workspace
stringclasses 4
values | channel
stringclasses 4
values | text
stringlengths 1
3.93k
| ts
stringlengths 26
26
| user
stringlengths 2
11
|
|---|---|---|---|---|
clojurians | clojure | also, lucky they did those better things, because they made spec! and now that validation is trivially added in a uniform way rather than ad-hoc | 2017-12-08T17:50:35.000227 | Aldo |
clojurians | clojure | <@Kareen> my line of question was in general not with respect to only core members. | 2017-12-08T17:53:44.000027 | Deandrea |
clojurians | clojure | it’s was more of “i’ve seen this GIGO thing be used as a rationale in the context of the clojure community and i’m curious what motivates it.” | 2017-12-08T17:54:18.000346 | Deandrea |
clojurians | clojure | right, I was assuming we were specifically talking about core | 2017-12-08T17:54:41.000169 | Kareen |
clojurians | clojure | <@Deandrea> The community cannot change clojure.core without the approval of one person. If someone else gives a rationale for what changes in clojure.core, or does not change in clojure.core, it is based on their interpretation/guesstimate of the reasons. | 2017-12-08T17:55:40.000292 | Micha |
clojurians | clojure | hey, someone write specs for the clojure.set functions and file a patch | 2017-12-08T17:57:26.000237 | Sonny |
clojurians | clojure | there you go :) | 2017-12-08T17:57:44.000253 | Kareen |
clojurians | clojure | My interpretation is that "can't be bothered to add the checks" seems unlikely for things like clojure.set functions, since many people would have been happy to provide patches for those years ago if they were desired by the core developer team. spec being a far more general tool is fantastic, and I'm glad it was created. | 2017-12-08T17:57:52.000244 | Micha |
clojurians | clojure | that’d be like, useful | 2017-12-08T17:57:59.000059 | Sonny |
clojurians | clojure | you don’t even have to wait for them to be in core to use them | 2017-12-08T17:58:10.000329 | Sonny |
clojurians | clojure | <@Micha> i understand that, however, my question was more broad in scope. i’m not calling out clojure core, the libraries, or the members. it was a question for the room in the context of the community purely because i’ve seen it be a rationale for doing something not bound by the performance reason. | 2017-12-08T17:58:16.000345 | Deandrea |
clojurians | clojure | just put ’em in a namespace and load them | 2017-12-08T17:58:20.000034 | Sonny |
clojurians | clojure | the typing thing, well, that’s effectively a fallacy (in my opinion). | 2017-12-08T17:59:13.000132 | Deandrea |
clojurians | clojure | to say “i won’t do _x_ because typing” can be a fine argument in certain situations but in general it’s very weak. | 2017-12-08T17:59:45.000272 | Deandrea |
clojurians | clojure | the point was simply that you see that kind of beahaviour less in typed languages (at an unacceptable cost). again. not a rationale. a statement | 2017-12-08T18:00:59.000054 | Aldo |
clojurians | clojure | in my experience, writing specs for core stuff can bring out many subtle questions. going through those and teasing apart what is expected, what works but is unexpected, and what doesn’t work now but we might want to work in the future has some measure of art to it. | 2017-12-08T18:01:00.000243 | Sonny |
clojurians | clojure | the way i read that argument is “my program won’t/will have this set of properties because (not) typing”. | 2017-12-08T18:01:04.000223 | Deandrea |
clojurians | clojure | I don't understand what "typing" you mean here :) the fingers on a keyboard one or the holy war one | 2017-12-08T18:01:48.000346 | Kareen |
clojurians | clojure | <@Kareen> fingers on the board. | 2017-12-08T18:02:07.000066 | Deandrea |
clojurians | clojure | i don’t believe in arguing about static vs dynamic typing. it’s a pointless endeavor. folks should be open to good ideas period and stop obsessing about typing discipline. | 2017-12-08T18:02:56.000163 | Deandrea |
clojurians | clojure | congrats on clojure 1.9.0! | 2017-12-08T18:03:01.000395 | Johana |
clojurians | clojure | i’m actually filled with a bit of dread now because it means i have to finish garden 2.0.0. :joy: | 2017-12-08T18:03:34.000081 | Deandrea |
clojurians | clojure | I’m celebrating with some Templeton Rye 6 year myself :) | 2017-12-08T18:03:43.000087 | Sonny |
clojurians | clojure | Your rye should be older than your programming language :slightly_smiling_face: | 2017-12-08T18:04:54.000028 | Bibi |
clojurians | clojure | I celebrated by upgrading our dev branch and pushed to staging | 2017-12-08T18:05:13.000072 | Johana |
clojurians | clojure | Or scotch | 2017-12-08T18:05:15.000289 | Bibi |
clojurians | clojure | I have some scotch that would qualify, but I keep coming back to this Templeton lately | 2017-12-08T18:05:37.000341 | Sonny |
clojurians | clojure | 3pm PST, I’m in for 1.9.0 scotch :smile: | 2017-12-08T18:06:07.000022 | Charity |
clojurians | clojure | line ’em up | 2017-12-08T18:06:23.000172 | Sonny |
clojurians | clojure | Congratulations Alex | 2017-12-08T18:06:27.000071 | Charity |
clojurians | clojure | "Ahhh, the 1.9.0 --- that was a good version for Scotch..." | 2017-12-08T18:06:42.000214 | Micha |
clojurians | clojure | ahaha | 2017-12-08T18:06:48.000217 | Deandrea |
clojurians | clojure | :smile: | 2017-12-08T18:06:51.000186 | Charity |
clojurians | clojure | they really don’t know how to do version numbers well in scotch | 2017-12-08T18:06:57.000055 | Sonny |
clojurians | clojure | branching strategies are crazy | 2017-12-08T18:07:09.000022 | Sonny |
clojurians | clojure | it’s like they were all drunk or something | 2017-12-08T18:07:31.000092 | Sonny |
clojurians | clojure | on to Clojure 1.10 :) | 2017-12-08T18:07:53.000037 | Sonny |
clojurians | clojure | <@Sonny> Out of curiosity, was it just waiting for bug reports or actually still fixing stuff in the last few weeks? | 2017-12-08T18:07:59.000156 | Johana |
clojurians | clojure | well the last few weeks has mostly been me rewriting the docs over and over at Rich’s direction :) | 2017-12-08T18:08:21.000317 | Sonny |
clojurians | clojure | Gotta make sure the docs don't over-promise :slightly_smiling_face: | 2017-12-08T18:08:43.000335 | Micha |
clojurians | clojure | but also giving it some soak time in case anything came up | 2017-12-08T18:08:46.000037 | Sonny |
clojurians | clojure | Sorry, I couldn't resist. Bad me[ | 2017-12-08T18:08:56.000173 | Micha |
clojurians | clojure | no, that’s true :) | 2017-12-08T18:09:05.000330 | Sonny |
clojurians | clojure | <@Micha> as long as the specs are valid :wink: | 2017-12-08T18:09:13.000090 | Johana |
clojurians | clojure | may all your specs be valid | 2017-12-08T18:09:21.000052 | Sonny |
clojurians | clojure | scottish blessing I think | 2017-12-08T18:09:33.000042 | Sonny |
clojurians | clojure | :thumbsup_all: to the CLI stuff. | 2017-12-08T18:10:15.000187 | Deandrea |
clojurians | clojure | Hi! Can you guys give me an idea of how to try to resolve this?
I have a test that calls a functions. This functions executes an async task. Locally if I use a `(Thread/sleep 4000)` I can assert my data just fine.
But, after I build my application on travis, the build stop in this test and never finalize the build. Someone hava already seen this? | 2017-12-08T18:12:04.000318 | Audie |
clojurians | clojure | <@Audie> are you able to patch your test (or code) in such a way that it does not require the task to be executed asynchronously? i know that’s not answering your question, however, it’s one i would ask myself in that situation. | 2017-12-08T18:14:03.000124 | Deandrea |
clojurians | clojure | <@Deandrea> I don’t see a way to do that. I have to call this async function and wait for the entire process to finish. Otherwise, when I go to the database the data is not there yet | 2017-12-08T18:16:37.000219 | Audie |
clojurians | clojure | `(Thread/sleep)` is the only way I thought how to do it | 2017-12-08T18:18:00.000254 | Audie |
clojurians | clojure | <@Audie> you could use a promise or a future. | 2017-12-08T18:18:31.000336 | Deandrea |
clojurians | clojure | <@Audie> actually i think a promise sounds like it might do the trick. you could create a promise, then execute your async task and, upon completion, deliver a value to the promise. at the end of the test you could simply dereference the promise as to block the thread and keep the build from exiting. | 2017-12-08T18:21:08.000253 | Deandrea |
clojurians | clojure | ```
(let [p (promise )]
(do-async-stuff
,,,
(deliver p some-val))
(deref p))
``` | 2017-12-08T18:21:48.000190 | Deandrea |
clojurians | clojure | someone else might have a more elegant solution. | 2017-12-08T18:22:57.000164 | Deandrea |
clojurians | clojure | but if use promise i’ll have to change the async task to put a value in the promisse right? | 2017-12-08T18:23:11.000288 | Audie |
clojurians | clojure | sounds like that’s the place in the code you need to patch. | 2017-12-08T18:23:34.000077 | Deandrea |
clojurians | clojure | so, yes. you need to be able to deliver to the promise when the async task completes. i’m guessing there’s probably a function or something that gets handed to whatever does the async work. | 2017-12-08T18:24:31.000251 | Deandrea |
clojurians | clojure | i hope i’m helping… :sweat_smile: | 2017-12-08T18:25:21.000101 | Deandrea |
clojurians | clojure | I see.. thanks for the help! I’ll try to fix this | 2017-12-08T18:29:38.000098 | Audie |
clojurians | clojure | ```(defn foo [x] {:post (some? %)} x)``` may look like it works but it actually asserts `some?` (truthy) and then asserts `%` (the value returned from `foo`) so it needs to be ```(defn foo [x] {:post [(some? %)]} x)``` although the following actually works too -- but looks odd: ```(defn foo [x] {:post (%)} x)``` (`[%]` works the same as `(%)` here -- it's just a sequence of forms to evaluate with `%` bound to the function result. _This was a surprise to me when I tried it in the REPL!_ | 2017-12-08T18:35:14.000023 | Daniell |
clojurians | clojure | for implementing a forth interpreter, is vector or list better matched for storing the stack ? | 2017-12-08T18:40:13.000041 | Berry |
clojurians | clojure | <@Daniell> yes, that was a typo. | 2017-12-08T18:40:30.000030 | Deandrea |
clojurians | clojure | thank you for catching that. | 2017-12-08T18:40:39.000156 | Deandrea |
clojurians | clojure | <@Berry> if you use conj/peek/pop, vectors and lists behave the same as stacks | 2017-12-08T18:41:21.000211 | Kareen |
clojurians | clojure | <@Deandrea> I was genuinely surprised it "worked" -- I rarely use `:pre`/`:post` so I nearly always have to look up the syntax! | 2017-12-08T18:41:34.000149 | Daniell |
clojurians | clojure | <@Kareen>: sure, but I suspect they have different performance characteristics | 2017-12-08T18:41:50.000112 | Berry |
clojurians | clojure | I tend to use vectors as stacks just because `pop-n` and `peek-n` are easier to implement efficiently on vectors | 2017-12-08T18:41:52.000171 | Kareen |
clojurians | clojure | <@Berry> +1 for vectors. and enjoy the concatenative journey. :slightly_smiling_face: | 2017-12-08T18:42:01.000107 | Deandrea |
clojurians | clojure | <@Berry> no, all those operations are constant time on both vectors and lists | 2017-12-08T18:42:07.000090 | Kareen |
clojurians | clojure | <@Kareen>: I think vector is log_32 #-elemes, which is <= 5 in most caeds, but technically log n | 2017-12-08T18:42:31.000019 | Berry |
clojurians | clojure | Hey <@Berry> If you want Clojure + Forth, take a look at <https://gershwin.github.io/> :slightly_smiling_face: | 2017-12-08T18:43:01.000135 | Daniell |
clojurians | clojure | I was actually sufficiently awed by Gershwin at one point that I tried it out at work and wrote a few functions in the "Forth" style of Clojure! :slightly_smiling_face: | 2017-12-08T18:43:51.000099 | Daniell |
clojurians | clojure | <@Berry> no that's not correct, conj/pop/peek on vectors are amortized constant time, not log_32, nth/assoc are log_32 on vectors | 2017-12-08T18:44:12.000247 | Kareen |
clojurians | clojure | <@Daniell>: I've lost track the # of afternoons I've lost due to you providing fasicnating links. | 2017-12-08T18:44:16.000167 | Berry |
clojurians | clojure | (but since it relies on a fork of Clojure and wasn't being updated, I quickly went back to pure Clojure) | 2017-12-08T18:44:31.000204 | Daniell |
clojurians | clojure | <@Kareen>: conj has to be log_32 time as it has to create a new node for every level of the b-tree | 2017-12-08T18:45:15.000020 | Berry |
clojurians | clojure | clojure's persistent vectors are not b-trees | 2017-12-08T18:45:36.000319 | Kareen |
clojurians | clojure | I thought they were btrees with 32 fanout -- if not btrees, what are they? | 2017-12-08T18:45:58.000254 | Berry |
clojurians | clojure | <http://hypirion.com/musings/understanding-persistent-vector-pt-1> | 2017-12-08T18:46:25.000004 | Kareen |
clojurians | clojure | <@Berry> if you’re interested have a look at factor, joy, cat, and kitten in this space as well. | 2017-12-08T18:46:50.000266 | Deandrea |
clojurians | clojure | factor is a ton of fun. | 2017-12-08T18:47:13.000243 | Deandrea |
clojurians | clojure | <@Berry> in particular <http://hypirion.com/musings/understanding-persistent-vector-pt-3> | 2017-12-08T18:47:17.000098 | Kareen |
clojurians | clojure | this is the optimization that makes conj amortized constant time | 2017-12-08T18:47:26.000236 | Kareen |
clojurians | clojure | same for pop/peek | 2017-12-08T18:47:55.000120 | Kareen |
clojurians | clojure | 1. the depth of the tree is log_32 n
2. when we do a conj, we have to create a new node at each level of the tree
^-- which is the two above statements is false ? -- because if they're both true, it's log_32 n time EVERY TIME | 2017-12-08T18:48:31.000363 | Berry |
clojurians | clojure | 2 is not true | 2017-12-08T18:49:07.000290 | Kareen |
clojurians | clojure | ```
Instead of keeping the rightmost leaf in the tree itself, we keep a direct reference to it in the vector header: That's the last block which has been added to the vector head since last blogpost. The reference to the rightmost leaf node is called the tail.
```
ah | 2017-12-08T18:49:55.000250 | Berry |
clojurians | clojure | that's clever; bronsa++ | 2017-12-08T18:49:59.000135 | Berry |
clojurians | clojure | well I didn't come up with any of this | 2017-12-08T18:50:15.000146 | Kareen |
clojurians | clojure | thanks for correcting me on this time, I've been thinking conj was log_32 n time | 2017-12-08T18:50:33.000086 | Berry |
clojurians | clojure | :smile: | 2017-12-08T18:50:33.000299 | Deandrea |
clojurians | clojure | wrt "what are they", I've heard the clj vector impl called bitmapped vector trie, not sure if there's a better name | 2017-12-08T18:50:49.000176 | Kareen |
clojurians | clojure | <@Daniell>: I feel like we can get most of forth in clojure by defining macro `f->`, where it behaves like `->` except
1. if it sees a constant (number, kw, string), it pushes it to the stack
2. all functionsin it take stack as input and produces stack as output | 2017-12-08T18:52:17.000298 | Berry |
clojurians | clojure | <@Berry> the thing is though you’ll need to come up with a way to call out how much of the stack to consume for fn’s that have multiple arities. | 2017-12-08T18:57:08.000126 | Deandrea |
clojurians | clojure | e.g. how do you interpret `["foo" "bar" "baz" str]`? | 2017-12-08T18:57:32.000165 | Deandrea |
clojurians | clojure | <https://twitter.com/brandonbloom/status/528262785642545153> this is still my favourite impl of stack programming in clojure :) | 2017-12-08T18:58:11.000004 | Kareen |
clojurians | clojure | <https://github.com/brandonbloom/factjor> for a more useful impl | 2017-12-08T18:58:49.000051 | Kareen |
clojurians | clojure | :+1: | 2017-12-08T18:59:53.000016 | Deandrea |
clojurians | clojure | <@Deandrea>: afaik, forth doesn't support var-arity functions, so f-> won't either | 2017-12-08T19:00:52.000297 | Berry |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.