that do a dummy update, just to simulate a\n> begin work, what a waste.\n\nI believe that this is not how Oracle works.\nI assumed that any DML statement will start transaction.\n\n> But this alone will not work, because a select statement, that needs\n> \n> a sort would have started the transaction earlier (Since it created\n> an implicit\n> temp table).\n ^^^^^^^^^^\nThis is changed - backend don't create temp table now. \n\n> I could go on, but you see I don't really like this ......\n> \n> If you want to change the behavior, I think we will need a\n> changeable default\n> at database level.\n\nOk, for compatibility reasons, I agreed that default\nmust be un-chained mode, with ability to switch in\ncompile/run time.\n\nVadim\n",
"msg_date": "Mon, 11 Jan 1999 00:05:08 +0700",
"msg_from": "Vadim Mikheev <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] MVCC works in serialized mode!"
},
{
"msg_contents": "Hello all,\n\n> -----Original Message-----\n> From: [email protected] [mailto:[email protected]]On Behalf Of Vadim\n> Mikheev\n> Sent: Monday, January 11, 1999 1:26 AM\n> To: Hiroshi Inoue\n> Cc: [email protected]\n> Subject: Re: [HACKERS] MVCC works in serialized mode!\n>\n>\n> Hiroshi Inoue wrote:\n> >\n> > I have a question about new default(isolation level / transaction mode).\n> >\n> > How do we upgrade existent programs,if default is different\n> > from current ?\n> > I think the isolation level of current PostgreSQL is SERIALIZABLE\n> > and the transaction mode is un-chained..\n> >\n[snip]\n>\n> 1. All our applications run in serializable mode now.\n> 2. \"Environment\" is changed from locking to\n> multi-versioning.\n>\n> This has to be explained in release notes.\n>\n> Should we implement ability to run backend in mode compatible\n> with old versions (it seems easy to do - just use AccessExclusive\n> Lock for UPDATE/INSERT/DELETE in Executor and don't release\n> AccessShare Lock in heap_endscan) ?\n>\n\nYes,if it's easy as you say.\nBut I don't know whether other people mind it or not.\nIf no one mind it,SERIALIZABLE is permissible for me.\n\nBTW before user's code of our own,there are no problems with\ninterface library such as ODBC/JDBC etc ?\nThose work well with chained mode or read committed isolation\nlevel ?\n\nThanks.\n\nHiroshi Inoue\[email protected]\n\n",
"msg_date": "Tue, 12 Jan 1999 09:40:22 +0900",
"msg_from": "\"Hiroshi Inoue\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "RE: [HACKERS] MVCC works in serialized mode!"
}
] |
[
{
"msg_contents": "Hi,\n\nI installed postgres from my User account on Solaris 2.6 and created some\ndatabases. If another user with a different Solaris account wants to access\nthe database that I created, what is the procedure? Right now, when we try \nto do that (psql <dbname> ), there is an error saying user-id not in pg_shadow\n\nCould anyone please help us on that ?\n\nThanks\n--shiby\n\n\n",
"msg_date": "Mon, 28 Dec 1998 13:25:43 -0500",
"msg_from": "Shiby Thomas <[email protected]>",
"msg_from_op": true,
"msg_subject": "Database access"
}
] |
[
{
"msg_contents": "My feeling (not that anyone asked) is that if one bit changes after an\nofficial release, then it's necessary to change the release name\n*somehow*. I'm at the ready to update the web site if Marc makes 6.4.2\nor 6.4.2.1 or 6.4.1-bis or whatever...\n\nThe Hermit Hacker <[email protected]> writes:\n\n> If you guys are *sure* you want to do this, let's do it...tag what is\n> required, let me know when done, I'll package up a new one and put it out,\n> removing v6.4.1 and then Hal can make appropriate changes to the rest of\n> the site :(\n\n",
"msg_date": "28 Dec 1998 14:18:09 -0600",
"msg_from": "Hal Snyder <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [CORE] Re: [HACKERS] postgresql-6.4.1.tar.gz"
}
] |
[
{
"msg_contents": "I've just applied a large number of changes to the sgml docs contributed\nby Vince Vielhaber. These mostly add \"tag completion\" to the markup as\nrequired by the newest versions of docbook and the style sheets.\n\nThe content changes are minimal; I've added mention of the \"NULL\" column\nconstraint to be added to gram.y (not yet committed though; waiting to\nget some more outer join parser stuff ready).\n\nThe changes required a good bit of time and effort to do; thanks Vince!\n\n - Tom\n",
"msg_date": "Tue, 29 Dec 1998 03:10:13 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "sgml tag completion cleanup"
}
] |
[
{
"msg_contents": "I'm trying to set up cvsup on a new machine to snarf the postgres cvs\nrepository. I've had my machine at home set up for so long, I can't\nremember how the $CVSROOT/CVSROOT part of my local repository directory\ntree came into being. Should it be imported from postgresql.org as is\nthe pgsql part of the repository? Or should I set it up locally? The\npgsql part of the tree shows up just fine; I'm just not sure how the\nother piece gets there. btw, I'm using an identical cvsup client\nconfiguration file on both machines.\n\n - Tom\n",
"msg_date": "Tue, 29 Dec 1998 05:59:01 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Forgot how to set up cvsup"
},
{
"msg_contents": "> I'm trying to set up cvsup on a new machine to snarf the postgres cvs\n> repository. I've had my machine at home set up for so long, I can't\n> remember how the $CVSROOT/CVSROOT part of my local repository directory\n> tree came into being. Should it be imported from postgresql.org as is\n> the pgsql part of the repository? Or should I set it up locally?\n\nI'd expect it to come in from the master site, as part of the update.\nI'm not sure how the collections are set up in postgresql. In the\nFreeBSD project, we have a collection \"src-base\" that contains\n$CVSROOT/CVSROOT and little else. Then there are separate collections\nfor various large parts of the source tree. Finally, there is a sort\nof mega-collection which is equivalent to all of that put together in\na single collection.\n\nJohn\n",
"msg_date": "Mon, 28 Dec 1998 22:06:57 -0800 (PST)",
"msg_from": "John Polstra <[email protected]>",
"msg_from_op": false,
"msg_subject": "RE: Forgot how to set up cvsup"
}
] |
[
{
"msg_contents": "Hello David, hello Thomas,\n\na couple of weeks ago I had a similar problem, I solved it by a little\nshell-script, (sorry, no pg-sql) which registers the missing operators.\nAs I thought it4s going to be fixed in the next version, I didn4t post\nit. But untill then, perhaps it helps ....\nFirst you need the int8cmp-Operator (I something like a\nsignature-function of the difference). Than you should register this\noperator.\n\n\nKind regards\n\nMichael Contzen\n\n-------------- int8cmp.c ---------------\n\n\n/*-------------------------------------------------------------------------\n\n *\n * int8cmp.c--\n * Internal 64-bit integer operations\n *\n *-------------------------------------------------------------------------\n\n */\n#include <stdio.h> /* for sprintf proto,\netc. */\n#include <stdlib.h> /* for strtod, etc. */\n#include <string.h>\n#include <ctype.h>\n#include <time.h>\n#include <math.h>\n#include <float.h>\n#include <limits.h>\n\n#include \"postgres.h\"\n#include \"utils/palloc.h\"\n\n#define MAXINT8LEN 25\n\n#if defined(__alpha) || defined(__GNUC__)\n#define HAVE_64BIT_INTS 1\n#endif\n\n#ifndef HAVE_64BIT_INTS\ntypedef char[8] int64;\n\n#elif defined(__alpha)\ntypedef long int int64;\n\n#define INT64_FORMAT \"%ld\"\n\n#elif defined(__GNUC__)\ntypedef long long int int64;\n\n#define INT64_FORMAT \"%Ld\"\n\n#else\ntypedef long int int64;\n\n#define INT64_FORMAT \"%ld\"\n#endif\n\nint32 int8cmp(int64 * val1, int64 * val2);\n\nint32\nint8cmp(int64 * val1, int64 * val2)\n{\n if (*val1 > *val2) return 1;\n if (*val1 < *val2) return -1;\n return 0;\n} /* int8cmp() */\n\n\n------------------ End of int8cmp.c ------------------\n\n\n\n\n\n------------------ initlongbtree.sh -----------------\n\n\ninitlongbtree.sh:\n\nSQL=\"psql -t MYDATABASE \"\n\nOPS=`$SQL -c \"SELECT am.amname, opc.opcname, opr.oprname\n FROM pg_am am, pg_amop amop,\n pg_opclass opc, pg_operator opr\n WHERE amop.amopid = am.oid AND\n amop.amopclaid = opc.oid AND\n amop.amopopr = opr.oid AND\n opc.opcname='int8_ops'\n ORDER BY am.amname, opc.opcname\"`\n\nif [ \"$OPS\" != \"\" ]; then\n echo \"int8-btree ops already registerd.\"\n exit 0\nfi\n\nTYPE=`$SQL -c \"select oid from pg_type where typname='int8'\"`\necho \"TYP=$TYPE\"\nif [ \"$TYPE\" = \"\" ]; then\n echo \"No int8-type found!\"\n exit -1\nfi\n\n$SQL </usr/local/pgsql/lib/int8cmp.sql\n\n$SQL -c \"delete from pg_opclass where opcname='int8_ops'\"\nOPCLOID=`$SQL -c \"insert into pg_opclass values('int8_ops', $TYPE)\" |\ncut -f2 -d\" \"`\n\nOP1=`$SQL -c \"select oid from pg_operator where oprname='<' and\noprleft=$TYPE and oprright=$TYPE\"`\nOP2=`$SQL -c \"select oid from pg_operator where oprname='<=' and\noprleft=$TYPE and oprright=$TYPE\"`\nOP3=`$SQL -c \"select oid from pg_operator where oprname='=' and\noprleft=$TYPE and oprright=$TYPE\"`\nOP4=`$SQL -c \"select oid from pg_operator where oprname='>=' and\noprleft=$TYPE and oprright=$TYPE\"`\nOP5=`$SQL -c \"select oid from pg_operator where oprname='>' and\noprleft=$TYPE and oprright=$TYPE\"`\nOP6=`$SQL -c \"select oid from pg_proc where proname='int8cmp' \"`\n\n\nAMOID=`$SQL -c \"select oid from pg_am where amname='btree'\"`\n\necho \"OPCLOID=$OPCLOID AMOID=$AMOID OP1..5: $OP1 $OP2 $OP3 $OP4 $OP5\ncmp=$OP6\"\n\nif [ \"$OP6\" = \"\" ]; then\n echo \"No int8cmp found!\"\n exit\nfi\n\n$SQL -c \"insert into pg_amop values ($AMOID, $OPCLOID, $OP1, 1,\n'btreesel', 'btreenpage')\"\n$SQL -c \"insert into pg_amop values ($AMOID, $OPCLOID, $OP2, 2,\n'btreesel', 'btreenpage')\"\n$SQL -c \"insert into pg_amop values ($AMOID, $OPCLOID, $OP3, 3,\n'btreesel', 'btreenpage')\"\n$SQL -c \"insert into pg_amop values ($AMOID, $OPCLOID, $OP4, 4,\n'btreesel', 'btreenpage')\"\n$SQL -c \"insert into pg_amop values ($AMOID, $OPCLOID, $OP5, 5,\n'btreesel', 'btreenpage')\"\n$SQL -c \"insert into pg_amproc(amid, amopclaid, amproc, amprocnum)\n values ($AMOID, $OPCLOID, $OP6, 1)\"\n\n------------------- end initlongbtree.sh ---------------\n\n\n",
"msg_date": "Tue, 29 Dec 1998 09:43:33 +0000",
"msg_from": "Michael Contzen <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] bug(?) if int8 as primary key"
}
] |
[
{
"msg_contents": ">From [194.58.47.100]!mcit.aris.ru!hub.org!owner-pgsql-hackers Mon Dec 28 10:36:11 GMT 1998 remote from doom\nReceived: from mcit.aris.ru ([194.58.47.100]) by doom.samara.su with SMTP id AA00380\n (5.65/IDA-simtel for <[email protected]>); Mon, 28 Dec 1998 10:36:11 +0400\nReceived: from hub.org by mcit.aris.ru with ESMTP id VAA16636;\n (8.8.5-MVC-230497/vak/1.9) Sun, 27 Dec 1998 21:39:34 +0300 (MSK)\nReceived: from localhost (majordom@localhost)\n\tby hub.org (8.9.1/8.9.1) with SMTP id NAA25039;\n\tSun, 27 Dec 1998 13:36:43 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: by hub.org (TLB v0.10a (1.23 tibbs 1997/01/09 00:29:32)); Sun, 27 Dec 1998 13:36:35 +0000 (EST)\nReceived: (from majordom@localhost)\n\tby hub.org (8.9.1/8.9.1) id NAA24939\n\tfor pgsql-hackers-outgoing; Sun, 27 Dec 1998 13:36:30 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: from enrico.ied.com (IDENT:[email protected] [129.97.130.69])\n\tby hub.org (8.9.1/8.9.1) with ESMTP id NAA24923\n\tfor <[email protected]>; Sun, 27 Dec 1998 13:35:58 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: from ann.ied.com ([email protected] [192.168.200.20])\n\tby enrico.ied.com (8.8.7/8.8.7) with ESMTP id NAA28127\n\tfor <[email protected]>; Sun, 27 Dec 1998 13:35:02 -0500\nDate: Sun, 27 Dec 1998 13:35:35 -0500 (EST)\nFrom: Jan Vicherek <[email protected]>\nTo: [email protected]\nSubject: [HACKERS] Re: int8 type substitute ?\nIn-Reply-To: <[email protected]>\nMessage-Id: <[email protected]>\nMime-Version: 1.0\nContent-Type: TEXT/PLAIN; charset=US-ASCII\nSender: [email protected]\nPrecedence: bulk\n\n\n/me goes red in face. 10 seconds after I sent this off I found\ncontrib/int8. That would answer one of my questions -- I.e. it would be\nreasonably simple to add my own data type. If somebody feels like\nanswering the other questions too, please go ahead.\n\n Jan\n\nOn Sun, 27 Dec 1998, Jan Vicherek wrote:\n\n> \n> Hello,\n> \n> I'm looking for a int8-like type or equivalent substitute.\n> \n> float8 is the first one that comes to mind, but a question arises :\n> isn't there a chance that if I store float8 values be integers, it may\n> on some occasions return to me a decimal ? I.e. I would store something\n> like\n> 2814749767106569\n> and on retrieval it would return something like\n> 2814749767106568.999999999998 instead of 2814749767106569 ?\n> \n> If there is such danger, what other suitable 8 byte data type (other\n> than char(8)) would there be in PG to give me the ablility to store\n> integers > 2^32 ?\n> \n> Or is it quite fast, easy, simple and performance-effective to add my\n> own 'int8' data type ?\n> \n> Thanx,\n> \n> Jan\n> \n> \n> -- Gospel of Jesus is the saving power of God for all who believe --\n> ## To some, nothing is impossible. ##\n> http://Vicherek.Waterloo.on.ca/\n> \n> \n\n -- Gospel of Jesus is the saving power of God for all who believe --\n ## To some, nothing is impossible. ##\n http://Vicherek.Waterloo.on.ca/\n\n\n",
"msg_date": "Tue, 29 Dec 1998 14:25:17 +0400 (SAMT)",
"msg_from": "[email protected] (Samara CIT)",
"msg_from_op": true,
"msg_subject": "None"
}
] |
[
{
"msg_contents": ">From [194.58.47.100]!mcit.aris.ru!hub.org!owner-pgsql-hackers Mon Dec 28 10:36:17 GMT 1998 remote from doom\nReceived: from mcit.aris.ru ([194.58.47.100]) by doom.samara.su with SMTP id AA00382\n (5.65/IDA-simtel for <[email protected]>); Mon, 28 Dec 1998 10:36:17 +0400\nReceived: from hub.org by mcit.aris.ru with ESMTP id VAA16622;\n (8.8.5-MVC-230497/vak/1.9) Sun, 27 Dec 1998 21:34:50 +0300 (MSK)\nReceived: from localhost (majordom@localhost)\n\tby hub.org (8.9.1/8.9.1) with SMTP id NAA24515;\n\tSun, 27 Dec 1998 13:32:02 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: by hub.org (TLB v0.10a (1.23 tibbs 1997/01/09 00:29:32)); Sun, 27 Dec 1998 13:30:56 +0000 (EST)\nReceived: (from majordom@localhost)\n\tby hub.org (8.9.1/8.9.1) id NAA24155\n\tfor pgsql-hackers-outgoing; Sun, 27 Dec 1998 13:30:43 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: from enrico.ied.com (IDENT:[email protected] [129.97.130.69])\n\tby hub.org (8.9.1/8.9.1) with ESMTP id NAA23958\n\tfor <[email protected]>; Sun, 27 Dec 1998 13:29:46 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: from ann.ied.com ([email protected] [192.168.200.20])\n\tby enrico.ied.com (8.8.7/8.8.7) with ESMTP id NAA28100\n\tfor <[email protected]>; Sun, 27 Dec 1998 13:28:50 -0500\nDate: Sun, 27 Dec 1998 13:29:23 -0500 (EST)\nFrom: Jan Vicherek <[email protected]>\nTo: [email protected]\nSubject: [HACKERS] int8 type substitute ?\nMessage-Id: <[email protected]>\nMime-Version: 1.0\nContent-Type: TEXT/PLAIN; charset=US-ASCII\nSender: [email protected]\nPrecedence: bulk\n\n\n Hello,\n\n I'm looking for a int8-like type or equivalent substitute.\n\n float8 is the first one that comes to mind, but a question arises :\n isn't there a chance that if I store float8 values be integers, it may\non some occasions return to me a decimal ? I.e. I would store something\nlike\n 2814749767106569\nand on retrieval it would return something like\n 2814749767106568.999999999998 instead of 2814749767106569 ?\n\n If there is such danger, what other suitable 8 byte data type (other\nthan char(8)) would there be in PG to give me the ablility to store\nintegers > 2^32 ?\n\n Or is it quite fast, easy, simple and performance-effective to add my\nown 'int8' data type ?\n\n Thanx,\n\n Jan\n\n\n -- Gospel of Jesus is the saving power of God for all who believe --\n ## To some, nothing is impossible. ##\n http://Vicherek.Waterloo.on.ca/\n\n\n",
"msg_date": "Tue, 29 Dec 1998 14:25:18 +0400 (SAMT)",
"msg_from": "[email protected] (Samara CIT)",
"msg_from_op": true,
"msg_subject": "None"
}
] |
[
{
"msg_contents": ">From [194.58.47.100]!mcit.aris.ru!hub.org!owner-pgsql-hackers Mon Dec 28 10:36:45 GMT 1998 remote from doom\nReceived: from mcit.aris.ru ([194.58.47.100]) by doom.samara.su with SMTP id AA00390\n (5.65/IDA-simtel for <[email protected]>); Mon, 28 Dec 1998 10:36:45 +0400\nReceived: from hub.org by mcit.aris.ru with ESMTP id OAA15851;\n (8.8.5-MVC-230497/vak/1.9) Sun, 27 Dec 1998 14:08:17 +0300 (MSK)\nReceived: from localhost (majordom@localhost)\n\tby hub.org (8.9.1/8.9.1) with SMTP id GAA24534;\n\tSun, 27 Dec 1998 06:05:30 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: by hub.org (TLB v0.10a (1.23 tibbs 1997/01/09 00:29:32)); Sun, 27 Dec 1998 06:05:10 +0000 (EST)\nReceived: (from majordom@localhost)\n\tby hub.org (8.9.1/8.9.1) id GAA24172\n\tfor pgsql-hackers-outgoing; Sun, 27 Dec 1998 06:04:54 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: from awesome.incredible.com (awesome.incredible.com [195.167.39.10])\n\tby hub.org (8.9.1/8.9.1) with ESMTP id GAA24123\n\tfor <[email protected]>; Sun, 27 Dec 1998 06:03:26 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: from localhost (angelos@localhost)\n\tby awesome.incredible.com (8.8.8/8.8.8) with ESMTP id MAA18669\n\tfor <[email protected]>; Sun, 27 Dec 1998 12:49:03 +0200 (EET)\nDate: Sun, 27 Dec 1998 12:49:02 +0200 (EET)\nFrom: Angelos Karageorgiou <[email protected]>\nTo: [email protected]\nSubject: [HACKERS] pg_hba.conf\nMessage-Id: <[email protected]>\nMime-Version: 1.0\nContent-Type: TEXT/PLAIN; charset=US-ASCII\nSender: [email protected]\nPrecedence: bulk\n\n\n\nI have been using postgres for about 4 months now, and I have to admit\nit is an awesome product , congrats to you all people but I do have a \nlingering question\n\n<pest mode>\nin pg_hba.conf when my netmask is 255.255.255.0 like below the authentication\nmechanism works fine\n\nhost databank x.y.z.12 255.255.255.0 ident sameuser \n\n\nbut if I change my netmask to 255.255.255.255 like below it does not work\nany more. Am I doing something wrong or did I not understand thing correctly ?\nI think it would be better security-wise to have 255.255.255.255 netmasks for\nindividual hosts. \nhost databank x.y.z.12 255.255.255.255 ident sameuser \n</pest mode>\n\nBTW I am working in Greece and postgres with the Greek locale on BSDI3.1 works\nfabulously ! Again congrats to the people who plugged in the locale code.\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\n",
"msg_date": "Tue, 29 Dec 1998 14:25:21 +0400 (SAMT)",
"msg_from": "[email protected] (Samara CIT)",
"msg_from_op": true,
"msg_subject": "None"
}
] |
[
{
"msg_contents": ">From [194.58.47.100]!mcit.aris.ru!hub.org!owner-pgsql-hackers Mon Dec 28 10:36:49 GMT 1998 remote from doom\nReceived: from mcit.aris.ru ([194.58.47.100]) by doom.samara.su with SMTP id AA00392\n (5.65/IDA-simtel for <[email protected]>); Mon, 28 Dec 1998 10:36:49 +0400\nReceived: from hub.org by mcit.aris.ru with ESMTP id NAA15833;\n (8.8.5-MVC-230497/vak/1.9) Sun, 27 Dec 1998 13:58:51 +0300 (MSK)\nReceived: from localhost (majordom@localhost)\n\tby hub.org (8.9.1/8.9.1) with SMTP id FAA23484;\n\tSun, 27 Dec 1998 05:56:02 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: by hub.org (TLB v0.10a (1.23 tibbs 1997/01/09 00:29:32)); Sun, 27 Dec 1998 05:55:27 +0000 (EST)\nReceived: (from majordom@localhost)\n\tby hub.org (8.9.1/8.9.1) id FAA23325\n\tfor pgsql-hackers-outgoing; Sun, 27 Dec 1998 05:55:24 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: from trends.net ([email protected] [209.47.148.2])\n\tby hub.org (8.9.1/8.9.1) with ESMTP id FAA23297\n\tfor <[email protected]>; Sun, 27 Dec 1998 05:55:15 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: from awesome.incredible.com (awesome.incredible.com [195.167.39.10])\n\tby trends.net (8.8.8/8.8.8) with ESMTP id FAA03367\n\tfor <[email protected]>; Sun, 27 Dec 1998 05:55:09 -0500 (EST)\nReceived: from localhost (angelos@localhost)\n\tby awesome.incredible.com (8.8.8/8.8.8) with ESMTP id MAA18601\n\tfor <[email protected]>; Sun, 27 Dec 1998 12:38:24 +0200 (EET)\nDate: Sun, 27 Dec 1998 12:38:24 +0200 (EET)\nFrom: Angelos Karageorgiou <[email protected]>\nTo: [email protected]\nSubject: Re: [HACKERS] psql \\d does not work\nIn-Reply-To: <[email protected]>\nMessage-Id: <[email protected]>\nMime-Version: 1.0\nContent-Type: TEXT/PLAIN; charset=US-ASCII\nSender: [email protected]\nPrecedence: bulk\n\nOn Sat, 26 Dec 1998, Angelos Karageorgiou wrote:\n\n> has anyone noticed that the backend still dies when you do a \"\\d\" from\n> psql ? This is with 6.4.1 on bsdi 3.1 \n> \n\nHate to respond to myself, but Bruce noted that other people's \\d on the\nback end did not crash.\n\nI had compiled postgres with shlicc2 which is the shared lib version of gcc2 \non bsdi. That version was crashing the back end. I recompiled with the\ndefault gcc2 and it works magnificently \n\nAny clues ?\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\n",
"msg_date": "Tue, 29 Dec 1998 14:25:22 +0400 (SAMT)",
"msg_from": "[email protected] (Samara CIT)",
"msg_from_op": true,
"msg_subject": "None"
}
] |
[
{
"msg_contents": ">From [194.58.47.100]!mcit.aris.ru!hub.org!owner-pgsql-hackers Mon Dec 28 10:36:52 GMT 1998 remote from doom\nReceived: from mcit.aris.ru ([194.58.47.100]) by doom.samara.su with SMTP id AA00394\n (5.65/IDA-simtel for <[email protected]>); Mon, 28 Dec 1998 10:36:52 +0400\nReceived: from hub.org by mcit.aris.ru with ESMTP id JAA15478;\n (8.8.5-MVC-230497/vak/1.9) Sun, 27 Dec 1998 09:45:26 +0300 (MSK)\nReceived: from localhost (majordom@localhost)\n\tby hub.org (8.9.1/8.9.1) with SMTP id BAA04767;\n\tSun, 27 Dec 1998 01:40:47 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: by hub.org (TLB v0.10a (1.23 tibbs 1997/01/09 00:29:32)); Sun, 27 Dec 1998 01:36:02 +0000 (EST)\nReceived: (from majordom@localhost)\n\tby hub.org (8.9.1/8.9.1) id BAA03806\n\tfor pgsql-hackers-outgoing; Sun, 27 Dec 1998 01:35:28 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: from trends.net ([email protected] [209.47.148.2])\n\tby hub.org (8.9.1/8.9.1) with ESMTP id BAA03232\n\tfor <[email protected]>; Sun, 27 Dec 1998 01:30:30 -0500 (EST)\n\t(envelope-from [email protected])\nReceived: from enrico.ied.com ([email protected] [129.97.130.69])\n\tby trends.net (8.8.8/8.8.8) with ESMTP id WAA26868\n\tfor <[email protected]>; Sat, 26 Dec 1998 22:37:23 -0500 (EST)\nReceived: from ann.ied.com ([email protected] [192.168.200.20])\n\tby enrico.ied.com (8.8.7/8.8.7) with ESMTP id WAA25889\n\tfor <[email protected]>; Sat, 26 Dec 1998 22:36:23 -0500\nDate: Sat, 26 Dec 1998 22:36:53 -0500 (EST)\nFrom: Jan Vicherek <[email protected]>\nTo: [email protected]\nSubject: [HACKERS] space usage of NULL fields\nMessage-Id: <[email protected]>\nMime-Version: 1.0\nContent-Type: TEXT/PLAIN; charset=US-ASCII\nSender: [email protected]\nPrecedence: bulk\n\n\n Hello,\n\n a quick question (I couldn't contact postgresql.org, so I didn't check\nthe FAQ, but I doubt it would be there.)\n\n When I create a table with field \"nnn float8\" and total row size, say\n140 bytes, and then I insert a row with NULL in this field, will my record\ntake up those 8 bytes in physical page (say 140 bytes) or will it take up\nless (like 140-8=132), i.e. will it be shortened by those 8 bytes, since\nthe value is NULL and no information is stored in that field ? ( I *hate*\nrun-on sentences :)\n\n In general, what are the savings/overheads associated in PG with\nNULL/not NULL fields ? I.e. if will I save the space if I define the field\nas varchar(8) instead of float8 ?\n\n Thanx,\n\n Jan\n\n -- Gospel of Jesus is the saving power of God for all who believe --\n ## To some, nothing is impossible. ##\n http://Vicherek.Waterloo.on.ca/\n\n\n",
"msg_date": "Tue, 29 Dec 1998 14:25:23 +0400 (SAMT)",
"msg_from": "[email protected] (Samara CIT)",
"msg_from_op": true,
"msg_subject": "None"
}
] |
[
{
"msg_contents": "\nWhile debugging the JDBC driver I've come across this:\n\ntemplate1=> create user test with password temppass;\nERROR: Bad abstime external representation ''\n*ABORT STATE*\ntemplate1=> create user test with password temppass valid until '2000 dec\n31';\nNOTICE: (transaction aborted): all queries ignored until end of transaction block\ntemplate1=> end;\nEND\ntemplate1=> create user test with password temppass valid until '2000 dec\n31';\nCREATE USER\n\nYou'll notice that I had to define a valid until date to get it working,\nand that when the error occured, I had to end a transaction, even though\none didn't exist before create user was called.\n\n-- \n Peter T Mount [email protected]\n Main Homepage: http://www.retep.org.uk\nPostgreSQL JDBC Faq: http://www.retep.org.uk/postgres\n Java PDF Generator: http://www.retep.org.uk/pdf\n\n",
"msg_date": "Tue, 29 Dec 1998 13:04:42 +0000 (GMT)",
"msg_from": "Peter T Mount <[email protected]>",
"msg_from_op": true,
"msg_subject": "Problem with create user in latest CVS"
}
] |
[
{
"msg_contents": "The recent report of success for NetBSD-m68k reminded me that there are\na few platforms on our \"supported list\" which have not had reports of\nsuccess for Postgres-v6.4 (or if there were reports I neglected to\nupdate the list).\n\nIf anyone has news or success on the following ports, please let us/me\nknow and I'll update the list. Machines which have no reports are likely\nto be listed as unsupported for v6.5 :(\n\nDGUX 5.4R4.11-m88k\n Anyone? Former maintainer lost machine\n\nIRIX 6.x-MIPS\n\nlinux 2.0.x-Alpha\n Recent progress on Alpha machines; does that include linux?\n\nNetBSD 1.3-VAX\n Tom's machine broke. Anyone else?\n\nSCO UnixWare 2.x\n There is another entry for Unixware 7. Is 2.x obsolete?\n\nSunOS (4.1.4?)\n Is non-Solaris SunOS defunct?\n\nSVR4-m88k\n\n\nbtw, we are currently listing 27 supported platforms, including those\nabove.\n\n - Thomas\n",
"msg_date": "Tue, 29 Dec 1998 15:08:10 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Stale porting list?"
},
{
"msg_contents": "\"Thomas G. Lockhart\" <[email protected]> writes:\n> If anyone has news or success on the following ports, please let us/me\n> know and I'll update the list. Machines which have no reports are likely\n> to be listed as unsupported for v6.5 :(\n\n> SunOS (4.1.4?)\n> Is non-Solaris SunOS defunct?\n\nI tested 6.4 on SunOS 4.1.4 and it more or less worked --- no int8\nsupport, and a number of regression tests show differences traceable\nto SunOS's atol() failing to report integer overflow. Stephen Kogge\nhas been looking into whether it's possible to get rid of those problems\nby using pieces of GNU libc, but I dunno whether it's really worth the\ntrouble.\n\nI did see a more serious problem, a backend coredump in the \"opr_sanity\"\ntest. That's been reported on at least one other platform (cf. message\nfrom Keith Parks on the hackers list, 14-Dec), so I think it may be an\nactual portability bug in the 6.4 code. I do not have the time or\nexpertise to track it down though.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Tue, 29 Dec 1998 12:19:22 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Stale porting list? "
},
{
"msg_contents": "> SunOS (4.1.4?)\n> Is non-Solaris SunOS defunct?\n\nI believe this works on 6.4.1.\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, 29 Dec 1998 13:12:48 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [PORTS] Stale porting list?"
},
{
"msg_contents": "\"Thomas G. Lockhart\" wrote:\n> The recent report of success for NetBSD-m68k reminded me that there are\n> a few platforms on our \"supported list\" which have not had reports of\n> success for Postgres-v6.4 (or if there were reports I neglected to\n> update the list).\n> \n> If anyone has news or success on the following ports, please let us/me\n> know and I'll update the list. Machines which have no reports are likely\n> to be listed as unsupported for v6.5 :(\n> \t[...]\n> SCO UnixWare 2.x\n> There is another entry for Unixware 7. Is 2.x obsolete?\n> \nI would say that the UnixWare 2.x is obsolete as it stands now. Most (if not \nall) of the changes made for Unixware 7 would apply to UnixWare 2.1 or greater \nif the SCO Universal Development Kit (UDK) was used to build it, but I no \nlonger have the means to test this.\n\nAlso, the binary generated by the UDK on UnixWare 7.0 should be able to run \nunder UnixWare 2.1+ and SCO OpenServer 5.x. I will obtain the 4.2 tar ball \nand generate a binary version (pkgadd format) that can be used to test this. \n(I already have someone willing to try the binaries on SCO OpenServer.)\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": "Wed, 06 Jan 1999 03:18:12 -0500",
"msg_from": "\"Billy G. Allie\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: Stale porting list? "
}
] |
[
{
"msg_contents": "> Here's the report I sent you earlier. 6.2 was the last version I had\n> any success with on the 88k :-(\n> ... but I would still like to be able to run Postgres on\n> the 88k.\n> > Attempted a build of v6.4 on the Motorola 88k SVR4---\n> > used \"./configure --with-template=svr4\".\n> > config.guess returns \"m88k-motorola-sysv4\".\n> > s_lock.c compile fails because of no slock_t and TAS definition.\n\nThanks Doug. Hopefully we will find a DGUX owner who is interested in\ngetting the m88k stuff to work. Have you been active on the mailing\nlists? Perhaps you could troll for folks who have the right kind of\nmachine to keep this going...\n\n - Tom\n",
"msg_date": "Tue, 29 Dec 1998 15:41:17 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [Fwd: postgresql-v6.4 - m88k svr4]"
}
] |
[
{
"msg_contents": "\nI upgraded my OS from FreeBSD 2.2.7 to 2.2.8 and rebuilt and updated\npostgreSQL to make sure there wouldn't be any library changes or anything\nelse like that to get me. Well the end result is this:\n\npostgres@paprika /usr/local/pgsql> psql template1\nConnection to database 'template1' failed.\nFATAL 1: Database template1 does not exist in pg_database\n\npostgres@paprika /usr/local/pgsql> createdb template1\nConnection to database 'template1' failed.\nFATAL 1: Database template1 does not exist in pg_database\n\ncreatedb: database creation failed on template1.\npostgres@paprika /usr/local/pgsql> \n\npostmaster's running: (here's from ps -awwwx)\n\n/usr/local/pgsql/bin/postmaster -D/usr/local/pgsql/data -i -S -o -F (postgres)\n\n\nAnd yes the kernel is built with the SYSV stuff in it. Everything I'm \nrunning was cvsup'd on 12/18/98.\n\n\nAny ideas? I do have the old binaries around that I might be able to \ninstall over it all and get a pgdump to reload it all, but it seems there\nshould be an easier way.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Searchable Campground Listings http://www.camping-usa.com\n \"There is no outfit less entitled to lecture me about bloat\n than the federal government\" -- Tony Snow\n==========================================================================\n\n\n",
"msg_date": "Tue, 29 Dec 1998 11:04:26 -0500 (EST)",
"msg_from": "Vince Vielhaber <[email protected]>",
"msg_from_op": true,
"msg_subject": "Open for suggestions - IOW: HELP!"
}
] |
[
{
"msg_contents": "Apologies if this has been raised before.\n\nThis:\n\ncreate table a12345678901234567890 ( b1234567890 serial );\n\ncauses this:\n\npqReadData() -- backend closed the channel unexpectedly.\n This probably means the backend terminated abnormally before or\nwhile processing the request.\nWe have lost the connection to the backend, so further processing is\nimpossible. Terminating.\n\n\nThis is on postgres 6.4.1 (the \"fixed\" 6.4.1, not the accidental\nsnapshot) on BSDI.\n\nThe same problem exists in 6.4 on FreeBSD.\n\n",
"msg_date": "Tue, 29 Dec 1998 12:32:20 -0500",
"msg_from": "\"Nat Howard\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "bug with long serial names"
},
{
"msg_contents": "> create table a12345678901234567890 ( b1234567890 serial );\n> causes this:\n> pqReadData() -- backend closed the channel unexpectedly.\n\nI hadn't heard about this. Will look at it. In the meantime, don't do\nthat :)\n\nThe upper limit on names is probably something like\n\n <table name length> + <column name length> + 5 <= 31\n\n - Tom\n",
"msg_date": "Tue, 29 Dec 1998 18:01:17 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] bug with long serial names"
},
{
"msg_contents": "> > create table a12345678901234567890 ( b1234567890 serial );\n> > causes this:\n> > pqReadData() -- backend closed the channel unexpectedly.\n> The upper limit on names is probably something like\n> <table name length> + <column name length> + 5 <= 31\n\nHere is a patch which should catch the problem:\n\npg=> create table a12345678901234567890 ( b1234567890 serial );\nERROR: CREATE TABLE/SERIAL implicit sequence name must be less than 32\ncharacters\n Sum of lengths of 'a12345678901234567890' and 'b1234567890' must\nbe less than 27\n\nWill be in the next release. Apply the patch by putting it into the\nbackend/parser/ directory, then typing:\n\n patch < analyze.c.patch\n\nbtw, the patch should apply cleanly but with offsets since my source\ncode has some other changes you will not want (yet).\n\n - Tom",
"msg_date": "Tue, 29 Dec 1998 18:34:26 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] bug with long serial names"
}
] |
[
{
"msg_contents": "I forgot to add the attachments--sorry for the wasted bandwidth!\n\nCarl G. Riches\nSoftware Engineer\nDepartment of Mathematics\nBox 354350\t\t\tvoice: 206-543-5082 or 206-616-3636\nUniversity of Washington\tfax: 206-543-0397\nSeattle, WA 98195-4350\t\tinternet: [email protected]\n\n---------- Forwarded message ----------\nDate: Mon, 28 Dec 1998 17:05:50 -0800 (PST)\nFrom: Carl Riches <[email protected]>\nTo: Bruce Momjian <[email protected]>\nCc: PostgreSQL-general <[email protected]>,\n PostgreSQL-development <[email protected]>\nSubject: Re: [GENERAL] Release of 6.4.1\n\nOn Thu, 24 Dec 1998, Bruce Momjian wrote:\n\n> Here are the list of changes for 6.4.1.\n> \n> Fix prototype for 64-bit platforms\n\nIf this is the prototyping problem in backend/port/snprintf.c that I\ncontributed, it did not help the problem of compiling 6.4.1 on Digital\nUnix 4.0d using either gcc or cc. Here is my environment:\n\n gcc 2.8.1\n gmake GNU Make version 3.77\n Digital cc 4.25\n\nTo set up cc as the compiler, I used the following command:\n\n ./configure --with-CC=cc --without-CXX\n\nThis failed, so I did the following:\n\n gmake distclean\n ./configure\n\nThis also failed, but for a different cause.\n\nI am attaching transcripts of the gmake session for each of the compilers. \nCan anyone provide a suggestion to get 6.4.1 to compile?\n\nThanks,\nCarl\n\nCarl G. Riches\nSoftware Engineer\nDepartment of Mathematics\nBox 354350\t\t\tvoice: 206-543-5082 or 206-616-3636\nUniversity of Washington\tfax: 206-543-0397\nSeattle, WA 98195-4350\t\tinternet: [email protected]",
"msg_date": "Tue, 29 Dec 1998 10:11:01 -0800 (PST)",
"msg_from": "Carl Riches <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [GENERAL] Release of 6.4.1 (fwd)"
},
{
"msg_contents": "> \n> I forgot to add the attachments--sorry for the wasted bandwidth!\n> \n> Carl G. Riches\n> Software Engineer\n> Department of Mathematics\n> Box 354350\t\t\tvoice: 206-543-5082 or 206-616-3636\n> University of Washington\tfax: 206-543-0397\n> Seattle, WA 98195-4350\t\tinternet: [email protected]\n> \n> ---------- Forwarded message ----------\n> Date: Mon, 28 Dec 1998 17:05:50 -0800 (PST)\n> From: Carl Riches <[email protected]>\n> To: Bruce Momjian <[email protected]>\n> Cc: PostgreSQL-general <[email protected]>,\n> PostgreSQL-development <[email protected]>\n> Subject: Re: [GENERAL] Release of 6.4.1\n> \n> On Thu, 24 Dec 1998, Bruce Momjian wrote:\n> \n> > Here are the list of changes for 6.4.1.\n> > \n> > Fix prototype for 64-bit platforms\n> \n> If this is the prototyping problem in backend/port/snprintf.c that I\n> contributed, it did not help the problem of compiling 6.4.1 on Digital\n> Unix 4.0d using either gcc or cc. Here is my environment:\n> \n> gcc 2.8.1\n> gmake GNU Make version 3.77\n> Digital cc 4.25\n> \n> To set up cc as the compiler, I used the following command:\n> \n> ./configure --with-CC=cc --without-CXX\n> \n> This failed, so I did the following:\n> \n> gmake distclean\n> ./configure\n> \n> This also failed, but for a different cause.\n\nI did not apply the entire patch you gave to 6.4.* because we only\nwanted bug fixes we were sure wouldn't break anything. I did apply the\nentires patch to 6.5 tree.\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, 29 Dec 1998 13:18:32 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [GENERAL] Release of 6.4.1 (fwd)"
}
] |
[
{
"msg_contents": "\nDunno if it's known or even important, but I don't recall seeing this \nbefore during cvsup:\n\nCannot calculate checksum for\n\"/usr/local/src/pgsql/pgsql/src/interfaces/odbc/config.h\": No such file or\ndirectory\n\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Searchable Campground Listings http://www.camping-usa.com\n \"There is no outfit less entitled to lecture me about bloat\n than the federal government\" -- Tony Snow\n==========================================================================\n\n\n",
"msg_date": "Tue, 29 Dec 1998 13:18:34 -0500 (EST)",
"msg_from": "Vince Vielhaber <[email protected]>",
"msg_from_op": true,
"msg_subject": "cvsup message"
},
{
"msg_contents": "On Tue, 29 Dec 1998, Vince Vielhaber wrote:\n\n> \n> Dunno if it's known or even important, but I don't recall seeing this \n> before during cvsup:\n> \n> Cannot calculate checksum for\n> \"/usr/local/src/pgsql/pgsql/src/interfaces/odbc/config.h\": No such file or\n> directory\n\n\tTry removing it and doing a cvsup again?\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Tue, 29 Dec 1998 14:32:29 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] cvsup message"
},
{
"msg_contents": "> Dunno if it's known or even important, but I don't recall seeing this\n> before during cvsup:\n> Cannot calculate checksum for\n> \"/usr/local/src/pgsql/pgsql/src/interfaces/odbc/config.h\":\n> No such file or directory\n\nThat file was being touched this morning. Try again. I'm not seeing the\nsymptom on my new v15.4 client. \n\nbtw, will be posting new static-linked clients and servers for both\nlibc5 and glibc2 linux boxes sometime soon.\n\n - Tom\n",
"msg_date": "Tue, 29 Dec 1998 18:41:57 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] cvsup message"
},
{
"msg_contents": "\nOn 29-Dec-98 The Hermit Hacker wrote:\n> On Tue, 29 Dec 1998, Vince Vielhaber wrote:\n> \n>> \n>> Dunno if it's known or even important, but I don't recall seeing this \n>> before during cvsup:\n>> \n>> Cannot calculate checksum for\n>> \"/usr/local/src/pgsql/pgsql/src/interfaces/odbc/config.h\": No such file or\n>> directory\n> \n> Try removing it and doing a cvsup again?\n\nThat worked.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Searchable Campground Listings http://www.camping-usa.com\n \"There is no outfit less entitled to lecture me about bloat\n than the federal government\" -- Tony Snow\n==========================================================================\n\n\n",
"msg_date": "Tue, 29 Dec 1998 14:19:52 -0500 (EST)",
"msg_from": "Vince Vielhaber <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] cvsup message"
}
] |
[
{
"msg_contents": "Hi,\n\nThis appears to have worked, but required a patch to your patch -- which I've attached.\nThe new version passes regression tests in a manner very similar to the old,\nso I'm assuming all is well, here, and that you mistakenly sent me one\nof the features I'm not ready for yet... (CONSTR_NULL was undefined).\n\nI'm just a novice here, so please, please, correct it if necessary!\n\nThanks very much for the quick service!\n\n*** analyze.c\t1998/12/29 18:43:50\t1.2\n--- analyze.c\t1998/12/29 18:48:53\n***************\n*** 7,13 ****\n *\n *\n * IDENTIFICATION\n! *\t $Header: /usr/local/ports/databases/postgres-6.4.1/work/postgresql-6.4.1/src/backend/parser/RCS/analyze.c,v 1.2 1998/12/29 18:43:50 pgsql Exp $\n *\n *-------------------------------------------------------------------------\n */\n--- 7,13 ----\n *\n *\n * IDENTIFICATION\n! *\t $Header: /usr/local/ports/databases/postgres-6.4.1/work/postgresql-6.4.1/src/backend/parser/RCS/analyze.c,v 1.3 1998/12/29 18:48:50 pgsql Exp $\n *\n *-------------------------------------------------------------------------\n */\n***************\n*** 592,606 ****\n \t\t\t\t\t\tconstraint = lfirst(clist);\n \t\t\t\t\t\tswitch (constraint->contype)\n \t\t\t\t\t\t{\n- \t\t\t\t\t\t\tcase CONSTR_NULL:\n- \t\t\t\t\t\t\t\t/* We should mark this explicitly,\n- \t\t\t\t\t\t\t\t * so we can tell if NULL and NOT NULL are both specified\n- \t\t\t\t\t\t\t\t */\n- \t\t\t\t\t\t\t\tif (column->is_not_null)\n- \t\t\t\t\t\t\t\t\telog(ERROR, \"CREATE TABLE/(NOT) NULL conflicting declaration\"\n- \t\t\t\t\t\t\t\t\t\t \" for %s.%s\", stmt->relname, column->colname);\n- \t\t\t\t\t\t\t\tcolumn->is_not_null = FALSE;\n- \t\t\t\t\t\t\t\tbreak;\n \n \t\t\t\t\t\t\tcase CONSTR_NOTNULL:\n \t\t\t\t\t\t\t\tif (column->is_not_null)\n--- 592,597 ----\n\n\n\n",
"msg_date": "Tue, 29 Dec 1998 14:12:04 -0500 (EST)",
"msg_from": "PostgreSQL pseudo-user <[email protected]>",
"msg_from_op": true,
"msg_subject": "RE: [PATCHES] Re: [HACKERS] bug with long serial names"
}
] |
[
{
"msg_contents": "\nDecided to start from scratch. Did a gmake distclean and a fresh cvsup.\nReinstalled and did an initdb. Starting the postmaster gives me this:\n\nIpcMemoryCreate: shmget failed (Invalid argument) key=5432001, size=894392,\npermission=600\nFATAL 1: ShmemCreate: cannot create region\n\nYes the kernel is built with the SYSV stuff. FreeBSD 2.2.8-RELEASE.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Searchable Campground Listings http://www.camping-usa.com\n \"There is no outfit less entitled to lecture me about bloat\n than the federal government\" -- Tony Snow\n==========================================================================\n\n\n",
"msg_date": "Tue, 29 Dec 1998 14:50:40 -0500 (EST)",
"msg_from": "Vince Vielhaber <[email protected]>",
"msg_from_op": true,
"msg_subject": "certainly not a banner day"
},
{
"msg_contents": "On Tue, 29 Dec 1998, Vince Vielhaber wrote:\n\n> \n> Decided to start from scratch. Did a gmake distclean and a fresh cvsup.\n> Reinstalled and did an initdb. Starting the postmaster gives me this:\n> \n> IpcMemoryCreate: shmget failed (Invalid argument) key=5432001, size=894392,\n> permission=600\n> FATAL 1: ShmemCreate: cannot create region\n\nerk...ipcs shows?\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Tue, 29 Dec 1998 16:24:00 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] certainly not a banner day"
},
{
"msg_contents": "\nOn 29-Dec-98 The Hermit Hacker wrote:\n> On Tue, 29 Dec 1998, Vince Vielhaber wrote:\n> \n>> \n>> Decided to start from scratch. Did a gmake distclean and a fresh cvsup.\n>> Reinstalled and did an initdb. Starting the postmaster gives me this:\n>> \n>> IpcMemoryCreate: shmget failed (Invalid argument) key=5432001,\n>> size=894392,\n>> permission=600\n>> FATAL 1: ShmemCreate: cannot create region\n> \n> erk...ipcs shows?\n\nBingo! That showed me that there were two keys sitting there (forgot all\nabout that command). I ipcsrm'd both and postmaster started.\n\nThanks!\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Searchable Campground Listings http://www.camping-usa.com\n \"There is no outfit less entitled to lecture me about bloat\n than the federal government\" -- Tony Snow\n==========================================================================\n\n\n",
"msg_date": "Tue, 29 Dec 1998 15:46:59 -0500 (EST)",
"msg_from": "Vince Vielhaber <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] certainly not a banner day"
}
] |
[
{
"msg_contents": "> > No time this year.\n> > If necessary I will do it next year.\n> > \n> > Thanks.\n> \n> Yes. I need to know if the code change between setitimer and select()\n> causes the difference you are seeing.\n\nOK, I have re-coded that section. I had majorly broken it.\n\nAs part of the fix, I changed the deadlock detection in include/config.h\nfrom 60 to 1. Remember, config.h is generated from config.h.in.\n\nCan you change it back to 60 to see that the timeout is working\nproperly? The one-second default is not going to be noticable during\ntesting.\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, 29 Dec 1998 14:58:36 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] MVCC works in serialized mode!"
},
{
"msg_contents": "> -----Original Message-----\n> From: Bruce Momjian [mailto:[email protected]]\n> Sent: Wednesday, December 30, 1998 4:59 AM\n> To: PostgreSQL-development\n> Cc: [email protected]; [email protected]\n> Subject: Re: [HACKERS] MVCC works in serialized mode!\n> \n> \n> > > No time this year.\n> > > If necessary I will do it next year.\n> > > \n> > > Thanks.\n> > \n> > Yes. I need to know if the code change between setitimer and select()\n> > causes the difference you are seeing.\n> \n> OK, I have re-coded that section. I had majorly broken it.\n> \n> As part of the fix, I changed the deadlock detection in include/config.h\n> from 60 to 1. Remember, config.h is generated from config.h.in.\n> \n> Can you change it back to 60 to see that the timeout is working\n> properly? The one-second default is not going to be noticable during\n> testing.\n>\n\nI changed DEADLOCK_CHECK_TIMER in config.h to 60 and compiled.\nDeadlock check takes about 60 seconds.\nSo timeout is set to 60 properly.\n\nAnd the result is as follows.\n\ncreate table t1 (key int,a int);\ninsert into t1 values (1,1);\n\nSession-1 Session-2\n\nbegin; begin;\nBEGIN BEGIN\nupdate t1 set a=1; \nUPDATE 1 \n update t1 set a=1;\n [ update is blocked ]\nabort;\nABORT\n\t\t [ immediately ]\n UPDATE 1\n\nIt's OK.\n\nThanks a lot.\n\nHiroshi Inoue\[email protected]\n",
"msg_date": "Tue, 5 Jan 1999 18:48:53 +0900",
"msg_from": "\"Hiroshi Inoue\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "RE: [HACKERS] MVCC works in serialized mode!"
}
] |
[
{
"msg_contents": "> I am currently running Postgresql, ver 6.4 (not yet 6.4.1) on a\n> new SGI Octane (IRIX 6.5-mips). The output from 'uname -a' is:\n> IRIX64 mahunui 6.5 11051213 IP30 mips\n\nThanks for the report! Will update the porting info...\n\nOh, and stay the h!@$ away from v6.4.1. It actually will work, but was\nunintentionally a packaging of our current development tree rather than\nour stable v6.4.x tree. v6.4.2 will be released very soon.\n\n - Tom\n",
"msg_date": "Wed, 30 Dec 1998 01:24:15 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: Postgresql v6.4 on IRIX"
}
] |
[
{
"msg_contents": "Hi all,\n\n the exact NUMERIC datatype materializes more and more. What I\n got so far are the four arithmetic base operators, all six\n comparision operators and these functions:\n\n ABS(n)\n CEIL(n)\n EXP(n)\n FLOOR(n)\n LN(n)\n MOD(m,n)\n ROUND(n[,scale])\n SIGN(n)\n SQRT(n)\n TRUNC(n[,scale])\n\n The trigonometric ones I left out for now, but since SQRT(),\n EXP() and LN() work, it wouldn't be that hard to do them\n later (the former produce the same results as bc(1) does -\n shortly tested up to 400 digits after the decimal point).\n\n The speed of the complex functions is IMHO acceptable. For\n example EXP() is 25% better than bc(1) on small numbers but\n needs up to 3 times on very big ones (999). Who ever needs\n EXP(999) or more? The result is about 7.2e433! SQRT() is a\n bit slow - so be it. Postgres shouldn't become a substitute\n for arbitrary precision calculators.\n\n So I think it's time now to move the stuff into the backend.\n Therefor I first need a bunch of OID's (about 30 C functions,\n 20 SQL functions and 10 operators for now). Should I fill up\n all the holes?\n\n For the next step then I need some help in the parser. I\n think it would be good to have anything that looks like a\n double/float and any integer that doesn't fit into an int4\n parsed into a NUMERIC, and if it really should be one of the\n others type-coerced back later. This is the only way to get\n rid of the single quotes for literal NUMERIC values without\n the danger of loosing digits (due to double's limited\n precision). But I'm not sure about all the things that might\n break if doing so.\n\n Rereading this mail I see that I forgot the aggregates. Just\n some more OID's that I need :-).\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, 30 Dec 1998 03:24:14 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": true,
"msg_subject": "NUMERIC needs OID's"
},
{
"msg_contents": "> Hi all,\n> \n> the exact NUMERIC datatype materializes more and more. What I\n> got so far are the four arithmetic base operators, all six\n> comparision operators and these functions:\n> \n> ABS(n)\n> CEIL(n)\n> EXP(n)\n> FLOOR(n)\n> LN(n)\n> MOD(m,n)\n> ROUND(n[,scale])\n> SIGN(n)\n> SQRT(n)\n> TRUNC(n[,scale])\n> \n> The trigonometric ones I left out for now, but since SQRT(),\n> EXP() and LN() work, it wouldn't be that hard to do them\n> later (the former produce the same results as bc(1) does -\n> shortly tested up to 400 digits after the decimal point).\n> \n> The speed of the complex functions is IMHO acceptable. For\n> example EXP() is 25% better than bc(1) on small numbers but\n> needs up to 3 times on very big ones (999). Who ever needs\n> EXP(999) or more? The result is about 7.2e433! SQRT() is a\n> bit slow - so be it. Postgres shouldn't become a substitute\n> for arbitrary precision calculators.\n> \n> So I think it's time now to move the stuff into the backend.\n> Therefor I first need a bunch of OID's (about 30 C functions,\n> 20 SQL functions and 10 operators for now). Should I fill up\n> all the holes?\n\nNo, I don't recommend it. I recommend getting a range of oid's. Oids\nare confusing enough, without trying to collect them scattered all over\na range of values.\n\nThe problem is that we don't have a range of 30 left anymore, and the\nneeds of future development are surely going to eat up the rest. One\nidea is to remove some of the conversion functions we have defined that\nare rarely used. We don't NEED them with Thomas's conversion stuff. \nThomas says native conversion is faster, but if we find some that we\nalmost never use, we could rip them out and use those.\n\nHowever, my recommendation is that we have to start thinking about\nincreasing the maximum allowable system oid.\n\n\t1 - 10\n\t100 - 101\n\t754\n\t842\n\t949\n\t1288 - 1295\n\t1597 - 1599\n\t1608 - 1610\n\t1619 - 1639\n\t1644 -\n\n\nThe max system oid is stored in transam.h as:\n\n\t/* ----------\n\t * note: we reserve the first 16384 object ids for internal use.\n\t * oid's less than this appear in the .bki files. the choice of\n\t * 16384 is completely arbitrary.\n\t * ----------\n\t */\n\t#define BootstrapObjectIdData 16384\n\nThis is 2^14.\n\nWe can increase this to 32k without any noticable problem, as far as I\nknow. That will give us a nice range of availble oids, and allow\nrenumbering if people want to clean up some of current oid mess.\n\nThe only problem is that loading a pg_dump -o is going to probably cause\nrows with oids in the 16k-32k range to duplicate those in the system\ntables. Is that a problem? I am not sure. As far as I know, there is\nno reason oid's have to be unique, especially if they are in different\ntables. contrib/findoidjoins will get messed up by this, but I am not\nsure if that is a serious problem.\n\nI can't figure out another way around it. We could expand by going very\nhigh, near 2 billion, assuming no one is up there yet, but the code will\nget very messy doing that, and I don't recommend it.\n\nI think going to 2^15 is going to become necessary someday. The\nquestion is, do we do it for 6.5, and if so, how will the duplicate oids\naffect our users?\n\nA new system would start assigning rows with oids > 2^15, so only system\ntable oids installed via pg_dump -o or COPY WITH OIDS FROM would be\nin the range 2^14-2^15.\n\npg_dump uses the max system oid to determine of a function is a user\nfunction, but because it gets the max oid from the template1 table oid,\nthis should be portable across the two oid systems.\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, 29 Dec 1998 23:52:31 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NUMERIC needs OID's"
},
{
"msg_contents": "Neat stuff!\n\n> > So I think it's time now to move the stuff into the backend.\n> > Therefor I first need a bunch of OID's (about 30 C functions,\n> > 20 SQL functions and 10 operators for now). Should I fill up\n> > all the holes?\n> No, I don't recommend it. I recommend getting a range of oid's. Oids\n> are confusing enough, without trying to collect them scattered all \n> over a range of values.\n> The problem is that we don't have a range of 30 left anymore, and the\n> needs of future development are surely going to eat up the rest. One\n> idea is to remove some of the conversion functions we have defined \n> that are rarely used. We don't NEED them with Thomas's conversion \n> stuff. Thomas says native conversion is faster, but if we find some \n> that we almost never use, we could rip them out and use those.\n\nThat won't get us very many OIDs, so doesn't solve the problem in the\nlong run.\n\n> However, my recommendation is that we have to start thinking about\n> increasing the maximum allowable system oid.\n> The max system oid is stored in transam.h as:\n> #define BootstrapObjectIdData 16384\n\nJan is asking for 20-50 OIDs. That is not enough to significantly affect\nthe size of the available \"reserved OID\" space. I'd recommend taking a\nblock of OIDs from somewhere just above the largest current pre-assigned\nOID.\n\n> I think going to 2^15 is going to become necessary someday. The\n> question is, do we do it for 6.5, and if so, how will the duplicate \n> oids affect our users?\n\nThis seems like a good subject for v7.0, when users might expect a more\ntroublesome transition. Anyway, by then we might have a different\nsolution.\n\n - Tom\n",
"msg_date": "Wed, 30 Dec 1998 06:23:45 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NUMERIC needs OID's"
},
{
"msg_contents": "> For the next step then I need some help in the parser. I\n> think it would be good to have anything that looks like a\n> double/float and any integer that doesn't fit into an int4\n> parsed into a NUMERIC, and if it really should be one of the\n> others type-coerced back later. This is the only way to get\n> rid of the single quotes for literal NUMERIC values without\n> the danger of loosing digits (due to double's limited\n> precision). But I'm not sure about all the things that might\n> break if doing so.\n\nI can help with this. I would suggest starting conservatively,\nconverting to NUMERIC only when one of the other existing types does not\nwork. Later, we can more tightly couple the numeric types to make this\nmore transparent.\n\nWhen you have the numeric stuff ready to go, commit it to the source\ntree and we can start working on the parsing issues.\n\n - Tom\n",
"msg_date": "Wed, 30 Dec 1998 14:55:44 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NUMERIC needs OID's"
},
{
"msg_contents": "> > The problem is that we don't have a range of 30 left anymore, and the\n> > needs of future development are surely going to eat up the rest. One\n> > idea is to remove some of the conversion functions we have defined \n> > that are rarely used. We don't NEED them with Thomas's conversion \n> > stuff. Thomas says native conversion is faster, but if we find some \n> > that we almost never use, we could rip them out and use those.\n> \n> That won't get us very many OIDs, so doesn't solve the problem in the\n> long run.\n> \n> > However, my recommendation is that we have to start thinking about\n> > increasing the maximum allowable system oid.\n> > The max system oid is stored in transam.h as:\n> > #define BootstrapObjectIdData 16384\n\nI am a dope. The first user oid is 16k, and unused_oids says:\n\n\t$ unused_oids\n\t1 - 10\n\t100 - 101\n\t754\n\t842\n\t949\n\t1288 - 1295\n\t1597 - 1599\n\t1608 - 1610\n\t1619 - 1639\n\t1644 -\n\nThere is nothing magical about 1644. We have always been packing the\noids into that range, while we have never used oids over 1644. Just\nstart at 1650 and head toward 16384. We have roughly 16200 unused\nsystem oids left.\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, 30 Dec 1998 11:53:08 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NUMERIC needs OID's"
},
{
"msg_contents": "Bruce Momjian wrote:\n\n> [...]\n> oids into that range, while we have never used oids over 1644. Just\n> start at 1650 and head toward 16384. We have roughly 16200 unused\n> system oids left.\n\n I've just committed the NUMERIC stuff to the CURRENT tree.\n Used OID's 1700-1767 up to now (some more will follow when\n implementing the trigonometric functions).\n\n Well folks - it's there now and can be stressed! Don't forget\n to initdb.\n\n I'll add a first precision check using some of the complex\n math functions with scale 1000 to the regression soon.\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, 30 Dec 1998 21:29:13 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": true,
"msg_subject": "NUMERIC stuff committed"
}
] |
[
{
"msg_contents": "Hi!\n\nI have to recover old db, produced by pgsql 6.1. The binaries are lost and \nfresh compiled postgres dumps core. Investigation narrowed the problem\ndown to postgresql-v6.1.1/src/backend/utils/cache/catcache.c:comphash\n\nWith debug turned on, I'm getting the following output:\n\n[akhavr@kbi akhavr]$ date; sudo -u postgres /home/akhavr/postgres \\\n-D /home/postgres/data -e -d 2 mailingad\nWed Dec 30 12:40:32 EET 1998\nFindBackend: found \"/home/akhavr/postgres\" using argv[0]\n\t---debug info---\n\tQuiet = f\n\tNoversion = f\n\tstable = f\n\ttimings = f\n\tdates = European\n\tbufsize = 64\n\tquery echo = f\n\tmultiplexed backend? = f\n\tDatabaseName = [mailingad]\n\t----------------\n\n\tInitPostgres()..\nbinding ShmemCreate(key=0, size=758552)\nDEBUG:InitSysCache: rid=0 id=0 nkeys=3 size=500\n[...]\nDEBUG:CatalogCacheInitializeCache: cache @081d3720\nDEBUG:CatalogCacheInitializeCache: called w/relname pg_user\nDEBUG:CatalogCacheInitializeCache: relid 1260, 1 keys\nDEBUG:CatalogCacheInitializeCache: load 1/1 w/1, -1\nDEBUG:CatalogCacheInit pg_user 0 -1 81d3720\nDEBUG:CatalogCacheComputeHashIndex pg_user 1 -1 0 81d3720\nDEBUG:comphash (-1,8187980)\nSegmentation fault (core dumped)\n[akhavr@kbi akhavr]$ sudo cp /home/postgres/data/base/mailingad/core .; \\\nsudo chown akhavr core\n[akhavr@kbi akhavr]$ gdb postgres core\nGDB is free software and you are welcome to distribute copies of it\n under certain conditions; type \"show copying\" to see the conditions.\nThere is absolutely no warranty for GDB; type \"show warranty\" for details.\nGDB 4.16 (i586-ksi-linux), Copyright 1996 Free Software Foundation, Inc...\nCore was generated by `/home/akhavr/postgres -D /home/postgres/data -e -d 2 mailingad'.\nProgram terminated with signal 11, Segmentation fault.\nReading symbols from /usr/lib/libcrypt.so.1...done.\nReading symbols from /lib/libdl.so.1...done.\nReading symbols from /lib/libm.so.5...done.\nReading symbols from /usr/lib/libreadline.so.2...done.\nReading symbols from /lib/libtermcap.so.2...done.\nReading symbols from /lib/libc.so.5...done.\nReading symbols from /lib/ld-linux.so.1...done.\n#0 comphash (l=1953333615, v=0x8187200 \"postgres\") at catcache.c:299\n299 i += *v++;\n(gdb) bt\n#0 comphash (l=1953333615, v=0x8187200 \"postgres\") at catcache.c:299\n#1 0x8136cec in CatalogCacheComputeHashIndex (cacheInP=0x81d2fa0)\n at catcache.c:334\n#2 0x813751b in SearchSysCache (cache=0x81d2fa0, v1=135819776, v2=0, v3=0,\n v4=0) at catcache.c:773\n#3 0x813a65e in SearchSysCacheTuple (cacheId=21, key1=135819776, key2=0,\n key3=0, key4=0) at syscache.c:416\n#4 0x813ef2b in SetUserId () at miscinit.c:331\n#5 0x813f40b in InitUserid () at postinit.c:341\n#6 0x813f661 in InitPostgres (name=0xbffffce1 \"mailingad\") at postinit.c:636\n#7 0x8107aaf in PostgresMain (argc=7, argv=0xbffffb94) at postgres.c:1237\n#8 0x80b144d in main (argc=7, argv=0xbffffb94) at main.c:68\n#9 0x806238e in _start ()\n(gdb)\n\nLooks like l is treated as unsigned.\n\nCan anyone suggest the way to fix the bug, so I can dump out the db and\nreload it to the last stable postgresql version?\n\n-- \nSY, Andrey V Khavryutchenko\thttp://www.kbi.kiev.ua/~akhavr\n\nShick's Law:\n\tThere is no problem a good miracle can't solve.\n",
"msg_date": "30 Dec 1998 13:05:30 +0200",
"msg_from": "Andrey V Khavryutchenko <[email protected]>",
"msg_from_op": true,
"msg_subject": "problems recovering 6.1 db"
},
{
"msg_contents": "> I have to recover old db, produced by pgsql 6.1. The binaries are \n> lost and fresh compiled postgres dumps core.\n> Can anyone suggest the way to fix the bug, so I can dump out the db \n> and reload it to the last stable postgresql version?\n\nOut of curiosity, what platform are you running on? I've still got v6.1\n(and v6.1.1) tar files and source trees on my linux/libc5 machine, and\ncould build binaries for you if they happen to match. Are you *sure* you\nhad a v6.1 installation, as opposed to a v6.1.1? Both releases had fixes\nfor some hashing operations, and I'll bet they need to match exactly\nwith your database.\n\nAlso, if your \"mailingad\" database is small, not confidential, and\nbinary compatible with my machine I could try reading it here.\n\nGood luck.\n\n - Tom\n",
"msg_date": "Wed, 30 Dec 1998 15:10:14 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] problems recovering 6.1 db"
},
{
"msg_contents": "Hi, Thomas! \n\n>>>>> \"TGL\" == Thomas G Lockhart writes:\n\n>> I have to recover old db, produced by pgsql 6.1. The binaries are lost\n>> and fresh compiled postgres dumps core. Can anyone suggest the way to\n>> fix the bug, so I can dump out the db and reload it to the last stable\n>> postgresql version?\n\nTGL> Out of curiosity, what platform are you running on? I've still got\nTGL> v6.1 (and v6.1.1) tar files and source trees on my linux/libc5\nTGL> machine, and could build binaries for you if they happen to match. Are\nTGL> you *sure* you had a v6.1 installation, as opposed to a v6.1.1? \n\nI'm running linux/libc5. And, yes, I'm not sure, that which version was\nused to create db. I've tried to use 6.1.1, now I'll try to build 6.1 and\nretry.\n\nTGL> Both releases had fixes for some hashing operations, and I'll bet they\nTGL> need to match exactly with your database.\n\nTGL> Also, if your \"mailingad\" database is small, not confidential, and\nTGL> binary compatible with my machine I could try reading it here.\n\nUnfortunately, it's impossible. It's several hunderds megabytes and\nlimited access. \n\nThanks! If I'll fail with 6.1 too, I'll repost.\n\n-- \nSY, Andrey V Khavryutchenko\thttp://www.kbi.kiev.ua/~akhavr\n\nShick's Law:\n\tThere is no problem a good miracle can't solve.\n",
"msg_date": "31 Dec 1998 12:39:15 +0200",
"msg_from": "Andrey V Khavryutchenko <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] problems recovering 6.1 db"
}
] |
[
{
"msg_contents": "Hi all\n\nI just noticed that when loading a dump file, all the echoes of the stuff\nin the dump file go to stderr.\n\nIs that really what you want? I would have thought errors would be sent\nto stderr and that every thing else would go to stdout.\n\nHave a great night\nTerry Mackintosh <[email protected]> http://www.terrym.com\nsysadmin/owner Please! No MIME encoded or HTML mail, unless needed.\n\nProudly powered by R H Linux 4.2, Apache 1.3.x, PHP 3.x, PostgreSQL 6.x\n-----------------------------------------------------------------------\nSuccess Is A Choice ... book by Rick Patino, get it, read it!\n\n",
"msg_date": "Wed, 30 Dec 1998 19:09:48 -0500 (EST)",
"msg_from": "Terry Mackintosh <[email protected]>",
"msg_from_op": true,
"msg_subject": "psql: stdout or stderr?"
},
{
"msg_contents": "> Hi all\n> \n> I just noticed that when loading a dump file, all the echoes of the stuff\n> in the dump file go to stderr.\n> \n> Is that really what you want? I would have thought errors would be sent\n> to stderr and that every thing else would go to stdout.\n> \n\nYes, this is new. As I remember, some stuff is sent to stdout, and some\nto stderr, and there was some reason for that, though I can't remember\nwhat it was. \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, 30 Dec 1998 19:42:46 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] psql: stdout or stderr?"
},
{
"msg_contents": "Bruce Momjian <[email protected]> writes:\n>> Is that really what you want? I would have thought errors would be sent\n>> to stderr and that every thing else would go to stdout.\n\n> Yes, this is new. As I remember, some stuff is sent to stdout, and some\n> to stderr, and there was some reason for that, though I can't remember\n> what it was. \n\nI've always thought that psql is pretty inconsistent, arbitrary, and\nunreasonable about what it chooses to write to stdout vs stderr.\n\nHowever, if we rejigger what goes where, we will surely break a lot of\nusers' shell scripts, since those already expect particular output to go\nto one or the other. (We'd also break all the regression tests, but\nat least those are within our power to fix.)\n\nOn the whole I'd vote against changing it --- a small improvement in\nconsistency is not worth the pain it will cause.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Thu, 31 Dec 1998 11:20:13 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] psql: stdout or stderr? "
},
{
"msg_contents": "> Bruce Momjian <[email protected]> writes:\n> >> Is that really what you want? I would have thought errors would be sent\n> >> to stderr and that every thing else would go to stdout.\n> \n> > Yes, this is new. As I remember, some stuff is sent to stdout, and some\n> > to stderr, and there was some reason for that, though I can't remember\n> > what it was. \n> \n> I've always thought that psql is pretty inconsistent, arbitrary, and\n> unreasonable about what it chooses to write to stdout vs stderr.\n> \n> However, if we rejigger what goes where, we will surely break a lot of\n> users' shell scripts, since those already expect particular output to go\n> to one or the other. (We'd also break all the regression tests, but\n> at least those are within our power to fix.)\n> \n> On the whole I'd vote against changing it --- a small improvement in\n> consistency is not worth the pain it will cause.\n\nOK, I will tell you what it does, and you tell me what it should do. \n\nSomewhere around 6.3, someone complained that he couldn't see how a\ndatabase reload was going, so he suggested that we output the table\ninformation to stdout, and the SQL queries to stderr, so:\n\n\techo \"select * from test;\" | psql demodb\n\nsends the text 'QUERY: select * from test;' to stderr, and the table\ncontents of the select to stdout. Now, if you do 'psql -q demodb', you\nwill not get anything on stderr, except error messages. -q suppresses\nthe query output.\n\nI believe this behavior also sends COPY FROM data rows to stdout, so a\npg_dump reload would probably want to see the SQL queries, so you can\nsee how far it had gotten, but not the data rows being loaded, which go\nto stdout.\n\nWhat we really need is three output streams from psql, one for data, one\nfor SQL queries, and one for errors.\n\nCurrently, the way we control this is to control how the SQL queries are\ndisplayed. Using -q causes them to be suppressed. Using -e causes them\nto be displayed twice.\n\nWhat is the best way to handle this? A bigger question is, \"Can you\nmake sense of an error message if you don't have the SQL query that\ncaused it?\"\n\nNow I can say that the error return code from psql should accurately\nshow if an error occurred. I did that code perhaps a year ago, and it\nshould still work.\n\nIdeally, you would want errors in a separate file, but also be able to\ngo back and see the query with the error message right next to it.\n\nSo, basically I am saying that the default behavior is to output the SQL\nqueries to stderr, but -q turns that off.\n\nI am interested in ways to improve 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": "Thu, 31 Dec 1998 15:18:08 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] psql: stdout or stderr?"
}
] |
[
{
"msg_contents": "\n-:)\n\nAnd, btw, :\n\ngcc -I../../../include -I../../../backend -O2 -m486 -pipe -Wall\n-Wmissing-prototypes -ggdb3 -I../.. -c numeric.c -o numeric.o\nnumeric.c:19: nan.h: No such file or directory\nnumeric.c: In function `numeric_float8':\nnumeric.c:1726: `NAN' undeclared (first use this function)\nnumeric.c:1726: (Each undeclared identifier is reported only once\nnumeric.c:1726: for each function it appears in.)\nnumeric.c: In function `numeric_float4':\nnumeric.c:1776: `NAN' undeclared (first use this function)\ngmake: *** [numeric.o] Error 1\n\n\n(FreeBSD 2.2.7)\n\nVadim\n",
"msg_date": "Thu, 31 Dec 1998 12:37:48 +0700",
"msg_from": "Vadim Mikheev <[email protected]>",
"msg_from_op": true,
"msg_subject": "Happy New Year!"
}
] |
[
{
"msg_contents": "Hi all,\n\nSeems that no one of the examples on 'man create_rule' does works.\nWhat about to try every example on man pages?\nSeems that's posssible to create a rule on event SELECT only to define\nviews\nAm I wrong ?\n\nI'm trying to create a rule like this:\n\ncreate rule my_rule as\n on select to employees\n do instead select INITCAP(ename),\n UPPER(empno,\n UPPER(job),\n hiredate,\n sal,\n comm,\n deptno,\n level,\n mgr\n from employees;\n\nAny ideas?\n\nThanks,\n\n-Jose'-\n",
"msg_date": "Thu, 31 Dec 1998 10:08:11 +0100",
"msg_from": "\"Jose' Soares\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "create rule"
},
{
"msg_contents": ">\n> Hi all,\n>\n> Seems that no one of the examples on 'man create_rule' does works.\n> What about to try every example on man pages?\n> Seems that's posssible to create a rule on event SELECT only to define\n> views\n> Am I wrong ?\n\n You're right and the man page for CREATE RULE is outdated.\n Rules on the SELECT event are 100% restricted to view\n definitions now (one single instead action that selects\n exactly the same attribute types and names the relation\n itself has).\n\n I must still update that man page, thanks.\n\n In the meantime you might want to take a look at the rule\n system documentation of the programmers manual, which I've\n written entirely from scratch after the fixes to the rule\n system for v6.4.\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, 4 Jan 1999 10:21:26 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] create rule"
}
] |
[
{
"msg_contents": "Here are some patches which fix the recently reported date/time problems\non linux/glibc2 machines. Note that these will also fix any machine\nwhich \"tests positive\" for HAVE_TM_ZONE, which might include SunOS and\nxBSD machines.\n\nThe symptom can be demonstrated by the following query:\n\n select datetime('2000-01-01'::date);\n\nwhich, if there is a problem, will return a time away from midnight.\nNote that your time zone must be set to something other than GMT for\nthis test to be meaningful.\n\nThanks Oleg for being persistant!\n\n - Tom",
"msg_date": "Thu, 31 Dec 1998 16:58:23 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Date/time fixes for HAVE_TM_ZONE platforms"
},
{
"msg_contents": "Hello!\n\nOn Thu, 31 Dec 1998, Thomas G. Lockhart wrote:\n> Here are some patches which fix the recently reported date/time problems\n[skip]\n> Thanks Oleg for being persistant!\n\n Cannot apply these patches. Which version/snapshot do I need?\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": "Sun, 3 Jan 1999 16:40:10 +0300 (MSK)",
"msg_from": "Oleg Broytmann <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: Date/time fixes for HAVE_TM_ZONE platforms"
},
{
"msg_contents": "> Cannot apply these patches. Which version/snapshot do I need?\n\nThe patches are probably for the current development tree, but both\nCURRENT and REL6_4 have already had the patches applied. So just refresh\nyour cvs snapshot if you have that available...\n\n\nWhat version of the tree do you have?\n\n - Tom\n",
"msg_date": "Mon, 04 Jan 1999 07:12:29 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: Date/time fixes for HAVE_TM_ZONE platforms"
},
{
"msg_contents": "Hello!\n\nOn Mon, 4 Jan 1999, Thomas G. Lockhart wrote:\n> What version of the tree do you have?\n\n I am not postgres hacker, just average user. I am running postgres for\nusing it (mostly as backend for CGIs). Currently I run 6.4 and do some\nexperiments with 6.4.1.\n\n> > Cannot apply these patches. Which version/snapshot do I need?\n> \n> The patches are probably for the current development tree, but both\n> CURRENT and REL6_4 have already had the patches applied. So just refresh\n> your cvs snapshot if you have that available...\n\n Does this mean 6.4.2 will have it applied?\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": "Tue, 5 Jan 1999 11:49:18 +0300 (MSK)",
"msg_from": "Oleg Broytmann <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: Date/time fixes for HAVE_TM_ZONE platforms"
},
{
"msg_contents": "Hello!\n\nOn Thu, 31 Dec 1998, Thomas G. Lockhart wrote:\n> Here are some patches which fix the recently reported date/time problems\n\nPostgres 6.4.2:\n\ntemplate1=> select datetime('1991-10-01', '11:00');\ndatetime \n----------------------------\nTue 01 Oct 11:00:00 1991 EET\n(1 row)\n\n Now the last step - timezone. I remember there was broken 6.4.1 that had\ncorrect timezone here...\n My timezone is, of course MSK/MSD. Global tzname really is EET, but\ntm->tz_name is MSK.\n\nOleg.\n---- \n Oleg Broytmann http://members.tripod.com/~phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n",
"msg_date": "Tue, 5 Jan 1999 19:09:29 +0300 (MSK)",
"msg_from": "Oleg Broytmann <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: Date/time fixes for HAVE_TM_ZONE platforms"
},
{
"msg_contents": "> > Here are some patches which fix the date/time problems\n> Postgres 6.4.2:\n> Now the last step - timezone. I remember there was broken 6.4.1 that \n> had correct timezone here...\n> My timezone is, of course MSK/MSD. Global tzname really is EET, but\n> tm->tz_name is MSK.\n\nYeah, so is mine. You will note that any year other than 1991 behaves as\nyou expect, and 1991 results are consistant with the Linux timezone\ndatabase (see below):\n\npostgres=> select datetime('1991-10-01'::date);\ndatetime\n----------------------------\nMon Sep 30 23:00:00 1991 EET\n(1 row)\n\npostgres=> select datetime('1992-10-01'::date);\ndatetime\n----------------------------\nThu Oct 01 00:00:00 1992 MSK\n(1 row)\n\npostgres=> select datetime('1990-10-01'::date);\ndatetime\n----------------------------\nMon Oct 01 00:00:00 1990 MSK\n(1 row)\n\nApparently 1991 had something different for timezone handling for Moscow\ntime. It isn't as dumb as it sounds; in the US in 1973-4 we ran on\ndaylight savings time through the winter season as an energy\nconservation measure:\n\npostgres=> select '1973-02-01'::datetime;\n?column?\n----------------------------\nThu Feb 01 00:00:00 1973 PST\n(1 row)\n\npostgres=> select '1974-02-01'::datetime;\n?column?\n----------------------------\nFri Feb 01 00:00:00 1974 PDT\n(1 row)\n\nAnyway, you can see the feature in the \"zic\" timezone database:\n\ngolem$ zdump -v -c 2000 Europe/Moscow\n<snip>\nEurope/Moscow Sat Sep 29 23:00:00 1990 GMT = Sun Sep 30 02:00:00 1990\nMSK isdst=0\nEurope/Moscow Sat Mar 30 22:59:59 1991 GMT = Sun Mar 31 01:59:59 1991\nMSK isdst=0\nEurope/Moscow Sat Mar 30 23:00:00 1991 GMT = Sun Mar 31 02:00:00 1991\nEET DST isdst=1\nEurope/Moscow Sat Sep 28 23:59:59 1991 GMT = Sun Sep 29 02:59:59 1991\nEET DST isdst=1\nEurope/Moscow Sun Sep 29 00:00:00 1991 GMT = Sun Sep 29 02:00:00 1991\nEET isdst=0\nEurope/Moscow Sat Jan 18 23:59:59 1992 GMT = Sun Jan 19 01:59:59 1992\nEET isdst=0\nEurope/Moscow Sun Jan 19 00:00:00 1992 GMT = Sun Jan 19 03:00:00 1992\nMSK isdst=0\nEurope/Moscow Sat Mar 28 22:59:59 1992 GMT = Sun Mar 29 01:59:59 1992\nMSK isdst=0\nEurope/Moscow Sat Mar 28 23:00:00 1992 GMT = Sun Mar 29 03:00:00 1992\nMSD isdst=1\n<snip>\n\n - Tom\n",
"msg_date": "Tue, 05 Jan 1999 17:16:03 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: Date/time fixes for HAVE_TM_ZONE platforms"
},
{
"msg_contents": "Hello!\n\n I have choosed nice test case, right? :)))\n Of course, I knew this before, just completely forgot when I got good\nsource for zic. I haven't looked into it for a while.\n Ok, I'll recreate patch for datetime regression test and submit it\ntomorrow.\n Thank you.\n\nOn Tue, 5 Jan 1999, Thomas G. Lockhart wrote:\n> Yeah, so is mine. You will note that any year other than 1991 behaves as\n> you expect, and 1991 results are consistant with the Linux timezone\n> database (see below):\n> \n> postgres=> select datetime('1991-10-01'::date);\n> datetime\n> ----------------------------\n> Mon Sep 30 23:00:00 1991 EET\n> (1 row)\n> \n> postgres=> select datetime('1992-10-01'::date);\n> datetime\n> ----------------------------\n> Thu Oct 01 00:00:00 1992 MSK\n> (1 row)\n> \n> postgres=> select datetime('1990-10-01'::date);\n> datetime\n> ----------------------------\n> Mon Oct 01 00:00:00 1990 MSK\n> (1 row)\n> \n> Apparently 1991 had something different for timezone handling for Moscow\n> time. It isn't as dumb as it sounds; in the US in 1973-4 we ran on\n> daylight savings time through the winter season as an energy\n> conservation measure:\n> \n> postgres=> select '1973-02-01'::datetime;\n> ?column?\n> ----------------------------\n> Thu Feb 01 00:00:00 1973 PST\n> (1 row)\n> \n> postgres=> select '1974-02-01'::datetime;\n> ?column?\n> ----------------------------\n> Fri Feb 01 00:00:00 1974 PDT\n> (1 row)\n> \n> Anyway, you can see the feature in the \"zic\" timezone database:\n> \n> golem$ zdump -v -c 2000 Europe/Moscow\n> <snip>\n> Europe/Moscow Sat Sep 29 23:00:00 1990 GMT = Sun Sep 30 02:00:00 1990\n> MSK isdst=0\n> Europe/Moscow Sat Mar 30 22:59:59 1991 GMT = Sun Mar 31 01:59:59 1991\n> MSK isdst=0\n> Europe/Moscow Sat Mar 30 23:00:00 1991 GMT = Sun Mar 31 02:00:00 1991\n> EET DST isdst=1\n> Europe/Moscow Sat Sep 28 23:59:59 1991 GMT = Sun Sep 29 02:59:59 1991\n> EET DST isdst=1\n> Europe/Moscow Sun Sep 29 00:00:00 1991 GMT = Sun Sep 29 02:00:00 1991\n> EET isdst=0\n> Europe/Moscow Sat Jan 18 23:59:59 1992 GMT = Sun Jan 19 01:59:59 1992\n> EET isdst=0\n> Europe/Moscow Sun Jan 19 00:00:00 1992 GMT = Sun Jan 19 03:00:00 1992\n> MSK isdst=0\n> Europe/Moscow Sat Mar 28 22:59:59 1992 GMT = Sun Mar 29 01:59:59 1992\n> MSK isdst=0\n> Europe/Moscow Sat Mar 28 23:00:00 1992 GMT = Sun Mar 29 03:00:00 1992\n> MSD isdst=1\n> <snip>\n> \n> - Tom\n> \n\nOleg.\n---- \n Oleg Broytmann http://members.tripod.com/~phd2/ [email protected]\n Programmers don't die, they just GOSUB without RETURN.\n\n",
"msg_date": "Tue, 5 Jan 1999 20:26:35 +0300 (MSK)",
"msg_from": "Oleg Broytmann <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: Date/time fixes for HAVE_TM_ZONE platforms"
}
] |
[
{
"msg_contents": "I have updated unused_oids so it now shows the max system oid in the\ndisplay, rather than just a dash,\n\n\t1 - 10\n\t100 - 101\n\t754\n\t842\n\t949\n\t1288 - 1295\n\t1597 - 1599\n\t1608 - 1610\n\t1619 - 1639\n\t1644 - 1699\n\t1768 - 16383\n\nThe last line used to be:\n\n\t1768 -\n\nwhich was obviously confusing me and probably others.\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, 31 Dec 1998 15:21:53 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Unused_oids"
}
] |
[
{
"msg_contents": "Two small patches are attached, both against current CVS:\n\n1. Some security, since we now have vsnprintf, I remade an old patch\n with some extra ugly sprintfs fixed. More work in this area is\n needed still. (pg.diff.gz)\n\n2. Is a experimental patch to libpq and psql to make XML output an\n alternative to to traditional and HTML3 output. (pg2.diff.gz)\n\nI been think about making XML an alternative FE/BE protocol, but\nfor now I implemented a hack on the client side. Check it out\nas a basis for discussion of structure etc.\n\n\tmvh,\n-- \n-----------------\nG�ran Thyni\nThis is Penguin Country. On a quiet night you can hear Windows NT\nreboot!",
"msg_date": "Thu, 31 Dec 1998 22:23:48 +0100",
"msg_from": "Goran Thyni <[email protected]>",
"msg_from_op": true,
"msg_subject": "2 patches"
},
{
"msg_contents": "[Charset iso-8859-1 unsupported, filtering to ASCII...]\n> \n> Two small patches are attached, both against current CVS:\n> \n> 1. Some security, since we now have vsnprintf, I remade an old patch\n> with some extra ugly sprintfs fixed. More work in this area is\n> needed still. (pg.diff.gz)\n\nApplied to CURRENT tree.\n\n> \n> 2. Is a experimental patch to libpq and psql to make XML output an\n> alternative to to traditional and HTML3 output. (pg2.diff.gz)\n\nNot applied, as you suggested.\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, 31 Dec 1998 23:50:31 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: 2 patches"
},
{
"msg_contents": "\nYesterday I posted a quick patch to make libpq able to output\nsome home brew (but valid) XML as an alternative to normal and HTML3\noutput.\n\nThe propose was to start a discussion how we could use XML in PgSQL.\n\nI mentioned in my previous mail I would like to use XML in the FE/BE\nprotocol. That would leav us with 3 protocols, the current one, CORBA\nand XML. But I would argue that CORBA and XML are more related since\nthey are both object based. In the long run the current protocol could \nbe dropped.\n\nPro and cons of using XML:\n+ Object-oriented (fits well with our ORDBMS project)\n+ Clear text (easily debugged)\n+ Easliy display in next generation network tools (even without\n middleware in some cases)\n+ In coupled to DOM, easily manipulatable both on server and/or client\nside.\n- In theory a little slower than an binary protocol.\n\nXML is a buzzword, but it has enormous potential,\nand it is not foreign to the pgsql community since\nwe are already using the father of XML (SGML) for\nthe docs.\n\nSuggestion:\nCORBA and XML should replace current FE/BE-protocol for v. 7.0\nComments?\n\n\tbest regards and a wishes for a good 1999,\n-- \n-----------------\nG�ran Thyni\nThis is Penguin Country. On a quiet night you can hear Windows NT\nreboot!\n",
"msg_date": "Fri, 01 Jan 1999 16:11:55 +0100",
"msg_from": "Goran Thyni <[email protected]>",
"msg_from_op": true,
"msg_subject": "XML (was: 2 patches)"
},
{
"msg_contents": "> XML is a buzzword, but it has enormous potential,\n> and it is not foreign to the pgsql community since\n> we are already using the father of XML (SGML) for\n> the docs.\n> \n> Suggestion:\n> CORBA and XML should replace current FE/BE-protocol for v. 7.0\n> Comments?\n\nI vote against it. Maybe nice to have as an extra, but the base\nprotocol seems to work well, and is fast.\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, 1 Jan 1999 15:44:06 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: XML (was: 2 patches)"
},
{
"msg_contents": "Bruce Momjian wrote:\n> \n> > XML is a buzzword, but it has enormous potential,\n> > and it is not foreign to the pgsql community since\n> > we are already using the father of XML (SGML) for\n> > the docs.\n> >\n> > Suggestion:\n> > CORBA and XML should replace current FE/BE-protocol for v. 7.0\n> > Comments?\n> \n> I vote against it. Maybe nice to have as an extra, but the base\n> protocol seems to work well, and is fast.\n\nYou are right,\nlet me rephrase myself:\n- the \"base protocol\" might become obsolite if and when we have\n a better object base alternative (CORBA and/or XML).\n- this might be the case for v. 7.0 or later.\n\nThis is an idea for a game plan:\n1. Make a proxy server\n2. Integrate it in postmaster when stable (code-wise and protocol-wise)\n3. Add closer integration with DB-objects\n\nThis game plan would perhaps work to get CORBA of the ground too,\nthere might even be sensible to make a common CORBA/XML-proxy.\n\nI will look in to this, but will be working out of town for 2 months\n(mainly feb-mars) so I will not promise anything for 6.5.\n\nAny help or input (right now comment on the DTD etc) welcome.\n\n\tregards,\n-- \n-----------------\nG�ran Thyni\nThis is Penguin Country. On a quiet night you can hear Windows NT\nreboot!\n",
"msg_date": "Sat, 02 Jan 1999 00:33:53 +0100",
"msg_from": "Goran Thyni <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Re: XML (was: 2 patches)"
},
{
"msg_contents": "On Sat, 2 Jan 1999, Goran Thyni wrote:\n\n> Bruce Momjian wrote:\n> > \n> > > XML is a buzzword, but it has enormous potential,\n> > > and it is not foreign to the pgsql community since\n> > > we are already using the father of XML (SGML) for\n> > > the docs.\n> > >\n> > > Suggestion:\n> > > CORBA and XML should replace current FE/BE-protocol for v. 7.0\n> > > Comments?\n> > \n> > I vote against it. Maybe nice to have as an extra, but the base\n> > protocol seems to work well, and is fast.\n> \n> You are right,\n> let me rephrase myself:\n> - the \"base protocol\" might become obsolite if and when we have\n> a better object base alternative (CORBA and/or XML).\n> - this might be the case for v. 7.0 or later.\n\nOkay, we have to sides to the coin here...Bruce, who feels that our\ncurrent protocol is both good and fast, and you, who wants to replace it\n:)\n\nMy question is simple...what benefit is gained by replacing current\nprotocol? And, what cons? \n\nReplacing something is great, if there is a reason...what is that reason\nand how does that improve us?\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Fri, 1 Jan 1999 19:48:45 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: XML (was: 2 patches)"
},
{
"msg_contents": "The Hermit Hacker wrote:\n> On Sat, 2 Jan 1999, Goran Thyni wrote:\n> > Bruce Momjian wrote:\n> > > I vote against it. Maybe nice to have as an extra, but the base\n> > > protocol seems to work well, and is fast.\n> >\n> > You are right,\n> > let me rephrase myself:\n> > - the \"base protocol\" might become obsolite if and when we have\n> > a better object base alternative (CORBA and/or XML).\n> > - this might be the case for v. 7.0 or later.\n> \n> Okay, we have to sides to the coin here...Bruce, who feels that our\n> current protocol is both good and fast, and you, who wants to replace it\n> :)\n> \n> My question is simple...what benefit is gained by replacing current\n> protocol? And, what cons?\n> \n> Replacing something is great, if there is a reason...what is that reason\n> and how does that improve us?\n\nMarc,\n\nThe pros, IMHO:\n* XML interfaces better to the object model of pgsql itself\n* XML interfaces better to XML-viewer, like next generation of\n web browsers.\n\nWith a little imagination we can at some point in the future forget\nabout making clients for the databases, just point your webbrowser\nat the serverport a make both SELECTs (displays) and INSERT/UPDATE\n(forms) with a few lines of javascript.\n\nThe con:\n* We have a good stable protocol but a little simplistic.\n Anything new is not as good until it has matured. \n\nThus I not suggesting replacing it until we have something\nbetter, sound resonable?\n\n\tmvh,\n-- \n-----------------\nG�ran Thyni\nThis is Penguin Country. On a quiet night you can hear Windows NT\nreboot!\n",
"msg_date": "Sat, 02 Jan 1999 01:29:14 +0100",
"msg_from": "Goran Thyni <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Re: XML (was: 2 patches)"
},
{
"msg_contents": "> The pros, IMHO:\n> * XML interfaces better to the object model of pgsql itself\n> * XML interfaces better to XML-viewer, like next generation of\n> web browsers.\n> \n> With a little imagination we can at some point in the future forget\n> about making clients for the databases, just point your webbrowser\n> at the serverport a make both SELECTs (displays) and INSERT/UPDATE\n> (forms) with a few lines of javascript.\n> \n> The con:\n> * We have a good stable protocol but a little simplistic.\n> Anything new is not as good until it has matured. \n> \n> Thus I not suggesting replacing it until we have something\n> better, sound resonable?\n\nIt is appealing to be able to talk to the database via some\nsgml/html/xml format. A proxy would allow this too, and it certainly\nsomething that would be nice. Certainly no reason we can't add it, and\nif we start out as a proxy, we don't have to add it to the backend\nitself.\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, 1 Jan 1999 20:01:28 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: XML (was: 2 patches)"
},
{
"msg_contents": "On Fri, 1 Jan 1999, Bruce Momjian wrote:\n\n> > The pros, IMHO:\n> > * XML interfaces better to the object model of pgsql itself\n> > * XML interfaces better to XML-viewer, like next generation of\n> > web browsers.\n> > \n> > With a little imagination we can at some point in the future forget\n> > about making clients for the databases, just point your webbrowser\n> > at the serverport a make both SELECTs (displays) and INSERT/UPDATE\n> > (forms) with a few lines of javascript.\n> > \n> > The con:\n> > * We have a good stable protocol but a little simplistic.\n> > Anything new is not as good until it has matured. \n> > \n> > Thus I not suggesting replacing it until we have something\n> > better, sound resonable?\n> \n> It is appealing to be able to talk to the database via some\n> sgml/html/xml format. A proxy would allow this too, and it certainly\n> something that would be nice. Certainly no reason we can't add it, and\n> if we start out as a proxy, we don't have to add it to the backend\n> itself.\n\n\tAnd, of course, as it matures, its something we can always roll\ninto the backend :)\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Fri, 1 Jan 1999 23:13:54 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: XML (was: 2 patches)"
},
{
"msg_contents": "Goran Thyni <[email protected]> writes:\n> The pros, IMHO:\n> * XML interfaces better to the object model of pgsql itself\n\nCORBA does that as well or better, I suspect.\n\n> * XML interfaces better to XML-viewer, like next generation of\n> web browsers.\n\nSo? I can't get excited about being able to view raw server output in\na web browser --- it seems to me that any interesting web application\nof Postgres is going to involve an intermediate layer of code anyway\n(to convert client actions into SQL commands, adorn the output with page\nheaders and buttons and so forth, etc etc). I don't see anything about\nXML that eliminates the need for that intermediate layer, and I don't\nsee much about XML that makes that layer easier to write.\n\n> With a little imagination we can at some point in the future forget\n> about making clients for the databases, just point your webbrowser\n> at the serverport a make both SELECTs (displays) and INSERT/UPDATE\n> (forms) with a few lines of javascript.\n\nI think you drastically underestimate the amount of glue code that will\nbe needed, and overestimate the advantage of having the results come\nback already formatted in an HTML-like fashion. (OK, it'd make direct\ndisplay a little easier, but what if you have to extract some data for\nprocessing? Writing an XML parser in Javascript doesn't sound like a\npiece of cake to me.)\n\nIf XML had big benefits for non-web-browser front ends, it might still\nbe a win, but I'm not seeing anything much there either.\n\nIt should also be pointed out that XML isn't a *protocol*, just a\ndata representation. If you choose to represent the results of SELECTs\nin XML, that still leaves you with all the other aspects of FE/BE\ncommunication yet to be designed.\n\nI thought the proposal to build a CORBA-based interface for Postgres\nwas pretty interesting and worthy of further study ... I'd rather see\nus put our time into that. (I am certainly not wedded to the current\nFE/BE protocol; it gets the job done, but it's crufty and has\nperformance limitations.)\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sat, 02 Jan 1999 12:25:12 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Re: XML (was: 2 patches) "
},
{
"msg_contents": "Tom Lane wrote:\n> \n> Goran Thyni <[email protected]> writes:\n> > The pros, IMHO:\n> > * XML interfaces better to the object model of pgsql itself\n> \n> CORBA does that as well or better, I suspect.\n> \n> > * XML interfaces better to XML-viewer, like next generation of\n> > web browsers.\n> \n> So? I can't get excited about being able to view raw server output in\n> a web browser --- it seems to me that any interesting web application\n> of Postgres is going to involve an intermediate layer of code anyway\n> (to convert client actions into SQL commands, adorn the output with page\n> headers and buttons and so forth, etc etc). I don't see anything about\n> XML that eliminates the need for that intermediate layer, and I don't\n> see much about XML that makes that layer easier to write.\n\nThis is what DOM is for, read on.\n \n> > With a little imagination we can at some point in the future forget\n> > about making clients for the databases, just point your webbrowser\n> > at the serverport a make both SELECTs (displays) and INSERT/UPDATE\n> > (forms) with a few lines of javascript.\n> \n> I think you drastically underestimate the amount of glue code that will\n> be needed, and overestimate the advantage of having the results come\n> back already formatted in an HTML-like fashion. (OK, it'd make direct\n> display a little easier, but what if you have to extract some data for\n> processing? Writing an XML parser in Javascript doesn't sound like a\n> piece of cake to me.)\n\nThe XML parser is in the browser. It parses to XML-code into a DOM which\nis easily accessed from javascript.\n\nThe step are:\n1/ a DTD defines the parsing of our document type into the DOM\n2/ a style sheet determines the layout on our output device\n3/ javascript access to the DOM for making tweak to layout and\nfunctions.\n\nThe client side is still immature and some parts are in still flux.\n\n> If XML had big benefits for non-web-browser front ends, it might still\n> be a win, but I'm not seeing anything much there either.\n\nThis is a chicken-egg problem,\nwe must have data to make clients interesting\nand\nwe must have clients to make server output interesting\n\nCORBA has the same problem, we must start somewhere.\nAs I am a \"server-side kind of guy\", I will start there\nand hope clients catch up.\n\nOn the client side I put most faith in \"Gecko\" \nthe new layout engine from mozilla.org.\n\n> It should also be pointed out that XML isn't a *protocol*, just a\n> data representation. If you choose to represent the results of SELECTs\n> in XML, that still leaves you with all the other aspects of FE/BE\n> communication yet to be designed.\n\nHTTP and/or IIOP/CORBA is useable for that.\n\n> I thought the proposal to build a CORBA-based interface for Postgres\n> was pretty interesting and worthy of further study ... I'd rather see\n> us put our time into that.\n\nOne does not rule out the other,\nI think that much of the object functionality \nof both ideas has much in common with seperate\noutput mechanisms, like:\n\n\t---------------------------------------\n\t| PgSQL Core | \n\t---------------------------------------\n\t| Object Data Access | Base |\n\t---------------------------- |\n\t| CORBA | XML | protocol |\n\t---------------------------------------\n\t| Clients | \n\t---------------------------------------\n\nAnyway, thanks for your input Tom.\nQuestioning my ideas, forces me to develop them\nand fill in the blanks, which is \"good thing(tm)\".\nKeep arguments, ideas and questions coming!\n\n\tmvh,\n-- \n-----------------\nG�ran Thyni\nThis is Penguin Country. On a quiet night you can hear Windows NT\nreboot!\n",
"msg_date": "Sat, 02 Jan 1999 19:22:05 +0100",
"msg_from": "Goran Thyni <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] Re: XML (was: 2 patches)"
},
{
"msg_contents": "Anyone doing serious SGML/XML playing (import/export) from PostgreSQL?\n\nI'm going to spend time writing a \"journal\", which is a generic\nrepository\nfor sequencital storage of incoming XML.\n\n\nThanks!\n\nClark\n",
"msg_date": "Tue, 05 Jan 1999 00:58:45 +0000",
"msg_from": "Clark Evans <[email protected]>",
"msg_from_op": false,
"msg_subject": "XML/SGML"
}
] |
[
{
"msg_contents": "I have modified the current tree so it will compile on BSD/OS, and\nprobably FreeBSD.\n\nI don't have a NAN define. My libm source looks like it assigns\nHUGE_VAL to those entries requiring a NAN.\n \nI have also commented out the inclusion of \"nan.h\" because I don't have\nthat either.\n\nThis will probably require further tweaking for other OS's.\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, 31 Dec 1998 17:36:55 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "NAN value"
}
] |
[
{
"msg_contents": "I have changed the NAN code, so if NAN is not defined, we just use 'num'\nas NAN, since we only get here if num is NAN. Seems like a good fix for\nplatforms that can't assign a NAN to a variable.\n\n---------------------------------------------------------------------------\n\n if (NUMERIC_IS_NAN(num))\n {\n result = (float32)palloc(sizeof(float32data));\n#ifdef NAN\n *result = NAN;\n#else\n *result = num;\n#endif\n return result;\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, 31 Dec 1998 23:16:28 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "NAN code"
},
{
"msg_contents": ">\n> I have changed the NAN code, so if NAN is not defined, we just use 'num'\n> as NAN, since we only get here if num is NAN. Seems like a good fix for\n> platforms that can't assign a NAN to a variable.\n>\n> ---------------------------------------------------------------------------\n>\n> if (NUMERIC_IS_NAN(num))\n> {\n> result = (float32)palloc(sizeof(float32data));\n> #ifdef NAN\n> *result = NAN;\n> #else\n> *result = num;\n> #endif\n> return result;\n> }\n\n And I where so happy to have finally found that damned NAN\n define :-}\n\n Anyway - that above is absolutely wrong.\n\n num is of type Numeric, which is a pointer to struct\n NumericData! So the above should result in some compiler\n warning (in fact it type casts a pointer into a double value\n - and that will give unpredictable results).\n\n Seems that only isnan() is defined as part of Posix. But not\n a definition that can force a NAN. So we have to find a\n portable way to define the value NaN for double and float.\n Does anybody know of such a way?\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, 4 Jan 1999 09:21:33 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NAN code"
},
{
"msg_contents": ">\n> I have changed the NAN code, so if NAN is not defined, we just use 'num'\n> as NAN, since we only get here if num is NAN. Seems like a good fix for\n> platforms that can't assign a NAN to a variable.\n>\n> ---------------------------------------------------------------------------\n>\n> if (NUMERIC_IS_NAN(num))\n> {\n> result = (float32)palloc(sizeof(float32data));\n> #ifdef NAN\n> *result = NAN;\n> #else\n> *result = num;\n> #endif\n> return result;\n> }\n\n And I where so happy to have finally found that damned NAN\n define :-}\n\n Anyway - that above is absolutely wrong.\n\n num is of type Numeric, which is a pointer to struct\n NumericData! So the above should result in some compiler\n warning (in fact it type casts a pointer into a double value\n - and that will give unpredictable results).\n\n Seems that only isnan() is defined as part of Posix. But not\n a definition that can force a NAN. So we have to find a\n portable way to define the value NaN for double and float.\n Does anybody know of such a way?\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\n",
"msg_date": "Mon, 4 Jan 1999 09:21:33 +0100 (MET)",
"msg_from": "[email protected]",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NAN code"
},
{
"msg_contents": "[email protected] writes:\n\n> Seems that only isnan() is defined as part of Posix. But not\n> a definition that can force a NAN. So we have to find a\n> portable way to define the value NaN for double and float.\n> Does anybody know of such a way?\n\nBruce changed it to defining NAN as (0.0/0.0) if NAN isn't already\ndefined. This works on FreeBSD. I don't know if it will break on other\nsystems that don't have NAN already.\n",
"msg_date": "04 Jan 1999 08:37:59 -0600",
"msg_from": "Hal Snyder <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NAN code"
},
{
"msg_contents": "> num is of type Numeric, which is a pointer to struct\n> NumericData! So the above should result in some compiler\n> warning (in fact it type casts a pointer into a double value\n> - and that will give unpredictable results).\n> \n> Seems that only isnan() is defined as part of Posix. But not\n> a definition that can force a NAN. So we have to find a\n> portable way to define the value NaN for double and float.\n> Does anybody know of such a way?\n\nSee my later postings. 0.0/0.0 seems to do 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, 4 Jan 1999 10:59:53 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] NAN code"
},
{
"msg_contents": "> > Seems that only isnan() is defined as part of Posix. But not\n> > a definition that can force a NAN. So we have to find a\n> > portable way to define the value NaN for double and float.\n> > Does anybody know of such a way?\n> Bruce changed it to defining NAN as (0.0/0.0) if NAN isn't already\n> defined. This works on FreeBSD. I don't know if it will break on other\n> systems that don't have NAN already.\n\nThe NaN concept is not portable on every platform, since it is an IEEE\n754 floating point feature. Older machines like VAX don't have it, and\nolder compilers don't support it cleanly. The divide-by-zero technique\nto get NaN is likely to result in a signaled FPE, where we want a quiet\nassignment of a special value (NaN on those platforms for which it is\navailable).\n\nHere is the code we already have in Postgres to support the date/time\ndata types; you'll see NAN, HUGE_VAL, and DBL_MAX used for special\nvalues:\n\n#ifdef NAN\n#define DT_INVALID (NAN)\n#else\n#define DT_INVALID (DBL_MIN+DBL_MIN)\n#endif\n#ifdef HUGE_VAL\n#define DT_NOBEGIN (-HUGE_VAL)\n#define DT_NOEND (HUGE_VAL)\n#else\n#define DT_NOBEGIN (-DBL_MAX)\n#define DT_NOEND (DBL_MAX)\n#endif\n#define DT_CURRENT (DBL_MIN)\n#define DT_EPOCH (-DBL_MIN)\n\n#define DATETIME_INVALID(j) {j = DT_INVALID;}\n#ifdef NAN\n#define DATETIME_IS_INVALID(j) (isnan(j))\n#else\n#define DATETIME_IS_INVALID(j) (j == DT_INVALID)\n#endif\n\nOf course, for historical reasons the date/time stuff has these special\nreserved values which may not be necessary for the numerics. If\nnecessary, I would suggest something similar for the numeric stuff, or\nperhaps on platforms without NAN and no way to generate it then just\nthrow an error.\n\nAs an aside, I worked on this a few years ago on my Alpha-dunix boxes\n(for converting back and forth between VAX and IEEE floating point\nbinary values) and found that for the DEC compiler you had to use a\nparticular compiler flag to keep NaN as an unsignaled value. Otherwise\nthe call to isnan() failed when the value was put on the stack to set up\nthe function call to isnan() :/\n\n - Tom\n",
"msg_date": "Mon, 04 Jan 1999 16:30:54 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NAN code"
},
{
"msg_contents": "Bruce Momjian wrote:\n\n>\n> > Seems that only isnan() is defined as part of Posix. But not\n> > a definition that can force a NAN. So we have to find a\n> > portable way to define the value NaN for double and float.\n> > Does anybody know of such a way?\n>\n> See my later postings. 0.0/0.0 seems to do it.\n\n Seen them. Just that I'm a little in doubt if this construct\n couldn't generate a SIGFPE on all of our supported\n platform/compiler combos. Still think we should add autoconf\n stuff to search for a NAN definition and only fallback to the\n above if that fails.\n\n While searching for the NAN definition I've noticed too that\n our float4/float8 datatypes can output 'NaN', but do not\n parse them back. They simply elog(ERROR, ...) if you try to\n use 'NaN' as an input string for a floating point attribute.\n Shouldn't all input functions be able to parse back any\n possible result of the corresponding output function? As of\n now, I cannot imagine a construct (except a user defined C\n function), that could result in a float8-NaN value stored in\n the database. But as soon as it happens, the database\n wouldn't any longer be dump/reloadable!\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, 4 Jan 1999 17:41:19 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NAN code"
},
{
"msg_contents": "> While searching for the NAN definition I've noticed too that\n> our float4/float8 datatypes can output 'NaN', but do not\n> parse them back.\n\nYeah, I just noticed that a few minutes ago too.\n\nI've got some patches which fix this, at least for machines with NaN\nsupport (and for HUGE_VAL/Infinity too). We'll need some more code to\nhandle non-IEEE machines...\n\n - Tom\n",
"msg_date": "Mon, 04 Jan 1999 18:17:50 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NAN code"
},
{
"msg_contents": ">\n> > While searching for the NAN definition I've noticed too that\n> > our float4/float8 datatypes can output 'NaN', but do not\n> > parse them back.\n>\n> Yeah, I just noticed that a few minutes ago too.\n>\n> I've got some patches which fix this, at least for machines with NaN\n> support (and for HUGE_VAL/Infinity too). We'll need some more code to\n> handle non-IEEE machines...\n\n The final might be that for machines/compilers with NaN\n support, NAN is\n\n #define NAN (0.0/0.0)\n\n or whatever is the correct thing for it (best choice is still\n the definition from a system header like nan.h if available),\n while for those who cannot support it it'll be\n\n #define NAN unsupported_nan_request()\n\n double unsupported_nan_request(void)\n {\n elog(ERROR, \"unsupported request for NaN as floating point value\");\n return 0.0;\n }\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, 4 Jan 1999 19:49:32 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] NAN code"
}
] |
[
{
"msg_contents": "I have found a way to assign NAN to a variable, thanks to the library\nsource code. Is this better than the code I just posted? Looks pretty\nplatform-specific, and we may be better with my earlier code to just use\nnum, which I have committed to the CURRENT tree.\n\nThe strange thing is that isnan(0/0) generates a floating-pointer error\nand stops the program, but this does not. Strange.\n\n---------------------------------------------------------------------------\n\t\n\t#include <stdio.h>\n\t#include <math.h>\n\t\n\tmain()\n\t{\n\t\tdouble x;\n\t\tx = 3;\n\t\t\n\t\tif (isnan((x-x)/(x-x)))\n\t\t\tprintf(\"nan\\n\");\n\t\telse\n\t\t\tprintf(\"not nan\\n\");\n\t}\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, 31 Dec 1998 23:23:48 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": true,
"msg_subject": "NAN code"
}
] |
[
{
"msg_contents": "DEAR SIR:\n WE HAVE LOT OF COMPUTER STOCK.WANT TO \nSALE IF YOU ARE INTERESTING PLS OPEN THE ATTACH\nFILE. YOU WILL FIND GOOD PRICE ABOUT \nMEMORY/CPU/MOBILE/APPLIANCE...............ETC\n\n\n MR ALBERT",
"msg_date": "Sat, 2 Jan 1999 04:11:45 +0800 (CST)",
"msg_from": "[email protected]",
"msg_from_op": true,
"msg_subject": "COMPUTER STOCK ON SALE /CPU/MEMORY/MOBILE.......ETC. "
}
] |
[
{
"msg_contents": "\nFor those interested in my XML-ideas,\nI stubled over this page/document talking about XML, data and RDBMSs:\n\nhttp://www.w3.org/TR/1998/NOTE-XML-data/\n",
"msg_date": "Sat, 02 Jan 1999 00:43:26 +0100",
"msg_from": "Goran Thyni <[email protected]>",
"msg_from_op": true,
"msg_subject": "XML-Data"
}
] |
[
{
"msg_contents": "\nI'm trying to add a unique index to a table. Notice the error message below:\n\n\nTable = ordertrack\n+----------------------------------+----------------------------------+------+\n| Field | Type |Length|\n+----------------------------------+----------------------------------+------+\n| ordernumber | int8 | 8|\n| total | money | 4|\n| cookie | text | var|\n| orderseq | int4 not null default nextval('o | 4|\n+----------------------------------+----------------------------------+------+\nIndex: ordertrack_orderseq_key\ncloud9=> create unique index ord1 on ordertrack(ordernumber);\nERROR: Can't find a default operator class for type 20.\ncloud9=> \n\n\nIs this telling me that I can't have an index of type int8? Why not?\n\nFreeBSD 2.2.8, PostgreSQL cvsup'd less than a week ago.\n\nVince.\n-- \n==========================================================================\nVince Vielhaber -- KA8CSH email: [email protected] flame-mail: /dev/null\n # include <std/disclaimers.h> TEAM-OS2\n Online Searchable Campground Listings http://www.camping-usa.com\n \"There is no outfit less entitled to lecture me about bloat\n than the federal government\" -- Tony Snow\n==========================================================================\n\n\n",
"msg_date": "Sat, 02 Jan 1999 17:18:42 -0500 (EST)",
"msg_from": "Vince Vielhaber <[email protected]>",
"msg_from_op": true,
"msg_subject": "Error creating index"
},
{
"msg_contents": "> Is this telling me that I can't have an index of type int8? Why not?\n> FreeBSD 2.2.8, PostgreSQL cvsup'd less than a week ago.\n\nBecause I added the int8 data type to the backend without knowing which\nplatforms would actually support it. (As an aside, I had tried to\nsolicit results of a small test program for various platforms on the\nhackers list, but can't recall getting any responses.)\n\nWithout knowing that, it wasn't clear to me whether the int8 would be\nable to stay, and I didn't want to embed these changes in the index\nsupport system tables if they would need to be removed. These tables are\nthe most difficult to configure and support, and it isn't as easy to\ntrace which entries support which types.\n\nAnyway, it seems int8 indices would be a good feature to have for v6.5,\neh? One could write some user-defined index operators as an interim\nsolution.\n\n - Tom\n",
"msg_date": "Mon, 04 Jan 1999 06:22:33 +0000",
"msg_from": "\"Thomas G. Lockhart\" <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Error creating index"
},
{
"msg_contents": "> Index: ordertrack_orderseq_key\n> cloud9=> create unique index ord1 on ordertrack(ordernumber);\n> ERROR: Can't find a default operator class for type 20.\n> cloud9=>\n\n Reminds me that the new NUMERIC type need's an operator class\n too. Will take a look at int8 while doing 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, 4 Jan 1999 09:24:01 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Error creating index"
},
{
"msg_contents": "> Index: ordertrack_orderseq_key\n> cloud9=> create unique index ord1 on ordertrack(ordernumber);\n> ERROR: Can't find a default operator class for type 20.\n> cloud9=>\n\n Reminds me that the new NUMERIC type need's an operator class\n too. Will take a look at int8 while doing 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\n",
"msg_date": "Mon, 4 Jan 1999 09:24:01 +0100 (MET)",
"msg_from": "[email protected]",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Error creating index"
}
] |
[
{
"msg_contents": "The blizzard hitting the Chicago area makes today ideal for hacking -\nas long as the power stays on! A thousand pardons if this is a known\nproblem - I'm just getting started with the current CVS set.\n\nI grabbed postgresql.snapshot.tar.gz for 8 am, Jan 2 and am trying to\nbuild it on my FreeBSD-2.2.7 system in order to test some things for\nthe web site. This seems to be the only source kit available from\nhub's anonymous FTP archive, BTW.\n\nHere's what happens:\n\n...\ngmake[2]: Entering directory `/pgsql-src/pgsql/src/backend/utils'\nfor i in adt cache error fmgr hash init misc mmgr sort time; do gmake -C $i SUBSYS.o; done\ngmake[3]: Entering directory `/pgsql-src/pgsql/src/backend/utils/adt'\ngcc -I../../../include -I../../../backend -O2 -m486 -pipe -Wall -Wmissing-prototypes -I../.. -c numeric.c -o numeric.o\nnumeric.c: In function `numeric_float8':\nnumeric.c:1729: incompatible types in assignment\nnumeric.c: In function `numeric_float4':\nnumeric.c:1783: incompatible types in assignment\ngmake[3]: *** [numeric.o] Error 1\ngmake[3]: Leaving directory `/pgsql-src/pgsql/src/backend/utils/adt'\n...\ngmake[2]: *** No rule to make target `adt/SUBSYS.o', needed by `SUBSYS.o'. Stop.\ngmake[2]: Leaving directory `/pgsql-src/pgsql/src/backend/utils'\ngmake[1]: *** [utils.dir] Error 2\ngmake[1]: Leaving directory `/pgsql-src/pgsql/src/backend'\ngmake: *** [all] Error 2\n\nI think the compiler is right in claiming incompatible types. In rev\n1.3 of backend/utils/adt/numeric.c, line 1783 says\n\n\t*result = num;\n\nbut \"result\" is a float32 (=*float), while \"num\" is Numeric (pointer\nto a custom PostgreSQL NumericData struct). What is intended here?\n\nHal\n",
"msg_date": "02 Jan 1999 16:34:31 -0600",
"msg_from": "Hal Snyder <[email protected]>",
"msg_from_op": true,
"msg_subject": "snapshot and FreeBSD-2.2.7 not building"
},
{
"msg_contents": "> The blizzard hitting the Chicago area makes today ideal for hacking -\n> as long as the power stays on! A thousand pardons if this is a known\n> problem - I'm just getting started with the current CVS set.\n> \n> I grabbed postgresql.snapshot.tar.gz for 8 am, Jan 2 and am trying to\n> build it on my FreeBSD-2.2.7 system in order to test some things for\n> the web site. This seems to be the only source kit available from\n> hub's anonymous FTP archive, BTW.\n> \n> Here's what happens:\n> \n> ...\n> gmake[2]: Entering directory `/pgsql-src/pgsql/src/backend/utils'\n> for i in adt cache error fmgr hash init misc mmgr sort time; do gmake -C $i SUBSYS.o; done\n> gmake[3]: Entering directory `/pgsql-src/pgsql/src/backend/utils/adt'\n> gcc -I../../../include -I../../../backend -O2 -m486 -pipe -Wall -Wmissing-prototypes -I../.. -c numeric.c -o numeric.o\n> numeric.c: In function `numeric_float8':\n> numeric.c:1729: incompatible types in assignment\n> numeric.c: In function `numeric_float4':\n> numeric.c:1783: incompatible types in assignment\n> gmake[3]: *** [numeric.o] Error 1\n> gmake[3]: Leaving directory `/pgsql-src/pgsql/src/backend/utils/adt'\n> ...\n> gmake[2]: *** No rule to make target `adt/SUBSYS.o', needed by `SUBSYS.o'. Stop.\n> gmake[2]: Leaving directory `/pgsql-src/pgsql/src/backend/utils'\n> gmake[1]: *** [utils.dir] Error 2\n> gmake[1]: Leaving directory `/pgsql-src/pgsql/src/backend'\n> gmake: *** [all] Error 2\n> \n> I think the compiler is right in claiming incompatible types. In rev\n> 1.3 of backend/utils/adt/numeric.c, line 1783 says\n> \n> \t*result = num;\n> \n> but \"result\" is a float32 (=*float), while \"num\" is Numeric (pointer\n> to a custom PostgreSQL NumericData struct). What is intended here?\n\nOK, I have applied a patch to the cvs tree. My old code was clearly\nwrong, now that I study it.\n\nI am now generating NAN by:\n\n\tdouble x;\n\n\treturn (x-x)/(x-x);\n\nThis seems to be the way my libm math library does it. The gcc2 -O4\ncompile does not seem to optimize it away, so looks like it will work. \nMy guess is that the compiler can not optimize floating-point\ncomputations.\n\nStill waiting for Jan to give me his comments on my fixes, but it is\nholiday time.\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, 2 Jan 1999 21:41:56 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] snapshot and FreeBSD-2.2.7 not building"
},
{
"msg_contents": "> I am now generating NAN by:\n> \n> \tdouble x=1;\n ^^\n> \n> \treturn (x-x)/(x-x);\n\nForgot the =1.\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, 2 Jan 1999 21:54:10 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] snapshot and FreeBSD-2.2.7 not building"
},
{
"msg_contents": "Bruce Momjian <[email protected]> writes:\n\n> I am now generating NAN by:\n> \n> \tdouble x;\n> \n> \treturn (x-x)/(x-x);\n> \n> This seems to be the way my libm math library does it. The gcc2 -O4\n> compile does not seem to optimize it away, so looks like it will work. \n> My guess is that the compiler can not optimize floating-point\n> computations.\n\nThanks for the follow-up. Suggestion:\n\nAt the top of numeric.c or in include/utils/numeric.h, put\n\n#ifndef NAN\n#define pgNAN (0.0/0.0)\n#endif\n\nThen in numeric.c, if NAN isn't defined,\n\n\treturn pgNAN;\n\n--\nHal\n",
"msg_date": "02 Jan 1999 21:50:57 -0600",
"msg_from": "Hal Snyder <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] snapshot and FreeBSD-2.2.7 not building"
},
{
"msg_contents": "> Thanks for the follow-up. Suggestion:\n> \n> At the top of numeric.c or in include/utils/numeric.h, put\n> \n> #ifndef NAN\n> #define pgNAN (0.0/0.0)\n> #endif\n> \n> Then in numeric.c, if NAN isn't defined,\n\nDone. Used:\n\n\t#ifndef NAN\n\t#define NAN (0.0/0.0)\n\t#endif\n\nat the top of numeric.c, and removed my other #ifdef's.\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, 3 Jan 1999 00:30:09 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] snapshot and FreeBSD-2.2.7 not building"
},
{
"msg_contents": ">\n> > Thanks for the follow-up. Suggestion:\n> >\n> > At the top of numeric.c or in include/utils/numeric.h, put\n> >\n> > #ifndef NAN\n> > #define pgNAN (0.0/0.0)\n> > #endif\n> >\n> > Then in numeric.c, if NAN isn't defined,\n>\n> Done. Used:\n>\n> #ifndef NAN\n> #define NAN (0.0/0.0)\n> #endif\n>\n> at the top of numeric.c, and removed my other #ifdef's.\n\n I hope this is portable and none of our supported platforms\n throws in a SIGFPE for it (cross fingers).\n\n OTOH it might be good to have configure checks if NAN is\n defined somewhere and include that header. Only if it isn't\n defined we should fallback to the above.\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\n",
"msg_date": "Mon, 4 Jan 1999 09:33:27 +0100 (MET)",
"msg_from": "[email protected]",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] snapshot and FreeBSD-2.2.7 not building"
},
{
"msg_contents": ">\n> > Thanks for the follow-up. Suggestion:\n> >\n> > At the top of numeric.c or in include/utils/numeric.h, put\n> >\n> > #ifndef NAN\n> > #define pgNAN (0.0/0.0)\n> > #endif\n> >\n> > Then in numeric.c, if NAN isn't defined,\n>\n> Done. Used:\n>\n> #ifndef NAN\n> #define NAN (0.0/0.0)\n> #endif\n>\n> at the top of numeric.c, and removed my other #ifdef's.\n\n I hope this is portable and none of our supported platforms\n throws in a SIGFPE for it (cross fingers).\n\n OTOH it might be good to have configure checks if NAN is\n defined somewhere and include that header. Only if it isn't\n defined we should fallback to the above.\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, 4 Jan 1999 09:33:28 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] snapshot and FreeBSD-2.2.7 not building"
}
] |
[
{
"msg_contents": "I've been tweaking the current sources to make them compile on HPUX 9\nagain (right now they don't, for lack of vsnprintf() in libc). All\nwent well until I tried to get rid of gcc's compile warnings about\nlack of a prototype for snprintf & vsnprintf. I put\n\n#ifndef HAVE_SNPRINTF\nextern int snprintf(char *str, size_t count, const char *fmt, ...);\n#endif\n\n#ifndef HAVE_VSNPRINTF\nextern int vsnprintf(char *str, size_t count, const char *fmt, va_list args);\n#endif\n\ninto include/c.h (they can't readily go into config.h because config.h\ndoesn't include any system header that defines size_t or va_list).\n\nThat almost works, but there are a few files that blow up because\nthey include c.h without having included config.h beforehand.\n\nNow I maintain that every file that uses *any* of the Postgres\ninclude files ought to include config.h. Not doing so is a recipe\nfor trouble as we become more and more dependent on autoconf.\nSo I propose changing these files to include postgres.h instead of\njust including c.h directly:\n\ncontrib/fulltextindex/fti.c\nsrc/backend/port/dynloader/aix.c\nsrc/backend/port/dynloader/hpux.c\nsrc/backend/port/dynloader/ultrix4.c\nsrc/backend/storage/buffer/s_lock.c\nsrc/backend/utils/error/format.c\nsrc/backend/utils/mmgr/oset.c\nsrc/backend/utils/mmgr/palloc.c\nsrc/include/libpq/pqcomm.h\nsrc/interfaces/libpq/pqsignal.h\n\nAny objections?\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 03 Jan 1999 19:55:12 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": true,
"msg_subject": "include-file unhappiness"
},
{
"msg_contents": "> I've been tweaking the current sources to make them compile on HPUX 9\n> again (right now they don't, for lack of vsnprintf() in libc). All\n> went well until I tried to get rid of gcc's compile warnings about\n> lack of a prototype for snprintf & vsnprintf. I put\n> \n> #ifndef HAVE_SNPRINTF\n> extern int snprintf(char *str, size_t count, const char *fmt, ...);\n> #endif\n> \n> #ifndef HAVE_VSNPRINTF\n> extern int vsnprintf(char *str, size_t count, const char *fmt, va_list args);\n> #endif\n> \n> into include/c.h (they can't readily go into config.h because config.h\n> doesn't include any system header that defines size_t or va_list).\n> \n> That almost works, but there are a few files that blow up because\n> they include c.h without having included config.h beforehand.\n> \n> Now I maintain that every file that uses *any* of the Postgres\n> include files ought to include config.h. Not doing so is a recipe\n> for trouble as we become more and more dependent on autoconf.\n> So I propose changing these files to include postgres.h instead of\n> just including c.h directly:\n> \n> contrib/fulltextindex/fti.c\n> src/backend/port/dynloader/aix.c\n> src/backend/port/dynloader/hpux.c\n> src/backend/port/dynloader/ultrix4.c\n> src/backend/storage/buffer/s_lock.c\n> src/backend/utils/error/format.c\n> src/backend/utils/mmgr/oset.c\n> src/backend/utils/mmgr/palloc.c\n> src/include/libpq/pqcomm.h\n> src/interfaces/libpq/pqsignal.h\n\nGo ahead. Good reasoning for doing 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": "Sun, 3 Jan 1999 22:28:05 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] include-file unhappiness"
}
] |
[
{
"msg_contents": "\tI am having problems reloading my 6.3 dump inot 6.4.1 Heres what I am\n\tseeing:\n\n\\connect template1 postgres\nconnecting to new database: template1 as user: postgres\ncreate database stan;\nQUERY: create database stan;\nCREATEDB\n\\connect stan postgres\nconnecting to new database: stan as user: postgres\n\\connect - stan\nconnecting as new user: stan\nFATAL 1: SetUserId: user 'stan' is not in 'pg_shadow'\n\nCould not connect to new database. exiting\n\n\tWhat am I doing wrong here?\n\n-- \nStan Brown [email protected] 843-745-3154\nWestvaco\nCharleston SC.\n-- \nWindows 98: n.\n\tuseless extension to a minor patch release for 32-bit extensions and\n\ta graphical shell for a 16-bit patch to an 8-bit operating system\n\toriginally coded for a 4-bit microprocessor, written by a 2-bit \n\tcompany that can't stand for 1 bit of competition.\n-\n(c) 1999 Stan Brown. Redistribution via the Microsoft Network is prohibited.\n",
"msg_date": "Sun, 3 Jan 1999 22:49:11 -0500 (EST)",
"msg_from": "\"Stan Brown\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Reloading a 6.3db into 6.4.1"
},
{
"msg_contents": "> \tI am having problems reloading my 6.3 dump inot 6.4.1 Heres what I am\n> \tseeing:\n> \n> \\connect template1 postgres\n> connecting to new database: template1 as user: postgres\n> create database stan;\n> QUERY: create database stan;\n> CREATEDB\n> \\connect stan postgres\n> connecting to new database: stan as user: postgres\n> \\connect - stan\n> connecting as new user: stan\n> FATAL 1: SetUserId: user 'stan' is not in 'pg_shadow'\n> \n> Could not connect to new database. exiting\n> \n> \tWhat am I doing wrong here?\n\nDid you use pg_dump preseving the users, but did not user pg_dumpall,\nwhich creates the users.\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, 3 Jan 1999 23:44:13 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Reloading a 6.3db into 6.4.1t"
}
] |
[
{
"msg_contents": "\nignore...\n\n",
"msg_date": "Mon, 4 Jan 1999 08:01:19 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\n",
"msg_date": "Mon, 4 Jan 1999 08:05:51 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\n",
"msg_date": "Mon, 4 Jan 1999 08:06:43 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\n",
"msg_date": "Mon, 4 Jan 1999 08:07:13 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\nignore ...\n\n",
"msg_date": "Mon, 4 Jan 1999 08:10:39 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\nignore ...\n\n",
"msg_date": "Mon, 4 Jan 1999 08:11:59 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\nignore \n\n",
"msg_date": "Mon, 4 Jan 1999 08:12:39 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\nignore ...\n\n",
"msg_date": "Mon, 4 Jan 1999 08:14:43 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\nignore ..\n\n",
"msg_date": "Mon, 4 Jan 1999 08:17:10 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\nif this works, the lists are working again...\n\n",
"msg_date": "Mon, 4 Jan 1999 08:23:20 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\nignore ...\n\n",
"msg_date": "Mon, 4 Jan 1999 08:25:01 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\nignore ...\n\n",
"msg_date": "Mon, 4 Jan 1999 08:29:48 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\n",
"msg_date": "Mon, 4 Jan 1999 08:30:34 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
}
] |
[
{
"msg_contents": "\n\nnow this is starting to annoy me...fix one \"bug\" and the next creeps in...\n\n",
"msg_date": "Mon, 4 Jan 1999 08:31:37 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "a test"
},
{
"msg_contents": "On Mon, 4 Jan 1999, Jan Wieck wrote:\n\n> > \n> > \n> > \n> > now this is starting to annoy me...fix one \"bug\" and the next creeps in...\n> \n> Pardon?\n\nI upgraded part of perl last night (~23:00ADT) and found out early this\nmorning that I had broken majordomo inadvertantly :(\n\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Mon, 4 Jan 1999 09:55:57 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] a test"
},
{
"msg_contents": "> \n> \n> \n> now this is starting to annoy me...fix one \"bug\" and the next creeps in...\n\n Pardon?\n\n> \n> \n> \n\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, 4 Jan 1999 14:56:04 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] a test"
},
{
"msg_contents": "> \n> \n> \n> now this is starting to annoy me...fix one \"bug\" and the next creeps in...\n\n Pardon?\n\n> \n> \n> \n\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, 4 Jan 1999 14:56:04 +0100 (MET)",
"msg_from": "[email protected]",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] a test"
},
{
"msg_contents": "On Mon, 4 Jan 1999, Jan Wieck wrote:\n\n> > \n> > On Mon, 4 Jan 1999, Jan Wieck wrote:\n> > \n> > > > \n> > > > \n> > > > \n> > > > now this is starting to annoy me...fix one \"bug\" and the next creeps in...\n> > > \n> > > Pardon?\n> > \n> > I upgraded part of perl last night (~23:00ADT) and found out early this\n> > morning that I had broken majordomo inadvertantly :(\n> \n> Ah - and I thought there are delivery problems on the debis\n> corporate network (just once again).\n> \n> Maybe I should remail some of the messages from today AM.\n\n\tThat would be prudent :)\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Mon, 4 Jan 1999 10:22:48 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] a test"
},
{
"msg_contents": "> \n> On Mon, 4 Jan 1999, Jan Wieck wrote:\n> \n> > > \n> > > \n> > > \n> > > now this is starting to annoy me...fix one \"bug\" and the next creeps in...\n> > \n> > Pardon?\n> \n> I upgraded part of perl last night (~23:00ADT) and found out early this\n> morning that I had broken majordomo inadvertantly :(\n\n Ah - and I thought there are delivery problems on the debis\n corporate network (just once again).\n\n Maybe I should remail some of the messages from today AM.\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, 4 Jan 1999 15:23:55 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] a test"
}
] |
[
{
"msg_contents": "Hi all,\n\nThe following query abort the backend:\n\nSELECT customer.name, reservation.rno, hotel.name\n FROM customer, reservation, hotel\n WHERE customer.name = 'Porter'\n AND customer.cno = reservation.cno\n AND reservation.hno = hotel.hno;\nERROR: cannot create temp_39296\n\n- the first time it gives me this message:\n ERROR: cannot create temp_NUMBER\n\n- the second time it gives me this message:\n\npqReadData() -- backend closed the channel unexpectedly.\n This probably means the backend terminated abnormally before or while\nprocessing the request.\nWe have lost the connection to the backend, so further processing is impossible.\nTerminating.\n\nHere the tables structure and data:\n\nTable = customer\n+----------------------------------+----------------------------------+-------+\n| Field | Type | Length|\n+----------------------------------+----------------------------------+-------+\n| cno | int4 | 4 |\n| title | char() | 8 |\n| state | char() | 4 |\n| zip | int2 | 2 |\n| city | varchar() | 0 |\n| firstname | char() | 7 |\n| name | char() | 8 |\n| account | float8 | 8 |\n+----------------------------------+----------------------------------+-------+\nselect * from customer;\n cno|title |state| zip|city |firstname|name |account\n----+--------+-----+----+--------+---------+--------+-------\n3391|Mr |NY |9001|New York|Joseph |Porter | 50\n3395|Mr |CH |9001|Chicago |Albert |Keen | 10\n3396|Mrs |DA |9024|Dallas |Joe |Nook | 60\n(3 rows)\n\nTable = reservation\n+----------------------------------+----------------------------------+-------+\n| Field | Type | Length|\n+----------------------------------+----------------------------------+-------+\n| cno | int4 | 4 |\n| rno | int4 | 4 |\n| hno | int4 | 4 |\n| arrival | date | 4 |\n| departure | date | 4 |\n+----------------------------------+----------------------------------+-------+\nselect * from reservation;\n cno|rno|hno| arrival| departure\n----+---+---+----------+----------\n3391| 2| 1|1995-12-04|1995-12-31\n3395| 3| 2|1997-11-04|1997-12-31\n3396| 4| 3|1996-12-14|1996-12-31\n(3 rows)\n\nTable = hotel\n+----------------------------------+----------------------------------+-------+\n| Field | Type | Length|\n+----------------------------------+----------------------------------+-------+\n| hno | int4 | 4 |\n| name | varchar() | 0 |\n| city | varchar() | 0 |\n+----------------------------------+----------------------------------+-------+\nselect * from hotel;\nhno|name |city\n---+--------+--------\n 1|Sheraton|Chicago\n 2|Hilton |New York\n 3|Ritz |Dallas\n(3 rows)\n \nAny ideas?\n\n-Jose'-\n\n\n",
"msg_date": "Mon, 4 Jan 1999 14:46:37 +0100",
"msg_from": "Sferacarta Software <[email protected]>",
"msg_from_op": true,
"msg_subject": "too many ANDs ?"
}
] |
[
{
"msg_contents": "Hi all,\n\nSeems there's a bug using GROUP BY not on select list on subselects:\n\nSELECT city, AVG(account)\n FROM customer\n GROUP BY city\n HAVING AVG(account) >= ALL\n (SELECT city,AVG(account)\n FROM customer group by city\n );\nERROR: parser: Subselect has too many or too few fields.\n\nInstead the following does work:\n\nSELECT city, AVG(account)\n FROM customer\n GROUP BY city;\ncity |avg\n--------+---\nChicago | 10\nDallas | 60\nNew York| 50\n(3 rows)\n\nSELECT AVG(account)\n FROM customer\n GROUP BY city;\navg\n---\n 10\n 60\n 50\n(3 rows)\n\n-Jose'-\n\n\n",
"msg_date": "Mon, 4 Jan 1999 14:53:35 +0100",
"msg_from": "Sferacarta Software <[email protected]>",
"msg_from_op": true,
"msg_subject": "bug on Subselect with GROUP BY clause"
}
] |
[
{
"msg_contents": "Hi all,\n\nI supposed the int4 range was from -2147483648 to +2147483647 but I\nhave the following message trying to update an int4 column:\n\nprova=> update t set i=-2147483648;\nNOTICE: Integer input '2147483648' is out of range; promoted to float\nUPDATE 1\n\nThe column was updated any way, but I don't understand the message.\n\nprova=> update t set i=-2147483647;\nUPDATE 1\n\nPS: I'm using v6.4 on Linux.\n\n-Jose'-\n\n\n",
"msg_date": "Mon, 4 Jan 1999 15:12:58 +0100",
"msg_from": "Sferacarta Software <[email protected]>",
"msg_from_op": true,
"msg_subject": "int4 range"
}
] |
[
{
"msg_contents": "Hi all,\n\nI have this funny result when I tried to multiply a money by a float:\n\nselect '($21,474,836.48)'::money / '1.0'::float8;\n?column? \n----------------\n($./,,),,(-*.,()\n(1 row)\n\nselect '$21,474,836.47'::money * '2.0'::float8;\n?column? \n----------------\n($./,,),,(-*.,()\n(1 row)\n\n--If I use integer instead of float it gives me an expected overflow\n--result:\n\nselect '$21,474,836.47'::money * '2'::int4;\n?column?\n--------\n($0.02)\n(1 row)\n\nI'm using v6.4 on Linux.\n\n-Jose'-\n\n\n",
"msg_date": "Mon, 4 Jan 1999 15:14:30 +0100",
"msg_from": "Sferacarta Software <[email protected]>",
"msg_from_op": true,
"msg_subject": "money bug"
}
] |
[
{
"msg_contents": "All,\n\tI've just read in DBMS magazine that Oracle, Informix, etc...\nare replacing there proprietary Procedural Language's (PL/SQL) with\nJVMs (JAVA virtual machine). Apparently according to the database\nvendors, this will obviate the need for middle ware products and to\nmake CORBA much more intuitive.\n\nMy personal opinion is this is yet another mechanism to slow down\nthe database engine. I don't mind using JAVA in the database engine\nas a procedural language, but I think there is still great merit in\nbreaking\nthis out as a separate process (Application Server).\n\nD.\n",
"msg_date": "Mon, 4 Jan 1999 09:26:08 -0500 ",
"msg_from": "Dan Gowin <[email protected]>",
"msg_from_op": true,
"msg_subject": "SQLJ"
},
{
"msg_contents": "Dan Gowin wrote:\n> \n> I've just read in DBMS magazine that Oracle, Informix, etc...\n> are replacing there proprietary Procedural Language's (PL/SQL) with\n> JVMs (JAVA virtual machine). Apparently according to the database\n> vendors, this will obviate the need for middle ware products and to\n> make CORBA much more intuitive.\n\nThis was a problem that I also wanted to talk about.\n \n> My personal opinion is this is yet another mechanism to slow down\n> the database engine. I don't mind using JAVA in the database engine\n> as a procedural language, but I think there is still great merit in\n> breaking\n> this out as a separate process (Application Server).\n\nI don't think that it would be slower than PlTcl for example.\n\nTalking about speed:\n\nI made some tests with some procedures written in PlPgsql and PlTcl!\nI understood also that PlPgsql is compiling the procedure body into a\nsort of bytecodes and the execution is somewhat faster.\nIn PlTcl, I feel (I didn't poke the sources) that a PlTcl procedure body\nis interpreted each time at execution time. For this reason, it doesn't\nbeneffit from Tcl 8.x compiler, making it slowly than PlPgsql. Am I\nright ?\n\nThe first question is : Is Tcl 8.x able to precompile a script and\nreturn the bytecodes that should be kept into the database and execute\nthem later?\n\n>From this point of view, I think that compiling a Java language\nprocedure and storing the resulting .class might be at execution time as\nfaster as PlPgsql. And , I think, perfectly possible.\n\nRight now, I am working at a big project using :\n1. Tcl/Tk as thin client\n2. Messaging server written in Java as a middle-tier with JDBC\n3. PostgreSQL RDBMS\n\nIt would be perfect from me if PostgreSQL will allow procedures written\nin Java language.\n\n-- \nConstantin Teodorescu\nFLEX Consulting Braila, ROMANIA\n",
"msg_date": "Mon, 04 Jan 1999 16:45:32 +0200",
"msg_from": "Constantin Teodorescu <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] SQLJ"
},
{
"msg_contents": "Constantin Teodorescu wrote:\n\n>\n> Dan Gowin wrote:\n> [...]\n>\n> > My personal opinion is this is yet another mechanism to slow down\n> > the database engine. I don't mind using JAVA in the database engine\n> > as a procedural language, but I think there is still great merit in\n> > breaking\n> > this out as a separate process (Application Server).\n>\n> I don't think that it would be slower than PlTcl for example.\n>\n> Talking about speed:\n>\n> I made some tests with some procedures written in PlPgsql and PlTcl!\n> I understood also that PlPgsql is compiling the procedure body into a\n> sort of bytecodes and the execution is somewhat faster.\n> In PlTcl, I feel (I didn't poke the sources) that a PlTcl procedure body\n> is interpreted each time at execution time. For this reason, it doesn't\n> beneffit from Tcl 8.x compiler, making it slowly than PlPgsql. Am I\n> right ?\n>\n> The first question is : Is Tcl 8.x able to precompile a script and\n> return the bytecodes that should be kept into the database and execute\n> them later?\n\n Speaking about the versions of PL/pgSQL and PL/Tcl shipped\n with Postgres v6.4 (and bugfix releases).\n\n PL/pgSQL does precompile the source of the function body the\n first time, a function is called in a backend. For most of\n the statements, even if they are a simple assignment like\n \"var := 0;\", a prepared SPI statement get's created. This\n makes PL/pgSQL damned slow and I plan to optimize that by\n handling simple expressions that have only one targetlist\n entry directly instead of using the SPI manager.\n\n PL/Tcl uses the precompiler of Tcl8.0. Inside the PL handler\n module, a safe Tcl interpreter is created at the first of all\n calls to one function written in PL/Tcl. All functions called\n are somewhat \"sourced\" into the safe interpreter. Thus, on\n multiple invocations of one and the same function, the\n Interpreter does reuse the precompiled bytecode it is holding\n inside for the procedures. But the programmer must arrange\n for usage of prepared and saved SPI execution plans for data\n access. The Tcl command \"spi_exec\", available in PL/Tcl,\n really calls SPI_exec - resulting in a parse-rewrite-\n optimize-execute sequence for the statement every time.\n Using spi_prepare (only on first call of the PL/Tcl function)\n and spi_execp instead will gain much performance increase!\n\n Another area that should be optimized sometimes in PL/Tcl is\n using the object interface which came with Tcl8.0. Up to now,\n the PL/Tcl handler uses the string level C-API only.\n\n I don't think it would be good to remember the generated\n bytecode of Tcl itself. This would mean to muck around with\n internals of Tcl objects. A clear violation of the OOP\n paradigm.\n\n Both languages have performance problems in scenarios like a\n WEB environment. By default, any WEB access creates a new\n Postgres backend, that has to compile the functions used.\n\n>\n> >From this point of view, I think that compiling a Java language\n> procedure and storing the resulting .class might be at execution time as\n> faster as PlPgsql. And , I think, perfectly possible.\n\n Should (and could) use the same interfacing mechanism as is.\n It is just another loadable procedural language handler. And\n the design changes we did in the fmgr for v6.3.2 where to\n enable these language handlers.\n\n>\n> Right now, I am working at a big project using :\n> 1. Tcl/Tk as thin client\n> 2. Messaging server written in Java as a middle-tier with JDBC\n> 3. PostgreSQL RDBMS\n>\n> It would be perfect from me if PostgreSQL will allow procedures written\n> in Java language.\n\n PL/Tcl and PL/pgSQL are the only two languages available,\n because noone else than me did anything in that corner up to\n now. You're welcome :-).\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, 4 Jan 1999 17:29:03 +0100 (MET)",
"msg_from": "[email protected] (Jan Wieck)",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] SQLJ"
},
{
"msg_contents": "Jan Wieck wrote:\n> \n> PL/Tcl and PL/pgSQL are the only two languages available,\n> because noone else than me did anything in that corner up to\n> now. You're welcome :-).\n\n:-)\n\nYeap! Easy to say! Hard to do!\n\nI'm not a \"core\" programmer. I am rather new to Java (only 2 weeks), and\nI don't know anything about inside PostgreSQL structures.\nWe would need someone with much more expertise in \"C\", \"C++\", Java and\nPostgreSQL than me.\n\n-- \nConstantin Teodorescu\nFLEX Consulting Braila, ROMANIA\n",
"msg_date": "Mon, 04 Jan 1999 18:30:51 +0200",
"msg_from": "Constantin Teodorescu <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] SQLJ"
},
{
"msg_contents": ">Jan Wieck wrote:\n>> \n>> PL/Tcl and PL/pgSQL are the only two languages available,\n>> because noone else than me did anything in that corner up to\n>> now. You're welcome :-).\n>\n>:-)\n>\n>Yeap! Easy to say! Hard to do!\n>\n>I'm not a \"core\" programmer. I am rather new to Java (only 2 weeks), and\n>I don't know anything about inside PostgreSQL structures.\n>We would need someone with much more expertise in \"C\", \"C++\", Java and\n>PostgreSQL than me.\n\nI've been interested in this topic too. My idea is having JVM as a\nseparate process from the backend. Maybe JServ(Apache's servlet\nextension package) is a good starting point. Also it would be nice if\nthe backend could use some functionalities of Java, such as encoding\nconversions.\n\nI am by no means a Java guru, thus this would be a great challenge for\nme:-)\n--\nTatsuo Ishii\n",
"msg_date": "Wed, 06 Jan 1999 11:07:53 +0900",
"msg_from": "Tatsuo Ishii <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] SQLJ "
},
{
"msg_contents": "On Wed, 6 Jan 1999, Tatsuo Ishii wrote:\n\n> >Jan Wieck wrote:\n> >> \n> >> PL/Tcl and PL/pgSQL are the only two languages available,\n> >> because noone else than me did anything in that corner up to\n> >> now. You're welcome :-).\n> >\n> >:-)\n> >\n> >Yeap! Easy to say! Hard to do!\n> >\n> >I'm not a \"core\" programmer. I am rather new to Java (only 2 weeks), and\n> >I don't know anything about inside PostgreSQL structures.\n> >We would need someone with much more expertise in \"C\", \"C++\", Java and\n> >PostgreSQL than me.\n> \n> I've been interested in this topic too. My idea is having JVM as a\n> separate process from the backend. Maybe JServ(Apache's servlet\n> extension package) is a good starting point. Also it would be nice if\n> the backend could use some functionalities of Java, such as encoding\n> conversions.\n> \n> I am by no means a Java guru, thus this would be a great challenge for\n> me:-)\n\nI havent touched the native side of Java (have stuck with just pure java\nso far, and will probably stay that way), but from what I understand on\nhow JNI works:\n\n1. If say PL/Java was implemented _AND_ enabled, then when the postmaster\nis first started, you can start a JVM as a separate process at that time.\nWhile waiting for connections, this JVM would be sitting dormant.\n\n2. When a connection is made to the backend, a Thread (Java not native) is\nstarted in the JVM, which creates any objects needed by the PL/Java\ncode. (For performance sake, I'd think this would be done by using a\nparameter set in the database - that way db's without the need of PL/Java\nsupport wouldn't be affected.\n\n3. When a call to a PL/Java procedure is made, then a fresh Thread is made\nin the JVM, and calls the method to handle the procedure.\n\nObviously this scheme would need the PL/Java functions pre-compiled using\nthe JDK, and the class files (or better still as a single jar file) being\nintroduced to the database.\n\nI'd thought that running EcmaScript (what is commonly known as JavaScript)\nwould be too slow.\n\n>From my tests using the latest JVM, at least under Windows, it's\nperformance has improved tremendously. It's still slow in starting up, but\nonce running it's pretty good. Also, when running without a GUI, it's\nbetter still.\n\nAnother plus would be CORBA. If, using JNI we had a way of getting at SPI\nfunctions, then you could create objects in Java that fed directly into\nthe backend. It would be limited (as only a small part of CORBA 2.0 is in\nJDK 1.2), but it would be a start.\n\n-- \n Peter T Mount [email protected]\n Main Homepage: http://www.retep.org.uk\nPostgreSQL JDBC Faq: http://www.retep.org.uk/postgres\n Java PDF Generator: http://www.retep.org.uk/pdf\n\n",
"msg_date": "Wed, 6 Jan 1999 23:20:11 +0000 (GMT)",
"msg_from": "Peter T Mount <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] SQLJ "
}
] |
[
{
"msg_contents": "Hello\n\nI tried create a small logging utility for postgresql (6.4, i use it on\nredhat-linux-5.2)\nI meand with triggers 'n SPI all database change can be hooked.\nBut this have small problem: in an aborted transaction the triggers run on\nchanges, but NOT run with the undos.\n4xmpl.: \n --\nmy=> Begin;\nmy=> insert into test values ( 'now',0,'Text' );\nNOTICE: trigf(): Ok, Insert event sent to log server.\nINSERT [5234]\nmy=> -- Succesfully logged. BUT:\nmy-> rollback;\nROLLBACK\nmy=> -- Not logged ;(\n --\nCan postgres call trigger beetween rollbacks too? (Nawp: all changes must be\nsuccesfully undoned, this can not depending on triggers)\n\nOk, that is not works, but realy want to logging to go outside the database.\n\n... can anybody give a solution to this?\n\n--\n NeKo@(kva.hu|Kornel.szif.hu) the servant of Crash\n\n\n",
"msg_date": "Mon, 4 Jan 1999 17:06:15 +0100 (NFT)",
"msg_from": "\"Vazsonyi Peter[ke]\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Q: Log, via triggers"
}
] |
[
{
"msg_contents": "\nHi everybody.\nI tried to use pg_options for monitoring sql-query.\npg_options: all=1\n/ets/syslog.conf:\n local5.* /var/log/postgres\nfor postgres start i use script src/contrib/linux/postgres.init.csh:\n(postmaster -d 2 2>&1 | logger -p local5.notice) &\n\nIn /var/log/postgres i have:\nJan 4 17:13:09 SantaMariya logger: StartTransactionCommand\n \n\nHow can i give syslog information in form:\n#timestamp #pid #message\n\nFor example:\nJan 4 17:13:09.173 [29271] StartTransactionCommand\n ^^^ ^^^^^\n I need this \n information.\n\nThank you for your answer.\n\n\n\n",
"msg_date": "Mon, 4 Jan 1999 18:14:57 +0200 (EET)",
"msg_from": "Blashko Alexander <[email protected]>",
"msg_from_op": true,
"msg_subject": "pg_options ?"
},
{
"msg_contents": "> \n> \n> Hi everybody.\n> I tried to use pg_options for monitoring sql-query.\n> pg_options: all=1\n> /ets/syslog.conf:\n> local5.* /var/log/postgres\n> for postgres start i use script src/contrib/linux/postgres.init.csh:\n> (postmaster -d 2 2>&1 | logger -p local5.notice) &\n> \n> In /var/log/postgres i have:\n> Jan 4 17:13:09 SantaMariya logger: StartTransactionCommand\n> \n> \n> How can i give syslog information in form:\n> #timestamp #pid #message\n> \n> For example:\n> Jan 4 17:13:09.173 [29271] StartTransactionCommand\n> ^^^ ^^^^^\n> I need this \n> information.\n> \n> Thank you for your answer.\n\nThere is a better way to do it. Compile postgres with -DUSE_SYSLOG, add\nsyslog=1 or syslog=2 into pg_options and it will output all elog messages\nto the syslog facility with the pid. See trace.c for details.\nBut you won't be able to see the milliseconds fields because syslog\ndoesn't handle this information.\n\n-- \nMassimo Dal Zotto\n\n+----------------------------------------------------------------------+\n| Massimo Dal Zotto email: [email protected] |\n| Via Marconi, 141 phone: ++39-0461534251 |\n| 38057 Pergine Valsugana (TN) www: http://www.cs.unitn.it/~dz/ |\n| Italy pgp: finger [email protected] |\n+----------------------------------------------------------------------+\n",
"msg_date": "Mon, 4 Jan 1999 18:36:57 +0100 (MET)",
"msg_from": "Massimo Dal Zotto <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: pg_options ?"
},
{
"msg_contents": "\n\nOn Mon, 4 Jan 1999, Massimo Dal Zotto wrote:\n\n> > \n> > \n> > Hi everybody.\n> > I tried to use pg_options for monitoring sql-query.\n> > pg_options: all=1\n> > /ets/syslog.conf:\n> > local5.* /var/log/postgres\n> > for postgres start i use script src/contrib/linux/postgres.init.csh:\n> > (postmaster -d 2 2>&1 | logger -p local5.notice) &\n> > \n> > In /var/log/postgres i have:\n> > Jan 4 17:13:09 SantaMariya logger: StartTransactionCommand\n> > \n> > \n> > How can i give syslog information in form:\n> > #timestamp #pid #message\n> > \n> > For example:\n> > Jan 4 17:13:09.173 [29271] StartTransactionCommand\n> > ^^^ ^^^^^\n> > I need this \n> > information.\n> > \n> > Thank you for your answer.\n> \n> There is a better way to do it. Compile postgres with -DUSE_SYSLOG, add\n> syslog=1 or syslog=2 into pg_options and it will output all elog messages\n> to the syslog facility with the pid. See trace.c for details.\n> But you won't be able to see the milliseconds fields because syslog\n> doesn't handle this information.\n\nThank you for your answer.\nBut in doc (pg-options.html) you got example:\n#timestamp #pid #message\n980127.17:52:14.173 [29271] StartTransactionCommand\n980127.17:52:14.174 [29271] ProcessUtility: drop table t;\n980127.17:52:14.186 [29271] SIIncNumEntries: table is 70% full\n\nHow can i get this information with milliseconds without using syslog ?\n \nThank you for your answer.\n> \n> -- \n> Massimo Dal Zotto\n> \n> +----------------------------------------------------------------------+\n> | Massimo Dal Zotto email: [email protected] |\n> | Via Marconi, 141 phone: ++39-0461534251 |\n> | 38057 Pergine Valsugana (TN) www: http://www.cs.unitn.it/~dz/ |\n> | Italy pgp: finger [email protected] |\n> +----------------------------------------------------------------------+\n> \n\n",
"msg_date": "Tue, 5 Jan 1999 10:10:07 +0200 (EET)",
"msg_from": "Blashko Alexander <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: pg_options ?"
}
] |
[
{
"msg_contents": "I see a 6.4.2 tarball on ftp.postgresql.org.\nIs it time to announce the release on the web site?\nWe've gone for a week with *no* release available.\n",
"msg_date": "04 Jan 1999 10:55:12 -0600",
"msg_from": "Hal Snyder <[email protected]>",
"msg_from_op": true,
"msg_subject": "6.4.2?"
},
{
"msg_contents": "On 4 Jan 1999, Hal Snyder wrote:\n\n> I see a 6.4.2 tarball on ftp.postgresql.org.\n> Is it time to announce the release on the web site?\n> We've gone for a week with *no* release available.\n\n\tYa, I announced it to postgresql-annoucne lastnight, after tar'ng\nit up ...\n\n\nMarc G. Fournier \nSystems Administrator @ hub.org \nprimary: [email protected] secondary: scrappy@{freebsd|postgresql}.org \n\n",
"msg_date": "Mon, 4 Jan 1999 13:16:03 -0400 (AST)",
"msg_from": "The Hermit Hacker <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.2?"
},
{
"msg_contents": "> On 4 Jan 1999, Hal Snyder wrote:\n> \n> > I see a 6.4.2 tarball on ftp.postgresql.org.\n> > Is it time to announce the release on the web site?\n> > We've gone for a week with *no* release available.\n> \n> \tYa, I announced it to postgresql-annoucne lastnight, after tar'ng\n> it up ...\n\nYou better do it again. I never saw it. Also, post to the general\nlist.\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, 4 Jan 1999 12:40:04 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] 6.4.2?"
},
{
"msg_contents": "The Hermit Hacker <[email protected]> writes:\n\n> \tYa, I announced it to postgresql-annoucne lastnight, after tar'ng\n> it up ...\n\nAck. Will do a quick adjustment now, fine tune tonight when there's\nmore time.\n",
"msg_date": "04 Jan 1999 11:58:02 -0600",
"msg_from": "Hal Snyder <[email protected]>",
"msg_from_op": true,
"msg_subject": "Re: [HACKERS] 6.4.2?"
}
] |
[
{
"msg_contents": "I'm unhappy to say that but ...\n\nstill there are some problems compiling 6.4.2 on Linux for libpgtcl and\nlibplpgsql libraries.\n\nConfiguration : RedHat Linux 5.1 i386 on Pentium\n\nAfter untarr-ing the final 6.4.2 distribution and configuring it with :\n\n./configure --with-tcl\n\n1. For libpgtcl, I had to manually add -lcrypt to SHLIB_LINK line\nMakefile in src/interfaces/libpgtcl\n2. For libplpgsql, I had to manually ad -lc to SHLIB_LINK line Makefile\nin src/pl/pgsql/src\n\n\nFor example, the original line for plpgsql was :\nSHLIB_LINK=-L($LIBPQDIR) -lpq\n\nCompiled that way, using plpgsql functions give the error \"cannot stat\nin plpgsql.so\" !\n\nAfter adding -lc as shown above, my functions worked corectly.\n\nBest regards,\n-- \nConstantin Teodorescu\nFLEX Consulting Braila, ROMANIA\n",
"msg_date": "Mon, 04 Jan 1999 19:47:19 +0200",
"msg_from": "Constantin Teodorescu <[email protected]>",
"msg_from_op": true,
"msg_subject": "Problems building some libraries in 6.4.2"
},
{
"msg_contents": "Constantin Teodorescu <[email protected]> writes:\n> still there are some problems compiling 6.4.2 on Linux for libpgtcl and\n> libplpgsql libraries.\n> Configuration : RedHat Linux 5.1 i386 on Pentium\n> 1. For libpgtcl, I had to manually add -lcrypt to SHLIB_LINK line\n> Makefile in src/interfaces/libpgtcl\n\nI have fixed this. It's ugly; libraries that call libpq should not have\nto know what libpq calls. But it seems we have no choice, for now.\nPerhaps GNU libtool will allow us to get rid of the explicit dependency.\n(I've decided to wait for libtool 1.3 before trying to use it for\nPostgres, however.)\n\n> 2. For libplpgsql, I had to manually ad -lc to SHLIB_LINK line Makefile\n> in src/pl/pgsql/src\n\nI don't understand this. If you are on a Linux box then\nsrc/Makefile.shlib should add -lc to SHLIB_LINK for you.\nIs PORTNAME not getting set to \"linux\"?\n\nWhat I *would* expect that plpgsql needs in SHLIB_LINK is -lcrypt,\nfor the same reason that libpgtcl needs it. I will change the makefile\naccordingly.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Sun, 07 Feb 1999 15:35:34 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Problems building some libraries in 6.4.2 "
}
] |
[
{
"msg_contents": "\tAT Tom Lanes sugestion I upgraded to 6.4.2. While this was undoubtably\n\ta good die, it did not solve the basic problem.\n\n\tI am runing as the postgres superuser, and trying to inport the data\n\tfrom my 6.3 databse inot my nice new 6.4.2 one. I am geting errors\n\talmost as soon as the first non-postgres user is encuntered. Here is\n\tthe tail end of the atempt:\n\n\\connect template1 postgres\nconnecting to new database: template1 as user: postgres\ncreate database stan;\nQUERY: create database stan;\nCREATEDB\n\\connect stan postgres\nconnecting to new database: stan as user: postgres\n\\connect - stan\nconnecting as new user: stan\nFATAL 1: SetUserId: user 'stan' is not in 'pg_shadow'\n\nCould not connect to new database. exiting\n\n\tI am using the example from the INSTALL document to import this data,\n\tthat is:\n\n\tpsql -e template1 < db.out\n\n\tI can connect to this db using psql, and I have run creatdb. \n\tWhat is this pg_shadow & how do I fix it?\n\n\tOr do I need to create the varios database first before doing the\n\timport?\n\n\tThanks.\n\n-- \nStan Brown [email protected] 843-745-3154\nWestvaco\nCharleston SC.\n-- \nWindows 98: n.\n\tuseless extension to a minor patch release for 32-bit extensions and\n\ta graphical shell for a 16-bit patch to an 8-bit operating system\n\toriginally coded for a 4-bit microprocessor, written by a 2-bit \n\tcompany that can't stand for 1 bit of competition.\n-\n(c) 1999 Stan Brown. Redistribution via the Microsoft Network is prohibited.\n",
"msg_date": "Mon, 4 Jan 1999 21:33:42 -0500 (EST)",
"msg_from": "\"Stan Brown\" <[email protected]>",
"msg_from_op": true,
"msg_subject": "Still having trouble importing 6.3 DB "
},
{
"msg_contents": "\"Stan Brown\" <[email protected]> writes:\n> \tI am runing as the postgres superuser, and trying to inport the data\n> \tfrom my 6.3 databse inot my nice new 6.4.2 one. I am geting errors\n> \talmost as soon as the first non-postgres user is encuntered. \n> \\connect - stan\n> connecting as new user: stan\n> FATAL 1: SetUserId: user 'stan' is not in 'pg_shadow'\n> \tWhat is this pg_shadow & how do I fix it?\n\nAh, I see what's going on. You're getting burnt by not having upgraded\nto 6.3.1 or 6.3.2 before moving to 6.4. The pg_user table was changed\nat 6.3.1, and pg_dumpall was modified accordingly --- but the 6.3\nversion of pg_dumpall produces a script that will not successfully load\nthe user definitions into 6.3.1 or later.\n\nI think you can make it work by just changing \"copy pg_user from stdin;\"\nto \"copy pg_shadow from stdin;\" a couple lines down from the top of\nthe dump script.\n\nMemo to hackers: we need a better way of coping with cross-version\nchanges that affect dump scripts...\n\n\t\t\tregards, tom lane\n",
"msg_date": "Mon, 04 Jan 1999 22:18:32 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Still having trouble importing 6.3 DB"
},
{
"msg_contents": "> Ah, I see what's going on. You're getting burnt by not having upgraded\n> to 6.3.1 or 6.3.2 before moving to 6.4. The pg_user table was changed\n> at 6.3.1, and pg_dumpall was modified accordingly --- but the 6.3\n> version of pg_dumpall produces a script that will not successfully load\n> the user definitions into 6.3.1 or later.\n\nOoh, good catch. You are correct.\n\n> \n> I think you can make it work by just changing \"copy pg_user from stdin;\"\n> to \"copy pg_shadow from stdin;\" a couple lines down from the top of\n> the dump script.\n> \n> Memo to hackers: we need a better way of coping with cross-version\n> changes that affect dump scripts...\n\nNot sure how to deal with that one.\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, 4 Jan 1999 22:30:40 -0500 (EST)",
"msg_from": "Bruce Momjian <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Still having trouble importing 6.3 DB"
},
{
"msg_contents": "On Mon, 4 Jan 1999, Tom Lane wrote:\n\n\n> I think you can make it work by just changing \"copy pg_user from stdin;\"\n> to \"copy pg_shadow from stdin;\" a couple lines down from the top of\n> the dump script.\n> \n\nI had this problem ypgrading from 6.4betaX to 6.4{0,1,2}\nA quick note , first do a pg_dumpall do dump all your databases the mv \ndata data.old recreate the data directory and do an initdb that will create\nall the relations properly , otherwise the pg_user view is not installed and\nyou will have some problems! \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": "Tue, 5 Jan 1999 11:14:14 +0200 (EET)",
"msg_from": "Angelos Karageorgiou <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Still having trouble importing 6.3 DB"
},
{
"msg_contents": "Bruce Momjian <[email protected]> writes:\n>> Memo to hackers: we need a better way of coping with cross-version\n>> changes that affect dump scripts...\n\n> Not sure how to deal with that one.\n\nI don't have any good general-purpose answers either, but I can think\nof one thing we probably oughta do: seems to me that using a direct\ncopy to a system table is just asking for cross-version trouble.\npg_dumpall ought to create users using the CREATE USER command, not\nby copy to pg_shadow, don't you think?\n\nI don't mind pg_dumpall doing a select from pg_shadow in order to get\nthe info; that should be OK on the premise that the dump utility will be\nrun against a database of the same version as the utility itself. The\nthing not to assume is that the output script will be reloaded into\nexactly the same version of Postgres.\n\n\t\t\tregards, tom lane\n",
"msg_date": "Tue, 05 Jan 1999 11:02:19 -0500",
"msg_from": "Tom Lane <[email protected]>",
"msg_from_op": false,
"msg_subject": "Re: [HACKERS] Still having trouble importing 6.3 DB "
}
] |
[
{
"msg_contents": "\nmaking sure that I didn't break things all over again, while trying to fix\nthings...\n\n",
"msg_date": "Tue, 5 Jan 1999 03:02:08 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "[test] ignore this"
}
] |
[
{
"msg_contents": "\n\n",
"msg_date": "Tue, 5 Jan 1999 03:03:08 -0500 (EST)",
"msg_from": "\"Marc G. Fournier\" <scrappy>",
"msg_from_op": true,
"msg_subject": "[test] ignore this"
}
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.