2014-06-08
| 00:35 | allenj12 | has anyone tried this little tutorial out? http://zaiste.net/2014/02/web_applications_in_clojure_all_the_way_with_compojure_and_om/ and got it to work? |
| 00:42 | catern | allenj12: are you having trouble with it? looks interesting, I might do it |
| 00:43 | allenj12 | catern: yea the widgets arent showing is all |
| 00:43 | allenj12 | catern: i thought it looked interesting too :) |
| 00:43 | catern | comment about it? |
| 00:47 | ddellacosta | allenj12: what version of Om are you using? |
| 00:47 | allenj12 | ddellacosta: [om "0.6.4"] i updated all the dependencies |
| 00:47 | catern | is clojure useful for anything but web development? |
| 00:48 | arrdem | I write compilers and data analysis code in clojure... |
| 00:48 | catern | ok |
| 00:48 | catern | just checking |
| 00:48 | ddellacosta | allenj12: yeah, that'll break it. There were some changes in how stuff gets passed into root and whatnot post 0.4.x I think, or maybe it was 0.5.x, I'm forgetting |
| 00:48 | arrdem | I've got microarchitecture simulators in clojure... |
| 00:48 | arrdem | there are IRC bots in clojure |
| 00:48 | arrdem | I'm writing a dogeon parser in clojure.. |
| 00:48 | arrdem | it's a general purpose lang ffs |
| 00:48 | ddellacosta | allenj12: also, I really don't recommend calling transact! in init-state like that is, it's strange |
| 00:49 | catern | yes I know |
| 00:49 | catern | I just needed a reminder because I want to write a FUSE filesystem for something and I think it would be a bad idea to do it in Clojure |
| 00:49 | allenj12 | ddellacosta: o really? yea i know really nothing of it. i was just hoping to get something running than figure out piece by piece by playing with it |
| 00:49 | catern | (because JVM) |
| 00:50 | allenj12 | ddellacosta: do you know of any similar/more modern tutorials? |
| 00:50 | ddellacosta | allenj12: yeah, I would just hesitate to start with that one--go straight for David Nolen's tutorials in the Om docs (if you haven't yet) |
| 00:51 | ddellacosta | allenj12: https://github.com/swannodette/om#tutorial |
| 00:51 | catern | ddellacosta: do you recommend Om? has it any viable competitors? |
| 00:51 | catern | (for someone not tied to react) |
| 00:51 | ddellacosta | catern: I haven't tried the alternative React wrappers in Clojure, if that's what you're asking about |
| 00:51 | allenj12 | ddellacosta: kk its not om in paticular i need i was just using it as a clj/cljs tutorial |
| 00:52 | ddellacosta | allenj12: oh, then definitely don't start with that one...haha. You could try mimmo cosenza's tutorials, I think they're decent, although not sure how up to date they are |
| 00:52 | ddellacosta | allenj12: https://github.com/magomimmo/modern-cljs |
| 00:53 | allenj12 | ddellacosta: kk cool, yea i just havent found one that did a backend and front end that acually did something. thank for the direction |
| 00:53 | ddellacosta | allenj12: but this may be best to start with: http://swannodette.github.io/2013/10/27/the-essence-of-clojurescript/ |
| 00:53 | allenj12 | ddellacosta: hmm kk |
| 00:53 | ddellacosta | allenj12: which you could follow with this: http://swannodette.github.io/2013/11/07/clojurescript-101/ |
| 00:54 | ddellacosta | allenj12: in general, I recommend poking around David Nolen's blog (what I just linked you to). And yeah, I agree that there is a gap in the "hooking up the backend to Om" tutorial space. Unfortunately I can't really recommend the one you were looking at, it looks like it has some not-so-standard ways of writing Om code |
| 00:55 | ddellacosta | allenj12: of course, just pop back here if you have questions. :-) |
| 00:55 | allenj12 | awsome thanks |
| 00:55 | allenj12 | :) |
| 00:56 | ddellacosta | catern: sorry I didn't really answer your question directly: I do recommend Om, if you want to build a very dynamic one-page app kind of front-end. |
| 01:00 | catern | okay, thanks |
| 01:45 | rpaulo | lobste.rs has a lot of clojure links |
| 02:19 | allenj12 | ddellacosta: that last one was a good little tutorial :) where the other links good? i lost them since my computer crashed |
| 02:20 | ddellacosta | allenj12: which one did you just look at? I can give you the other links again |
| 02:21 | ddellacosta | allenj12: first I suggested looking at this one: http://swannodette.github.io/2013/10/27/the-essence-of-clojurescript/ Then, http://swannodette.github.io/2013/11/07/clojurescript-101/ |
| 02:21 | allenj12 | ddellacosta: looked at this one http://swannodette.github.io/2013/11/07/clojurescript-101/ |
| 02:21 | allenj12 | ahh kk |
| 02:21 | ddellacosta | allenj12: and for a more in depth guide: https://github.com/magomimmo/modern-cljs |
| 02:21 | allenj12 | that was it |
| 02:22 | ddellacosta | allenj12: oh, ok. Yeah, also just poke around dnolen's blog, there is a lot of very good, authoritative info there: http://swannodette.github.io/ |
| 02:22 | ddellacosta | allenj12: and if you are interested in Om, do check out the tutorials on the github page: https://github.com/swannodette/om#tutorial |
| 02:22 | ddellacosta | allenj12: but you may want to get comfortable with the other stuff first...or not. :-) |
| 02:23 | Frozenlock | allenj12: Also make sure to check Reagent, Om closest competition :-p |
| 02:23 | ddellacosta | Frozenlock: give him a tutorial link if you have one |
| 02:23 | allenj12 | ddellacosta: yea like i said i dont acually care to much about om :) i will have to incorporate mysql tho (or some sql...) |
| 02:24 | allenj12 | i guess i could use it later tho.. hmmm |
| 02:24 | allenj12 | lol |
| 02:24 | Frozenlock | ddellacosta: Cmon, I'm sure he can google.... http://holmsand.github.io/reagent/ |
| 02:24 | allenj12 | Frozenlock: nope please link :p |
| 02:25 | john2x | if I do `cider-jack-in`, and do something else while it's `Starting nREPL server...`, does it cancel starting the server? |
| 02:25 | allenj12 | john2x: pretty sure not |
| 02:26 | allenj12 | john2x: havent used emacs in a while tho |
| 02:26 | Frozenlock | john2x: Not in my experience |
| 02:26 | john2x | Frozenlock: how do I cancel the server startup |
| 02:26 | john2x | I think it's downloading a lot of dependencies, and with no feedback, it's hard to know what's happening |
| 02:27 | Frozenlock | john2x: I don't know... let it do its stuff and close it after :-p |
| 02:27 | john2x | or is there a buffer for cider-messages or some sort? |
| 02:27 | ddellacosta | allenj12: if you want to do back-end Clojure stuff, I would take a look at clojure.java.jdbc first: https://github.com/clojure/java.jdbc |
| 02:27 | ddellacosta | allenj12: er, rather, if you want to connect to a DB |
| 02:28 | allenj12 | ddellacosta: that i acually already did :) hehe but not much further i basically just created a db a table and added some rows lol :) |
| 02:29 | ddellacosta | allenj12: ah, okay. Well, that's a start. :-) |
| 02:30 | allenj12 | ddellacosta: when im done with this, i kinda wanna make a tutorial that goes over both front-end back-end together. start to finish on a little real world "seeable" website |
| 02:32 | ddellacosta | allenj12: do it! But throw it at folks in IRC/mailing list too so you can get some constructive criticism--for example, that first link you gave me had some weird stuff in it, and that can be really confusing to beginners. I want to help people avoid a similar fate. |
| 02:33 | ddellacosta | allenj12: but that's premature. First thing to do is just give it a shot. :-) |
| 02:34 | allenj12 | ddellacosta: indeed ill definatly do it, might be a while in the future. gotta finish this project and then my masters project. maybe end of next semester |
| 02:34 | ddellacosta | allenj12: hey, no pressure from me...haha |
| 02:36 | allenj12 | ddellacosta: :) ill probably do it just to spread more functional programming at my school. very few of us here who are willing to try it. i did convert some people tho talking about machine learning and how clojure makes things very simple |
| 02:36 | ddellacosta | allenj12: awesome. |
| 02:37 | allenj12 | :) |
| 02:39 | allenj12 | ooo reagent looks pretty nice |
| 02:39 | MagBo[MURT] | Hello people, a ranty question: can you please name one "atomic" things about clojure atoms? Anyone has a link about rationale behind picking such a term? |
| 02:40 | MagBo[MURT] | Why not "synchronous agents" or something less counter-intuitive? |
| 02:41 | allenj12 | MagBo[MURT]: well not sure but heres my guess: atoms are a type that can be multiple things very much like a common lisp atom |
| 02:41 | amalloy | changes made to atoms are atomic |
| 02:41 | MagBo[MURT] | amalloy: touche |
| 02:41 | amalloy | allenj12: no, that's quite wrong. lisp's atoms were named so because they're the indivisible particles at the bottom of a lisp data structure |
| 02:42 | MagBo[MURT] | ^ |
| 02:42 | tolstoy | So was rhickey punning a bit because of that? |
| 02:42 | allenj12 | amalloy: kk yea tbh i dont know why i ever thought that was right lol |
| 02:44 | allenj12 | Frozenlock: is there any good reason to use reagent vs om? it seems cool but i have no knowledge comparing the two |
| 02:45 | Frozenlock | allenj12: I find it more clojure-y |
| 02:46 | allenj12 | Frozenlock: hmmmm good enough for me lol |
| 02:46 | Frozenlock | :-p |
| 02:50 | john2x | how do I add a jar file as a dependency in project.clj? |
| 02:55 | ddellacosta | allenj12, Frozenlock: saying something is "more clojure-y" is quite subjective--not really a strong reason to choose one library over the other. |
| 02:56 | allenj12 | ddellacosta: lol yea :p im just looking at both of them |
| 02:57 | Frozenlock | ddellacosta: Both are just wrappers to react.... |
| 02:57 | ddellacosta | Frozenlock: it's not that simple, which David Nolen explains in detail in his posts about Om |
| 02:58 | ddellacosta | Frozenlock, allenj12: For one, Om takes advantage of ClojureScript's immutable data structures to make React's life cycle rendering more efficiently, but checking on referential equality by default in ShouldComponentUpdate |
| 02:58 | ddellacosta | *more efficient |
| 02:59 | allenj12 | has anyone here used both? |
| 02:59 | ddellacosta | Frozenlock, allenj12: Om also takes advantage of protocols to allow for flexibly extending the base React lifecycle methods in a very "clojure-y" way. |
| 03:00 | Frozenlock | ddellacosta: http://holmsand.github.io/reagent/news/reagent-is-async.html |
| 03:01 | ddellacosta | Frozenlock: your point? Om uses requestAnimationFrame...which is mentioned in that article, and seems to be in fact where he got the idea. |
| 03:01 | Frozenlock | Yes, showing there's less difference than you implied |
| 03:02 | Frozenlock | Both are using immutable data structure, because... well... clojurescript :-) |
| 03:02 | ddellacosta | Frozenlock: just because they are both CLJS doesn't mean they are both doing the same thing with them. Does reagent also do referential equality checks in ShouldComponentUpdate? I didn't know that if so. |
| 03:03 | Frozenlock | I honestly don't know. |
| 03:06 | allenj12 | somone had to have compared the two sometime no? |
| 03:06 | ddellacosta | allenj12: well, in the end, I suggest you try them both, learn about their different strengths, and see which you prefer. My basic takeaway is that Om gives you more flexibility and extensibility at the cost of a higher learning curve than Reagent |
| 03:08 | Frozenlock | I'd honestly like to see an Om app that can't be done in Reagent. Perhaps one of the custom lifecyle methods? |
| 03:08 | ddellacosta | Frozenlock: it looks like reagent does have a similar strategy for updating: https://github.com/holmsand/reagent/blob/master/src/reagent/impl/template.cljs#L142-L146 Not sure how wrap-component is used though |
| 03:09 | ddellacosta | Frozenlock: I don't know, I can't speak to that. I can only say that I've found Om's protocols to be a very good way to extend React lifecycle methods, and I use them extensively. |
| 03:10 | Frozenlock | o_O |
| 03:10 | Frozenlock | By lifecycle, you mean the component-did-update, component-will-update and all? |
| 03:10 | ddellacosta | Frozenlock: yes |
| 03:11 | Frozenlock | Intriguing... when were React methods not enough? |
| 03:12 | ddellacosta | Frozenlock: ...when was JS not enough? Not sure I understand what you're getting at |
| 03:13 | Frozenlock | well, you already have component-will-mount, did-mount, will-receive-props, will-update, did-update, will-unmount... what's a possible 'other' lifecycle? |
| 03:14 | ddellacosta | Frozenlock: it's the same lifecycle. My point was that protocol semantics are a very nice way to provide implementations of those lifecycle functions when it is necessary to do so, nothing more. |
| 03:15 | Frozenlock | Oh |
| 03:15 | ddellacosta | Frozenlock: yes, I see what the confusion was--I probably should have said "implementing" vs. "extending" |
| 03:16 | arrdem | ping |
| 03:16 | ddellacosta | arrdem: pong |
| 03:16 | arrdem | just testing a server side logging configuration. |
| 03:16 | arrdem | nothing to see here. |
| 03:16 | ddellacosta | ah. |
| 05:16 | john2x | I"m getting "Error: (error "Can't find nREPL middleware providing op stacktrace. Please, install cider-nrepl and restart CIDER")" messages in my *Messages* buffer.. but there's no cider-nrepl package in marmalade? |
| 05:21 | danielcompton | john2x I prefer melpa to marmalade, the packages seem to be more up to date there |
| 05:24 | john2x | ohh it's refering to a lein plugin |
| 05:24 | john2x | forgot to backup my profiles.clj :( |
| 05:39 | danielcompton | john2x Ah yes, there's a new cider-nrepl plugin |
| 05:40 | danielcompton | john2x forgot about that, even though I just upgraded a week ago |
| 05:47 | hariravula | hello all |
| 06:09 | mi6x3m | hey clojure, is there a way to force the execution of a future without deref? |
| 06:09 | mi6x3m | something like a hint to being immediately |
| 06:19 | ddellacosta | mi6x3m: I don't believe so, you explicitly cede control over execution when you use a future. |
| 06:19 | ddellacosta | mi6x3m: and deref isn't forcing execution as far as I know; it's just blocking until execution is complete |
| 06:44 | john2x | is there something similar to Python's `dir` function? to list public methods/properties of a Java object? |
| 06:46 | pyrtsa | Not directly but you can play with clojure.reflect to create your own. (No idea if the following works here:) |
| 06:46 | pyrtsa | ,(->> Object clojure.reflect/reflect :members (filter (comp :public :flags)) (map :name)) |
| 06:46 | clojurebot | #<CompilerException java.lang.ClassNotFoundException: clojure.reflect, compiling:(NO_SOURCE_PATH:0:0)> |
| 06:47 | pyrtsa | john2x: Just replied to you while you were away. X-) |
| 06:47 | pyrtsa | Not directly but you can play with clojure.reflect to create your own. |
| 06:47 | pyrtsa | ,(do (require 'clojure.reflect) (->> Object clojure.reflect/reflect :members (filter (comp :public :flags)) (map :name))) |
| 06:47 | clojurebot | (getClass wait java.lang.Object equals notifyAll ...) |
| 07:01 | pwiki | Is there a way to make instaparse more greedy to prevent it from making unnecessary output nodes, other than using regular expressions? |
| 08:45 | expez | Why does (source app.foo.bar/baz) return nil? I have src/clj/app/foo/bar.clj with (ns app.foo.bar) and the fn baz in it and :source-paths in project.clj contains "src/clj". |
| 08:45 | Jarda | hey all! |
| 08:46 | Jarda | anyone know something about lobos? |
| 08:49 | Jarda | I can't even require lobos.connectivity |
| 08:49 | Jarda | user=> (use '(lobos connectivity core schema)) |
| 08:49 | Jarda | CompilerException java.lang.RuntimeException: Unable to resolve symbol: *db* in this context, compiling:(lobos/connectivity/jdbc_2.clj:14:1) |
| 09:13 | no7hing | https://www.refheap.com/86377 on line 6, is that expected behaviour in clojurescript? |
| 09:14 | no7hing | the compiler complains on non-existing namespaces, but not non-existing vars |
| 09:36 | john2x | what does java.lang.IllegalArgumentException mean in this context https://www.refheap.com/86378? |
| 09:39 | expez | john2x: it's not expecting a String |
| 09:40 | expez | As far as I can google it's expecting a List<String> |
| 09:41 | john2x | expez: I tried (.setArguments chrome-options ["foo"]) as well |
| 09:41 | Glenjamin | expez: source prints to console |
| 09:41 | Glenjamin | ,(doc source) |
| 09:41 | clojurebot | "([n]); Prints the source code for the given symbol, if it can find it. This requires that the symbol resolve to a Var defined in a namespace for which the .clj is in the classpath. Example: (source filter)" |
| 09:41 | john2x | does it need to be a Java List? |
| 09:42 | Glenjamin | john2x: that error says that it can't find method on the ChromeOptions class |
| 09:43 | john2x | Glenjamin: oh ok. but it shows up in reflect? |
| 09:43 | Glenjamin | as expez said, it looks like it should take a List<String> |
| 09:43 | Glenjamin | http://javadox.com/org.seleniumhq.selenium/selenium-chrome-driver/2.40.0/org/openqa/selenium/chrome/ChromeOptions.html#addArguments%28java.util.List%29 |
| 09:44 | Glenjamin | so (.addArguments chrome-options ["foo"]) should work |
| 09:44 | Glenjamin | ,(implements? java.util.List clojure.lang.Vector) |
| 09:44 | clojurebot | #<CompilerException java.lang.RuntimeException: Unable to resolve symbol: implements? in this context, compiling:(NO_SOURCE_PATH:0:0)> |
| 09:44 | Glenjamin | erm |
| 09:44 | Glenjamin | (instance? java.util.List []) |
| 09:44 | Glenjamin | ,(instance? java.util.List []) |
| 09:44 | clojurebot | true |
| 09:44 | Glenjamin | yeah, vector should be fine |
| 09:52 | john2x | oh this is unfortunate.. so it throws the same IllegalArgumentException for both a non-existent instance member and bad arguments to an existing instance member.. |
| 09:57 | Glenjamin | yeah, because of method overloading, methods with different typehints are different methods |
| 09:57 | Bronsa | john2x: eh, error reporting on interop forms is not really consistent |
| 09:58 | Bronsa | ,(Integer/parseInt :foo :bar) |
| 09:58 | clojurebot | #<ClassCastException java.lang.ClassCastException: clojure.lang.Keyword cannot be cast to java.lang.String> |
| 09:58 | Bronsa | ,(Integer/parseInt :foo :bar :baz) |
| 09:58 | clojurebot | #<CompilerException java.lang.IllegalArgumentException: No matching method: parseInt, compiling:(NO_SOURCE_PATH:0:0)> |
| 09:58 | john2x | wat |
| 09:58 | Bronsa | ,(.foo "bar") |
| 09:58 | clojurebot | #<IllegalArgumentException java.lang.IllegalArgumentException: No matching field found: foo for class java.lang.String> |
| 09:58 | Bronsa | ,(.foo "bar" 1) |
| 09:58 | clojurebot | #<IllegalArgumentException java.lang.IllegalArgumentException: No matching method found: foo for class java.lang.String> |
| 09:59 | Bronsa | ,(.toString "bar" 1) |
| 09:59 | clojurebot | #<IllegalArgumentException java.lang.IllegalArgumentException: No matching method found: toString for class java.lang.String> |
| 09:59 | Bronsa | mh |
| 09:59 | Bronsa | well, whatever |
| 10:23 | pandeiro | anyone know what i can do to get rid of the message 'Java HotSpot(TM) Client VM warning: TieredCompilation is disabled in this release.' ? |
| 10:30 | AWizzArd | pandeiro: where do you see this message? |
| 10:31 | pandeiro | AWizzArd: running lein with Oracle JDK 1.8.0_05 |
| 10:31 | AWizzArd | pandeiro: have you specified any options? Such as :jvm-opts ["-XX:+TieredCompilation"] |
| 10:33 | pandeiro | AWizzArd: no, let me try that - can it be done on a lein-wide basis in :jvm-opts in profiles.clj? |
| 10:34 | AWizzArd | pandeiro: should be possible, though I wouldn’t suggest that, for reasons of repeatability of your build. |
| 10:35 | pandeiro | AWizzArd: not fixed - from the error i am thinking i need to disable lein's use of tiered compilation, no? |
| 10:36 | pandeiro | i'm less concerned about build repeatability since i am mostly working with cljs - i just want this error to disappear from every lein invocation i do |
| 10:36 | AWizzArd | pandeiro: well, with the above option you want to activate tiered compilation. |
| 10:37 | AWizzArd | pandeiro: you could try to make the ‘+’ in front of ‘Tiered’ into a minus ‘-’. |
| 10:38 | AWizzArd | pandeiro: but I think your message is coming from the lein.bat or lein.sh script. |
| 10:38 | AWizzArd | See: https://github.com/technomancy/leiningen/blob/master/bin/lein#L136 |
| 10:38 | pandeiro | AWizzArd: yep, that doesn't work either... i also tried :jvm-opts ^:replace [] in profiles |
| 10:38 | pandeiro | i'll check that out, thanks |
| 10:38 | AWizzArd | Leiningen’s starter shell script itself issues this option. |
| 10:39 | pandeiro | k so i go in and patch $HOME/bin/lein itself maybe |
| 10:40 | AWizzArd | pandeiro: I guess this should do the trick. You probably just have to comment out that part, or make that + into a -. |
| 10:40 | pandeiro | also find it weird that googling for TieredCompilation being deprecated in Java 8 doesn't pull up anything... |
| 10:40 | pandeiro | i wonder if this is just something about this build i am using |
| 10:43 | AWizzArd | pandeiro: I am using Java 8 and can’t remember having seen this message. |
| 10:43 | pandeiro | AWizzArd: Oracle? |
| 10:43 | AWizzArd | But most likely I am not using the newest Leiningen, and perhaps it is different in the lein.bat file, dunno. |
| 10:44 | pandeiro | alright, i guess i will live with this for now |
| 10:44 | pandeiro | AWizzArd: thanks for the help |
| 10:54 | technomancy | pandeiro: you can set LEIN_JVM_OPTS iirc |
| 10:55 | technomancy | pandeiro: also: I had no idea there was a client jvm for java 8. is this a 32-bit machine? |
| 10:55 | pandeiro | technomancy: yes, 32-bit |
| 10:57 | pandeiro | can i set LEIN_JVM_OPTS somewhere within lein's config (eg profiles.clj)? |
| 11:04 | technomancy | pandeiro: yeah, export it in ~/.lein/leinrc |
| 11:04 | technomancy | pandeiro: |
| 11:04 | technomancy | pandeiro: what happens if you edit bin/lein to remove -client near the bottom? |
| 11:05 | pandeiro | technomancy: i haven't tried that; let me see |
| 11:05 | technomancy | that's basically a leftover attempt to get faster launch that only works on 32-bit; should probably get rid of it anyway |
| 11:06 | pandeiro | technomancy: yes removing that flag removes the error |
| 11:06 | technomancy | cool; will get rid of it |
| 11:07 | technomancy | pandeiro: does it affect boot time at all? |
| 11:07 | pandeiro | technomancy: from what i can tell with a few `lein new` commands, no |
| 11:08 | technomancy | how about `time lein version`? |
| 11:09 | pandeiro | technomancy: avg 0.1s different over five runs |
| 11:09 | technomancy | very nice; thank you |
| 11:10 | pandeiro | technomancy: thanks for the assistance |
| 11:10 | technomancy | np |
| 11:10 | technomancy | it'll be in 2.4.0, hopefully going out in the next two days |
| 11:11 | technomancy | is this a 32-bit arm or 32-bit intel, out of curiosity |
| 11:11 | pandeiro | technomancy: intel |
| 11:11 | pandeiro | 2011-era laptop |
| 11:12 | AWizzArd | Everybody knows this, but it still is worth a re-read: http://therobot.tumblr.com/post/26900072/emacs-outshines-all-other-editing-software-in |
| 11:13 | pandeiro | AWizzArd: yep, unfortunately true |
| 11:13 | technomancy | pandeiro: cool; fun when you can make use of old machines |
| 11:13 | pandeiro | i say unfortunately b/c i have adolescent students who based on aesthetics alone will not touch emacs b/c it looks like crap next to modern IDE's |
| 11:14 | TimMc | technomancy: "old" |
| 11:14 | pandeiro | technomancy: i wish i were having less "fun" in this case but this is what i got! |
| 11:14 | technomancy | actually wayt |
| 11:14 | technomancy | wait |
| 11:14 | technomancy | that's like three years newer than the laptop I'm using right now |
| 11:15 | pandeiro | technomancy: i hope you aren't having to compile cljs :) |
| 11:15 | technomancy | lucky me =) |
| 11:18 | Glenjamin | people actually avoid emacs because of looks? |
| 11:18 | technomancy | but I feel like the two laptops I had before this one were also 64-bit |
| 11:18 | Glenjamin | my only reason to avoid it is the keybindings are foreign, and i can't bring myself to invest the time to learn them |
| 11:18 | technomancy | Glenjamin: people come up with all kinds of excuses not to learn new things |
| 11:19 | technomancy | it's a defense mechanism |
| 11:19 | catern | IDEs look like crap compared to emacs to me |
| 11:19 | catern | do they even have themes? :D |
| 11:19 | Glenjamin | every time i accidentally open emacs, i discover i've forgotten the keybinding for kill-emacs |
| 11:20 | Glenjamin | i know there's a g in it |
| 11:20 | Glenjamin | C-c g? |
| 11:22 | ToxicFrog | Glenjamin: ctrl-alt-F1 pkill emacs >.> |
| 11:24 | Glenjamin | bah, there isn't even a g in it |
| 11:24 | Glenjamin | one day, i will learn this |
| 11:24 | Glenjamin | unless lighttable continues to improve, maybe it'll do |
| 11:27 | technomancy | one thing at a time |
| 11:29 | TimMc | Glenjamin: C-g is "shit shit shit get me out of this" |
| 11:29 | TimMc | maybe that's what you're thinking of |
| 11:29 | Glenjamin | yeah, i think so |
| 11:31 | catern | Glenjamin: it's C-x C-c |
| 11:31 | Bronsa | TimMc: ahah, I feel like that could be an accurate docstring for C-g |
| 11:38 | mdeboard | well if anyone ever wanted the edges & vertices of the Ticket to Ride board game (original) as Clojure data, here you go https://gist.github.com/mattdeboard/08fbe157b029355dbc94 |
| 11:39 | mdeboard | I just spent a half-hour of my life on this |
| 11:46 | technomancy | it sounds like the github releases feature requires you to omit the `git push --tags` step, as it creates a tag for you? |
| 11:47 | Glenjamin | on the releases pages it says " Choose an existing tag, or create a new tag on publish " |
| 11:48 | arrdem | mornin' |
| 11:48 | Glenjamin | i think if you push tags, they appear on the releases page, then if you create a matching release is fleshes out the detail |
| 11:48 | technomancy | Glenjamin: oh, the dummy repo I'm testing with doesn't have any |
| 11:48 | technomancy | that would explain it |
| 11:49 | technomancy | trying to get off my own S3 bucket for lein downloads as the monthly bill keeps creeping up |
| 11:49 | Glenjamin | yeah, it offers tag autocomplete, then gives you a green "Existing tag" indicator |
| 11:52 | mdeboard | This is frustrating. Is CIDER with cider-nrepl 0.7.0 misbehaving for anyone else? |
| 11:52 | mdeboard | misses* |
| 11:53 | arrdem | mdeboard: define misbehaving |
| 11:53 | mdeboard | cider-jack-in results in a Class Not Found Error |
| 11:53 | arrdem | .... did you just upgrade to 0.7.0-SNAPSHOT? |
| 11:54 | mdeboard | Unhandled java.lang.ClassNotFoundException cider.nrepl |
| 11:54 | mdeboard | Yes |
| 11:54 | mdeboard | Because I was getting the same thing with 0.6.1 |
| 11:54 | arrdem | because if so this is expected. Cider 0.7.0 requires that you add a lein plugin to your default profile. |
| 11:54 | mdeboard | sheesh. |
| 11:54 | mdeboard | ok |
| 11:55 | arrdem | lemme grab it for ya.. |
| 11:55 | arrdem | {:user {:plugins [cider/cider-nrepl "0.7.0-SNAPSHOT"]}} should do for ya |
| 11:55 | mdeboard | Ok, thanks |
| 11:56 | mdeboard | Oh, I actually had it in my profiles.clj already (:dev instead of :user) but I hadn't upgraded there. |
| 11:56 | mdeboard | I was just adding clojure-nrepl to project.clj |
| 12:18 | pepijndevos | I suppose a record would serialize to json just as well as any map? |
| 12:28 | mi6x3m | hey clojure, how would you name a function argument which is a function which gets proxy'd into some kind of a handler? |
| 12:28 | mi6x3m | (defn default-...-handler [f] ....) |
| 12:28 | mdeboard | f |
| 12:29 | mi6x3m | yeah, would call it 'f' as well |
| 12:43 | Glenjamin | does anyone else get the same result as this: "lein repl :headless; ctrl+z; bg" -> says "Stopped" |
| 12:49 | Glenjamin | as in, i cant get lein repl :headless to run in the background via job control |
| 12:50 | arrdem | unless you know what your doing lein is not usually recommended for running deployment code... |
| 13:07 | engblom | I have decided to to try to give emacs a chance (with evil added). Could anyone point me to any up to date guide for how to configure emacs24 for editing clojure? All tutorials seem to be many years old and the git repositories often mentioned have not been seeing any changes in 4 years. |
| 13:08 | dnolen_ | ClojureScript users, 1.6.0 branch is ready to be tested, feedback welcome https://github.com/clojure/clojurescript/tree/1.6.0 |
| 13:08 | engblom | What should I use at this moment? Is slime still the tool? |
| 13:09 | engblom | Emacs24 apparently have a package manager. Is there anything useful directly from it? |
| 13:09 | Shayanjm | engblom: what do you use right now? |
| 13:09 | Shayanjm | (just a random q) |
| 13:09 | engblom | Shayanjm: Just vim with rainbow parentheses added. |
| 13:09 | Shayanjm | gotcha |
| 13:10 | arrdem | engblom: fireplace.vim is supposed to be good |
| 13:10 | gtrak | I take back anything bad I said about Swift, knowing Lattner was the lead means it's gotta be hella fast :-). |
| 13:10 | arrdem | engblom: https://github.com/tpope/vim-fireplace |
| 13:11 | engblom | arrdem: Thanks |
| 13:11 | arrdem | engblom: baring that, yeah the instructions for getting running with emacs/cider/clojure are pretty bad I'll try and do a writeup later today |
| 13:12 | arrdem | or gtrak could cause you know he's actually a cider maintainer :P |
| 13:12 | gtrak | hehe, I've been wayyyy too busy. |
| 13:12 | gtrak | I hope in the next week or two to push on solidifying stuff for a release. |
| 13:12 | arrdem | dude I wrote a DSON parser. http://github.com/arrdem/dogeon #summervacation |
| 13:13 | arrdem | engblom: if you'd like it I can quickly share the relevant parts of my .emacs config, but sharing the whole thing is nontrivial. |
| 13:14 | engblom | So (> cider slime) |
| 13:14 | systemfault | arrdem: Awesome O_O |
| 13:14 | engblom | I will have to look up cider |
| 13:14 | gtrak | arrdem: that's speshal |
| 13:15 | arrdem | engblom: yeah. slime was succeeded by nrepl.el which got renamed to cider. |
| 13:16 | gtrak | engblom: we're targeting a 'stable' release in the next couple of weeks, it's had a ton of churn lately. |
| 13:16 | gtrak | features: clojurescript support, java source navigation, stacktraces, to name a few. |
| 13:17 | gtrak | the 'inspector' |
| 13:17 | gtrak | much more featureful completion. |
| 13:18 | engblom | Is this a good way for setting up cider? http://clojure-doc.org/articles/tutorials/emacs.html |
| 13:19 | engblom | It will install it from marmalade |
| 13:19 | arrdem | yeah that's what you want to do. |
| 13:19 | arrdem | get cider and clojure-mode from marmalade and you should be good to go. |
| 13:20 | arrdem | smartparens and ac-cider are both nice but optional. |
| 13:21 | arrdem | for anyone interested https://github.com/arrdem/dot-emacs |
| 13:23 | mdeboard | Anyone used the "Neocons" lib for neo4j? I'm tryig to sort out how to just send a raw cypher query to the backend, but idk if the trouble i'm having is because of neocons ceremony or if I'm misunderstanding the neo4j cypher tutorial |
| 13:26 | mdeboard | Like is "START" a special thing for Neocons? I have not run across it in the docs yet |
| 13:26 | mdeboard | in the neo4j |
| 13:36 | mdeboard | arrdem: I think ac-cider was breaking up to 0.7.0, something with el-docs |
| 13:37 | mdeboard | (I haven't tried it since upping to 0.7.0) |
| 13:37 | arrdem | mdeboard: maybe, no idea. |
| 13:37 | arrdem | I think my ac-cider is still working but I haven't checked. |
| 13:37 | arrdem | not a feature I leverage heavily. |
| 13:39 | mdeboard | It was pretty nasty, made emacs freeze completely |
| 13:39 | mdeboard | but it seems to be working now, that's good |
| 13:46 | engblom | With cider, how do I get the form I have the cursor inside of to end up at repl? |
| 13:47 | arrdem | M-x cider-eval-last-sexp-to-repl? |
| 13:47 | arrdem | there's probably a shortcut for that but I don't know it. |
| 13:47 | engblom | (foo (bar xy (baz z))) <--- to evaluate (bar...) |
| 13:48 | engblom | And not foo... |
| 13:48 | arrdem | dude I just answered that. |
| 13:49 | engblom | Thanks |
| 13:54 | mdeboard | lol |
| 14:03 | mdeboard | (nn/ |
| 14:06 | engblom | It looks like clojure+emacs+evil is a bad combination. cw (change word) when you have [.... foo] and intending to replace foo, will take away ] also |
| 14:07 | engblom | Somehow it is not understanding [] are not words. |
| 14:08 | catern | that works for me |
| 14:08 | catern | have you changed the handling of words? |
| 14:09 | catern | when I do cw on with my cursor on foo, it eliminates just foo and not ] |
| 14:10 | engblom | No. I was doing the koans I found somewhere and there you need to replace __ with what you think is right. If you have [foo __] it will replace the whole "__]" section. |
| 14:10 | Jaood | the marmalade website used to have a search feature :/ |
| 14:15 | catern | engblom: ah, it's special handling for _ |
| 14:15 | catern | _ is counted as not-a-word |
| 14:17 | catern | engblom: if you want to not count _ as pseudo-whitespace/not-a-word, find the setting for it in evil |
| 14:18 | catern | this is a difference between vim and evil, _ is not counted as part of a word in evil (because it's using emacs's native word detection) and it is in vim, but you can change it to whatever you like |
| 14:22 | allenj12 | hey for lighttable live programming with chrome after i add the script tag a just type Command-ENTER after the expression? |
| 14:23 | engblom | catern: Thanks! |
| 14:23 | allenj12 | it dosnt seem to be working |
| 14:30 | RedOnion | is there any way to read protocol-buffer file in Clojure? |
| 14:31 | RedOnion | like the *pbf files |
| 14:44 | arrdem | cbp: ^ clojure protobufs go |
| 14:50 | cbp | RedOnion: you have to compile those |
| 14:50 | cbp | you can use lein protobuf |
| 14:50 | RedOnion | I did compile them with 'lein protobuf' |
| 14:50 | RedOnion | but I need to read from a .pbf file now |
| 14:52 | cbp | RedOnion: the .pbf file is just a spec that you compile into .java files |
| 14:52 | cbp | and possibly .class files afterwards |
| 14:53 | RedOnion | Well I've the .proto files with the description and they compiled to java files. |
| 14:53 | cbp | lein protobuf should compile them to .class files and place them in /target |
| 14:53 | RedOnion | Now i've a sample.pbf file which is binary, and I need to read it and import it's data |
| 14:56 | RedOnion | Maybe I'm asking the wrong question. I need to read a binary file as Java InputStream |
| 14:58 | gtrak | RedOnion: clojure.java.io/input-stream |
| 14:58 | RedOnion | Ok, let me check that |
| 15:13 | iamjarvo | hi all, so i am trying to understand -> vs ->> and a failure i am getting http://pastie.org/private/psgvzpgwuarzyayfionww can anyone provide some clarity please |
| 15:13 | iamjarvo | total lisp / clojure noob |
| 15:14 | gtrak | collection functions take colls in their last argument, so you probably want to use (->> for those) |
| 15:14 | cbp | iamjarvo: you are doing (drop x 3) when it should be (drop 3 x) |
| 15:14 | iamjarvo | ahh so -> puts it in the vs positionw hereas ->> puts it in the last? |
| 15:15 | iamjarvo | cbp: ^ |
| 15:15 | gtrak | -> is pronounced thread-first. |
| 15:15 | arrdem | iamjarvo: correct. |
| 15:15 | gtrak | ->> is thread-last. |
| 15:15 | iamjarvo | thanks guys |
| 15:15 | iamjarvo | i read something like that but didnt fully grock it |
| 15:21 | fifosine | What do people recommend for a JSON library? |
| 15:22 | gtrak | fifosine: cheshire |
| 15:22 | fifosine | gtrak: Ok, thanks |
| 16:43 | gfredericks | are there cheshire v data.json perf tests results somewhere? |
| 16:45 | MatthiasL | is there a way to let core/async channels communicate over the network? |
| 16:46 | seancorfield | gfredericks: dakrone's repo claims cheshire is 2x faster than data.json |
| 16:48 | seancorfield | Googling doesn't turn up any specific benchmarks that I find, however... |
| 16:49 | gfredericks | I should submit a ticket to data.json that just replaces it with cheshire |
| 17:00 | gtrak | MatthiasL: lots of ways |
| 17:02 | MatthiasL | gtrak: good to know ;) |
| 17:02 | gtrak | pick a transport, someone's probably done it :-) |
| 17:05 | MatthiasL | i'm pretty new to clojure and i'm trying to make a multiplayer game. so far all "clients" are running in the same jvm. now i need to figure out how to do the network communication. kinda clueless about that too. any recommendations on what to look at? |
| 17:05 | justin_smith | MatthiasL: consider making a "command" out of a clojure map, and converting that to json for communication with clients |
| 17:05 | justin_smith | http is probably the most generic choice |
| 17:07 | technomancy | MatthiasL: different games might need different communication styles, but if you can get away with modeling it as request/response, then json over http would be the easiest |
| 17:07 | Glenjamin | does anyone know if there's a way in compojure to perform functions on matched parameters before the route function runs? |
| 17:08 | Glenjamin | like casting to interger, turning :user into a User record etc |
| 17:09 | MatthiasL | json seems like an odd choice for two clojure applications to talk to each other. :p why no edn? |
| 17:09 | justin_smith | Glenjamin: sounds like you would want to wrap "handler" in your function aop style before passing that to ring |
| 17:11 | justin_smith | (fn [handler] (fn [request] (update-in request ...))) - something along those lines |
| 17:11 | justin_smith | where you call handler inside ... of course |
| 17:12 | Glenjamin | mm, essentially middleware |
| 17:12 | justin_smith | right, but wrapping outside compojure's handler |
| 17:13 | Glenjamin | annoyingly, (context) does something similar, but only wraps with a route match, no extra behaviour |
| 17:15 | justin_smith | Glenjamin: it's been ages since I used compojure, but in caribou the default setup has a handler expression to describe stuff that should happen both before and after routing to the controller |
| 17:16 | justin_smith | (this is with caribou's polaris routing lib though, but compojure should not be much different) |
| 17:16 | dbasch | Glenjamin: it is normal to write custom middlewares to do stuff with requests in compojure (e.g. authentication) |
| 17:17 | Glenjamin | yeah, i think i need to just structure everything as middleware |
| 17:18 | Glenjamin | http://weavejester.github.io/compojure/compojure.core.html#var-context |
| 17:18 | justin_smith | what about (context "/" ...) |
| 17:18 | Glenjamin | context looks like what i want, but i'd want to do something with the matched :id before the handlers |
| 17:19 | Glenjamin | under the hood, context is implemented with a wrap-context middleware, but it's marked as private :( |
| 17:25 | zand | hey folks - anyone used "lein new heroku <app-name>" with success? |
| 17:25 | zand | I'm getting an error when trying to do 'lein repl' in that project: |
| 17:25 | zand | https://www.refheap.com/86382 |
| 17:27 | justin_smith | zand: did you actually cd to the shredstir dir? |
| 17:27 | zand | yup - lein downloads the reqs but then hung on JsonFactory |
| 17:28 | zand | I added cheshire explicitly to the the deps and now it's hanging on the "SchemeLayeredSocketFactory" |
| 17:28 | justin_smith | wait, I see nothing about JsonFactory in your paste |
| 17:28 | justin_smith | oh, ok |
| 17:29 | justin_smith | technomancy may be able to help, he works for heroku iirc |
| 17:29 | zand | I'm just following the tutorial at : |
| 17:29 | zand | https://devcenter.heroku.com/articles/getting-started-with-clojure |
| 17:29 | zand | thx for the input justin |
| 17:30 | technomancy | zand: can you run `lein deps :tree` |
| 17:31 | zand | technomancy: https://www.refheap.com/86383 |
| 17:32 | zand | And here's my ~/.lein/profiles.clj: |
| 17:32 | zand | https://www.refheap.com/86384 |
| 17:33 | technomancy | oh, that's a lot of user dependencies |
| 17:33 | zand | I'll try with a clean profiles.clj |
| 17:33 | technomancy | I would try getting rid of leiningen and cd-client, if that doesn't fix it, try bisecting the list |
| 17:34 | technomancy | yeah, it works fine here with my user profile |
| 17:34 | technomancy | I will probably get rid of drawbridge from the heroku template though; it's just not worth the hassle |
| 17:34 | zand | ok - I just mv'd my profiles.clj out of the way |
| 17:35 | zand | did a lein clean && lein deps |
| 17:35 | zand | and it starts the repl fine |
| 17:35 | zand | so it must be something in my profiles.clj that's the problem |
| 17:36 | zand | seems so obvious now that I look at how much stuff I had in there :-) |
| 17:36 | technomancy | yeah, that's the danger with stuff like that |
| 17:36 | technomancy | I try to keep from adding anything with a nontrivial dependency tree to :user |
| 17:36 | zand | technomancy: BTW - thanks for all your great work on Leiningen!! :-) |
| 17:37 | technomancy | glad you like it =) |
| 17:37 | zand | 'lein' is my answer for "what's the killer app for Clojure?" :-) |
| 17:38 | zand | thanks gents - BFN |
| 17:41 | Hodapp | nbeloglazov: https://github.com/overtone/overtone/issues/293 - the issue I mentioned looks like it may be related to Quil, but not strictly a Quil issue, so I filed it there. |
| 17:41 | allenj12 | any benefit of datomic over sql? |
| 17:43 | benkay | is there a SLIME "who-calls" equivalent in Cider? |
| 17:43 | technomancy | benkay: I keep trying to rope people into implementing that |
| 17:43 | technomancy | hasn't worked so far afaik |
| 17:44 | benkay | wouldn't even know where to start... |
| 17:44 | Hodapp | what does who-calls do? |
| 17:44 | technomancy | benkay: tools.analyzer |
| 17:45 | benkay | Hodapp: tells you who calls a function :) |
| 17:46 | allenj12 | https://github.com/swannodette/om/wiki/Advanced-Tutorial best om tutorial? |
| 17:46 | Hodapp | benkay: a dynamic thing, or from static analysis of sorts? |
| 17:48 | benkay | Hodapp: you'd have to ask someone who actually knows how SLIME and Lisps work. |
| 17:48 | technomancy | should be static |
| 17:49 | technomancy | swank-clojure basically just does a grep =( |
| 17:49 | benkay | allenj12: try it and let us know |
| 17:49 | benkay | grep is fast! |
| 17:49 | allenj12 | benkay: its a empty page pretty much lol |
| 17:49 | benkay | related: a friend was doing a stripe ctf, one of the challenges was to index a directory of files, wrote this tidy, somewhat performant node thing, was blown out of the water by his buddy shelling out to grep. |
| 17:50 | Hodapp | hah |
| 17:50 | technomancy | if we had a self-hosting compiler it would be a lot easier to mark every var reference in metadata at compile time |
| 17:51 | bbloom | technomancy: why do you need the whole compiler? isn't tools analyzer enough? |
| 17:52 | technomancy | bbloom: well, I'd rather have a single tool to learn than two =) |
| 17:52 | bbloom | technomancy: you'd be interfacing with the analyzer part of any self hosting compiler anyway, since you don't care about the code generator |
| 17:52 | technomancy | but given that tools.analyzer is basically a prerequisite to self-hosting, yeah it doesn't make a difference |
| 17:52 | andyf | tools.analyzer is becoming part of a self-hosting compiler, or at least I thought that is the plan. Not there yet |
| 17:52 | bbloom | tools.analyzer + tools.emitter essentially == self-hosting compiler |
| 17:53 | technomancy | it was more "I wish this stuff was already done" |
| 17:53 | andyf | understood :) |
| 17:53 | bbloom | technomancy: my understanding is that the jvm analyzer is pretty solid already |
| 17:53 | bbloom | technomancy: of course Bronsa can comment more intelligently about that |
| 17:54 | andyf | Eastwood has a test suite that beats on it fairly heavily, which is where a lot of bugs were discovered back in Dec/Jan time frame. It is pretty stable these days for most code bases. |
| 17:54 | technomancy | it's not done till you can trivially implement who-calls, and no one has implemented who-calls yet, so it must not be done. =) |
| 17:54 | andyf | Still some bugs, I believe, but they are getting harder to find. |
| 17:54 | technomancy | </troll-mode> |
| 17:55 | bbloom | technomancy: so go implement it, it should be a pretty trivial tree walk and just swap! in to an atom in a dynamic var & return that value when you're done |
| 17:55 | Bronsa | uhm well it's now used by eastwood, core.typed and core.async which are fairly big and projects and I haven't received any major bug report for the last months so yeah, I'd say it's pretty stable at this point |
| 17:55 | technomancy | bbloom: I was trying to get benkay to do it for me |
| 17:56 | andyf | bbloom: Um, no offense intended here, but suggesting technomancy is a do-nothing complainer who doesn't implement stuff is something that could leave others with their jaws hanging on the floor. |
| 17:56 | Bronsa | (inc andyf) ;; for the crucible |
| 17:56 | lazybot | ⇒ 3 |
| 17:57 | allenj12 | hey so i went through the om tutorial. i was wondering tho if there is a good open source example of a website both backend and frontend that i can play with? |
| 17:57 | allenj12 | thats somwhat polished |
| 17:57 | bbloom | andyf: i'm certainly not accusing him of complaining |
| 17:57 | benkay | <technomancy> bbloom: I was trying to get benkay to do it for me // you don't want me hacking on clojure toolchain. trust me. |
| 17:57 | justin_smith | allenj12: maybe refheap or 4clojure? |
| 17:58 | bbloom | andyf: it seemed like he didn't realize how close to a reality an analyzer-backed who-calls is |
| 17:58 | justin_smith | allenj12: https://github.com/Raynes/refheap |
| 17:58 | bbloom | andyf: knowing that technomancy is in fact a doer of things, i thought he might just do it, given this new knowledge :-) |
| 17:58 | andyf | bbloom: I withdraw the comment. Misunderstood the thrust of your comment, then. |
| 17:58 | cbp` | justin_smith: i think he wants an Om example |
| 17:59 | Hodapp | Om? |
| 17:59 | technomancy | bbloom: no, you're right. I definitely haven't been keeping up with the changes to tooling land in the past few months. |
| 17:59 | justin_smith | cbp`: ahh, reading comprehension fail |
| 17:59 | Hodapp | last time I heard Om in the context of software was some audio synthesis software drobilla created |
| 17:59 | bbloom | andyf: besides, technomancy even utilized troll mode ;-) |
| 18:00 | justin_smith | Hodapp: hah, I made some ladspa plugins specifically designed to go with Om, I remember that well |
| 18:00 | technomancy | it's all fair game |
| 18:00 | allenj12 | justin_smith: awsome thanks :) and dosnt have to be with om i just wanna see a project that has basically the full package, i just thought om was pretty much used everywhere |
| 18:00 | andyf | Bronsa: Considering adding core.typed to the list of projects in the Eastwood crucible, if I can get the warnings down to a reasonable number. Too noisy right now, and it isn't core.typed's fault. |
| 18:00 | justin_smith | allenj12: om is fairly new |
| 18:00 | Bronsa | technomancy: bbloom I believe arrdem might have already done or will do something like who-calls for his oxcart project, you might want to bug him for that :P |
| 18:01 | justin_smith | and is specifically on the client side, while most clojure stuff deployed is probably still server side |
| 18:01 | arrdem | technomancy/bbloom what's up? |
| 18:01 | technomancy | Bronsa: oh cool. |
| 18:01 | technomancy | another person I can't give a hard time about because he's literally in the middle of making contributions to parts of the toolchain. =) |
| 18:03 | Bronsa | andyf: yeah, I'm a bit busy this period with the GSoC and exams so I don't have time to do that myself but you're more than welcome to report all the bugs you can find :) |
| 18:03 | Hodapp | justin_smith: I'm poking around with Overtone right now... |
| 18:03 | Hodapp | Om was a neat project though |
| 18:03 | allenj12 | justin_smith: o, am i blind, or is there no .cljs in refheap.... |
| 18:03 | Bronsa | andyf: even though tbh core.typed did some funky stuff last time I looked w/ circular dependencies so it might not be all tools.analyzer's fault if it fails :P |
| 18:04 | arrdem | Bronsa: oh who-calls yeah that's on the todo list. I already have calls-what, I just haven't built the mapping inversion. |
| 18:04 | justin_smith | allenj12: I had misenterpreted your question, it is not a cljs project, just clojure |
| 18:04 | justin_smith | sorry |
| 18:04 | justin_smith | Hodapp: cool |
| 18:05 | andyf | Bronsa: Your latest suggested change to Eastwood has, I *think*, gotten it to go through all of core.typed's code without throwing any more exceptions. Now it is a matter of going through the warnings and either fixing small things in core.typed, or figuring out a way to make Eastwood not complain about something that isn't a bug. |
| 18:05 | Bronsa | arrdem: make sure to tell technomancy once you have it working :P |
| 18:05 | arrdem | Bronsa: he can bloody well open an issue if he wants updates :P |
| 18:05 | arrdem | technomancy: ^ |
| 18:05 | allenj12 | justin_smith: o ok i guess i should have been more clear. im still surprised tho there was no need for it? |
| 18:05 | danielcompton | arrdem do you have a turns-down-for-what function planned? |
| 18:05 | justin_smith | allenj12: it's not a js heavy site, apart from some off the shelf libs |
| 18:05 | arrdem | danielcompton: (constantly false), next! |
| 18:06 | justin_smith | hah |
| 18:06 | danielcompton | ,(inc arrdem) |
| 18:06 | clojurebot | #<CompilerException java.lang.RuntimeException: Unable to resolve symbol: arrdem in this context, compiling:(NO_SOURCE_PATH:0:0)> |
| 18:06 | Bronsa | andyf: oh cool |
| 18:06 | danielcompton | ,(inc 'arrdem) |
| 18:06 | clojurebot | #<ClassCastException java.lang.ClassCastException: clojure.lang.Symbol cannot be cast to java.lang.Number> |
| 18:06 | justin_smith | (inc arrdem) |
| 18:06 | lazybot | ⇒ 29 |
| 18:06 | justin_smith | it's a pseudo-syntax |
| 18:06 | danielcompton | No comma? |
| 18:06 | arrdem | danielcompton: we aren't vars in clojurebot yo |
| 18:06 | justin_smith | it's not really clojure, just a bot command triggered by (inc |
| 18:06 | arrdem | s/clojurebot/lazybot/g |
| 18:07 | justin_smith | (inc unbalanced |
| 18:07 | justin_smith | I guess I never tried that before |
| 18:07 | metellus | ) |
| 18:07 | danielcompton | ahh, I thought everyone was a var too |
| 18:07 | bbloom | arrdem: not sure i understand where oxcart fits in. is a bag of extra passes gluing together tools analyzer & emitter into a proper compiler? does emitter not act as a compiler on its own? |
| 18:08 | allenj12 | justin_smith: i guess im not even completely, how much .cljs i need lol so its still a good reference |
| 18:08 | bbloom | arrdem: oh, the caveats seems to make it a bit clearer... it's the *static* subset of clojure? |
| 18:08 | arrdem | bbloom: okay so Oxcart is essentially a research project into what can be done to make clojure run fast |
| 18:08 | bbloom | arrdem: cool |
| 18:09 | arrdem | bbloom: the language subsetting I'm doing is to enable specific performance transforms that are illegal or more difficult when I may need to be able to give you a real class reference for any var |
| 18:09 | bbloom | arrdem: ok, i get it now. very interesting |
| 18:10 | arrdem | bbloom: now. all the infrastructure for the crazy var squashing emitter isn't built yet. Oxcart is, for now, just a bunch of passes that I hope I can build into a compiler server like technomancy described because I think that editor and tooling compiler introspection is a good thing. |
| 18:10 | bbloom | arrdem: i'm curious what you come up with. i haven't gotten a chance to work on eclj much lately, but my intention was to identify a comparable set of fast stuff, but add deoptimization guards to preserve the dynamic functionality where required |
| 18:10 | bbloom | arrdem: but i'm starting with an interpreter, not a compiler |
| 18:11 | bbloom | (so that i have something to deoptimize to!) |
| 18:11 | arrdem | bbloom: I'll be writing a paper on what I find for my undergrad thesis :P also your interpreter is on my "to play with list" because I think it'd be hillarious to implement a dogescript to t.a.jvm frontend and then run it atop clojure with your interp. |
| 18:12 | bbloom | arrdem: godel would be proud |
| 18:12 | Hodapp | s/godel/Gödel/ |
| 18:12 | arrdem | bbloom: I think the real answer is that Oxcart will turn into several optimization targeted t.a(.jvm)? passes that are pre-emission things, and probably two or three different either emitters or emitter modes that can play different tricks compared to the standard t.a.jvm or clojure core emitter. |
| 18:13 | arrdem | how the core team winds up packaging all that I don't know, hence the monolithic oxcart for now. |
| 18:13 | arrdem | </wall> |
| 18:14 | deathknight | Is there a zero-hero guide for making API calls with clj-http or one of its alternatives? |
| 18:14 | bbloom | arrdem: sounds like fun. good luck! i'll keep an eye on it |
| 18:14 | arrdem | bbloom: thanks! I'm just hoping that I manage not to hang myself with scope :P |
| 18:14 | bbloom | psha, naaaaah |
| 18:15 | bbloom | i mean, whole program scheme optimizers are well studied. you just add some fancy pants stuff for clojure-specific bits like pervasive usage of maps and what not, then badda bing badda boom graduate degree |
| 18:16 | arrdem | "bada bing bada boom" |
| 18:16 | arrdem | excuse me I need more beer to appreciate that appropriately. |
| 18:16 | Hodapp | story of my life. |
| 18:16 | arrdem | bbloom: the alternate title for my project is "stealing everything we can from GHC" |
| 18:17 | bbloom | arrdem: ha, nice. |
| 18:17 | bbloom | going to do stream fusion? :-) |
| 18:17 | Hodapp | hey, I found that stream fusion paper interesting! |
| 18:17 | arrdem | 0_0 ghc is magic shit |
| 18:17 | danielcompton | deathknight the tutorial for http://http-kit.org is pretty good |
| 18:17 | Bronsa | callen is going to be proud |
| 18:17 | bbloom | ghc's optimizer is pretty impressive |
| 18:17 | deathknight | daniel, thanks a bunch. checking it out now |
| 18:18 | arrdem | Bronsa: hehe he's still trying to convince me to just drop the project and join the dark side. |
| 18:18 | augustl | anyone got a favorite path matcher/router for building single page web apps with om or something else? :) |
| 18:24 | nbeloglazov | Hodapp: I'll take a look |
| 18:25 | Hodapp | nbeloglazov: I'd dig deeper myself but I don't even know where to start :-/ |
| 18:25 | arrdem | I should amend my summer schedule to include two weeks to just read & evaluate GHC papers... |
| 18:25 | Hodapp | nbeloglazov: but, like the issue says, I can load just Quil and not have the issue |
| 18:26 | nbeloglazov | Hodapp: I've commented on the issue |
| 18:27 | arrdem | I think that the real answer is that when all is said and done Bronsa, tbc++, ambrosebs and I need to just sit down and figure out what could even be done to leverage types and to mitigate structural sharing so that the compiler can introduce transients. |
| 18:27 | Hodapp | oh, I'd not realized Quil 2.0.0 was a thing |
| 18:31 | andyf | arrdem: Do you mean, so the compiler can introduce transients automatically, when it can prove it is safe, without the programmer having used transients? |
| 18:32 | arrdem | andyf: yep, that's something I'd like to do research on when the school year starts up again. |
| 18:32 | arrdem | waaay out of scope for GSoC |
| 18:33 | mdeboard | 7 |
| 18:33 | mdeboard | ) |
| 18:33 | mdeboard | gah sorry |
| 18:35 | arrdem | hum... looks like eastwood does a ton of what I'm doing already... |
| 18:35 | arrdem | well. some. |
| 18:36 | andyf | I didn't realize |
| 18:36 | arrdem | andyf: like the def in def stuff, insanities which normal clojure supports that I have to detect as well. |
| 18:37 | arrdem | although I forget what I decided to do with def in def.. |
| 18:37 | andyf | The nice thing about implementing Eastwood is that we don't do anything fancy about it -- just detect and report it. |
| 18:37 | Bronsa | andyf: some recent additions to ta/taj could make some of the pattern matching code for eastwood way easier btw |
| 18:37 | arrdem | andyf: true that, but it'd be nice if the detection code could be shared. |
| 18:38 | andyf | Bronsa: I'm open to suggestions. |
| 18:38 | andyf | Bronsa: But by that statement, I do not mean to pressure you into implementing any of them :) |
| 18:39 | Bronsa | andyf: e.g. defonce-or-defmulti-macro-expansion? could be written as (some (fn [form] (and (seq? form) ('#{defonce defmulti} (first form)))) (:raw-forms node)) |
| 18:39 | arrdem | hehe |
| 18:39 | arrdem | yeah I need to start using that... |
| 18:40 | andyf | Bronsa: Thanks. I will leave myself an issue to remind me. |
| 18:41 | andyf | Bronsa: Eventually Eastwood will be t.a(.jvm) plus about 30 lines of code :) |
| 18:41 | arrdem | andyf: this is a good thing. |
| 18:41 | Bronsa | andyf: http://sprunge.us/TgcI?clj |
| 18:41 | arrdem | Oxcart will one day share this fate |
| 18:43 | Bronsa | andyf: well it means I must be doing something right then :P |
| 18:44 | Bronsa | arrdem: if you need :raw-forms be sure to use 0.2.1/2-SNAPSHOT btw 0.2.0 had some passes that discarded it |
| 18:44 | arrdem | Bronsa: does t.e.jvm support definline? |
| 18:45 | arrdem | Bronsa: thanks for the heads up I'll do new local installs tomorrow. |
| 18:45 | Bronsa | arrdem: it should |
| 18:45 | arrdem | Bronsa: but by var as a value or by inlining? |
| 18:45 | arrdem | 'cause if the former I need to build the latter. |
| 18:46 | Bronsa | arrdem: both last I checked, if not it's a bug |
| 18:46 | arrdem | kk. good to know. |
| 18:46 | Bronsa | arrdem: supporting :inline-fn is the only thing that makes t.a.j *require* using c.c/eval |
| 18:46 | technomancy | deleting code is the best thing ever |
| 18:47 | Bronsa | which I don't like, but it's the only way to split the analyzer and the emitter |
| 18:47 | Hodapp | hah |
| 18:47 | Bronsa | in Compiler.java analysis and emission are intertwined |
| 18:48 | Hodapp | deleting code, though, would not have gone over well in many environments where I've developed, where money was involved... |
| 18:49 | andyf | Bronsa: I see :raw-forms is a sequence of forms. In general, is it always 2 forms, or is it some varying number depending upon how many macroexpansions it takes to hit a fixed point? |
| 18:49 | Bronsa | andyf: the latter |
| 18:50 | arrdem | Bronsa: hum. okay. |
| 18:53 | allenj12 | is it weird after going through several tutorials, im still not completely sure how much of a website i should actually do in html and how much in clsj for front end |
| 19:05 | allenj12 | also what is a good lein template for a website, i asked this before and liked the answer but totally forgot which one it was |
| 19:05 | andyf | You can check the logs: http://logs.lazybot.org/irc.freenode.net/%23clojure |
| 19:07 | allenj12 | o cool |
| 19:07 | allenj12 | andyf: thanks |
| 19:15 | gtrak | allenj12: probably luminus |
| 19:31 | Janiczek | Hi guys, where to put JARs for Leiningen to see them? |
| 19:31 | clojurebot | Cool story bro. |
| 19:32 | nDuff | Janiczek, your local Maven repository -- which is where "mvn install" and "lein install" will put things by default |
| 19:33 | nDuff | Janiczek, ...if you aren't using something that speaks Maven, (1) you should, but (2) for the meantime, look at the lein-localrepo plugin. |
| 19:34 | technomancy | nDuff: as of 2.3 or so you actually don't need localrepo |
| 19:34 | nDuff | Ahh -- then I'm out-of-date. |
| 19:35 | technomancy | clojurebot: lein-localrepo is now unnecessary since lein deploy's second arity can deploy jars from disk directly. |
| 19:35 | clojurebot | Ik begrijp |
| 19:37 | Janiczek | So if I have a jar that isn't in any maven repo, what's the easiest way for my Clojure project to see it? Some lein command, or just right path? |
| 19:39 | Janiczek | I have a ~/.m2 if that's where it should be... |
| 19:44 | Janiczek | Maybe it was unneeded but I managed to do something with https://www.pgrs.net/2011/10/30/using-local-jars-with-leiningen/ |
| 19:56 | andyf | Bronsa: Looking at new with-env in t.a(.jvm) 0.2.1, including example atom called e1 in the test suite referred to in the namespace docs. Should Eastwood be using a similar thing, perhaps without any mappings for "foo" in that example? |
| 19:57 | andyf | Here in particular: https://github.com/clojure/tools.analyzer/blob/master/src/test/clojure/clojure/tools/analyzer/core_test.clj#L45 |
| 19:58 | Bronsa | andyf: t.a.j has a global-env function, you might want to (env/with-env (taj/global-env) ALL-THE-ANALYSIS-HERE) |
| 19:58 | Bronsa | andyf: to be clear, tej will work even if you don't do that |
| 19:59 | andyf | thx. Reading... |
| 19:59 | Hodapp | Overtone is working wonders to remind me that I know little about Clojure and less about music theory. |
| 19:59 | Bronsa | but I'd rather if you did that, cljs will require it so it's more consistent to do it for taj too |
| 20:01 | andyf | ALL-THE-ANALYSIS-HERE would be analyzing one form, or the loop for read/analyze/emit/eval ? Currently Eastwood creates a fresh empty env for each analyze call on each top level form. |
| 20:02 | Bronsa | andyf: http://sprunge.us/VjXg?clj it should look like this |
| 20:03 | Bronsa | andyf: analyze/analyze+eval take care of providing the analyze-env for you, there's no need to do that manually now |
| 20:04 | Bronsa | add a recur there |
| 20:04 | andyf | Also, Eastwood is currently using a custom run-passes modified from what is in taj, to avoid errors that occur in some of the passes. I am guessing I should continue doing that. |
| 20:06 | Bronsa | andyf: let me look at the code for a second |
| 20:12 | andyf | I've read taj/analyze+eval, and it looks a lot like eastwood.analyze-ns/analyze-file, except for all the extra debug printing in analyze-file. I think I can see how to update eastwood.analyze-ns/analyze-file to work like analyze+eval, leaving the extra debugs in. |
| 20:13 | andyf | If I tried using taj/analyze+eval, I suspect I would end up copying it and adding in the extra debug printing, anyway. |
| 20:13 | Bronsa | andyf: yeah I thought about updating eastwood to analyze+eval, the only problem is that currently there's no way to retrieve the AST of e.g. (throw (Exception. "")) with analyze+eval |
| 20:13 | Bronsa | I might change that in the next release though |
| 20:14 | Bronsa | andyf: anyway, you can replace eastwood's analyze with http://sprunge.us/NIIH now |
| 20:15 | Bronsa | andyf: wrap the loop of analyze-ns with (env/with-env (ana.jvm/global-env) ..) and you're ready to go with 0.2.1 |
| 20:16 | Bronsa | andyf: but if you wait a couple of days for 0.2.2 you might get an analyze+eval that you can work with |
| 20:17 | andyf | Should your example http://sprunge.us/NIIH have #'ana.jvm/run-passes in place of ana.jvm/run-passes, or as you wrote it? If it is as you wrote it, I didn't realize clojure.core/binding could do that. |
| 20:17 | andyf | Sure, I can wait :) |
| 20:17 | Bronsa | andyf: yeah binding works like that |
| 20:17 | Bronsa | it's with-bindings that takes a map of Var -> value |
| 20:20 | antonv | hi |
| 20:20 | Bronsa | andyf: I'll probably go with http://sprunge.us/SLXh?diff but I'll sleep over it to see if there's a better approach |
| 20:21 | antonv | The :arglists metadata of a funtion var holds the function signature (argument list) |
| 20:21 | antonv | Can I rely on the :arglist name? |
| 20:21 | antonv | Or it is a private Clojure metadata attribute, and I can't expect it will never change? |
| 20:21 | antonv | I have a macro, which expands into a function definition |
| 20:22 | antonv | I want the function defined by the marco to have proper argument list metadata |
| 20:23 | antonv | I don't see anywhere in Clojure docs, that :arglist is guaranteed to always work like this |
| 20:24 | Bronsa | antonv: :arglists is documented http://clojure.org/special_forms#def so it's not private |
| 20:24 | Bronsa | antonv: unfortunately nothing enforces :arglists to be in the correct format '([..]*) |
| 20:24 | Bronsa | and some libraries overwrite :arglists to something different, which really annoys me. |
| 20:26 | technomancy | Bronsa: ouch, like what? |
| 20:27 | antonv | Bronsa: I would call such libraries unfriendly |
| 20:27 | Bronsa | technomancy: like, for single arity functions [args..] rather than ([args..]) |
| 20:27 | Bronsa | antonv: agreed. |
| 20:27 | technomancy | oof |
| 20:27 | technomancy | actually I change :arglists, but only for partially-applied lein tasks |
| 20:27 | technomancy | I wish c.c/partial would do the same |
| 20:28 | Bronsa | technomancy: java.jdbc was doing that and I believe seancorfield changed that after I/andyf asked him to do so |
| 20:29 | andyf | Several macros in core do it, for documentation purposes. No functions that I can recall |
| 20:30 | andyf | core.typed has a fun thing where it chose to name the first arg &form, and clojure.core/defn removes the first two args when that happens, I guess because it assumes it must be the function "beneath" a macro that takes &form and &env as the first 2 args. |
| 20:30 | Bronsa | technomancy: the thing that annoys me the most is that because it's not enforced, tools like eastwood cannot use :arglists to warn users about invoke wrong arities at compile time without the risk of warning about false positives |
| 20:31 | andyf | Hiccup's defelem does it. |
| 20:32 | andyf | got to go. Thanks for all the info, Bronsa. |
| 20:32 | Bronsa | andyf: np |
| 20:33 | danielcompton | I'm working on a problem where I need to find if an array contains any point where the sum of the left parts of the array = the sum of the right parts of the array, excluding the element that you are on. I thought this would be a perfect use of a zipper and reduce |
| 20:33 | Bronsa | technomancy: Andy made a ml post asking for introducing :doc-arglists or something like that so that :arglists could be enforced and one could use :doc-arglists for documentation purposed but we got no response from the core team :/ |
| 20:34 | danielcompton | A naive implementation of (= (reduce lhs) (reduce rhs)) works but gets very slow when the vector gets to 10000 elements |
| 20:35 | danielcompton | I was thinking of building a binary tree with values at each node for the sum of the elements below it. Is there a clojure function/datastructure similar to this or do I need to build it myself? |
| 20:39 | hyPiRion | danielcompton: You can do that even easier: The most efficient way is to just use an array storing cumulative sums, and calculate diffs based on that. |
| 20:39 | Jachy | danielcompton: I think my approach would be to make two prefix sum arrays, one from the left and one from the right. |
| 20:40 | hyPiRion | danielcompton: If you need to do real time updates on the array, have a look at Fenwick trees/BIT |
| 20:40 | danielcompton | hyPiRion what would the array look like then? I need to calculate it at each point in the list |
| 20:40 | hyPiRion | danielcompton: if you have [1 2 3 4], the cumulative vector is [1 3 6 10] |
| 20:40 | Jachy | http://en.wikipedia.org/wiki/Prefix_sum |
| 20:41 | danielcompton | hyPiRion And then do a reverse array? |
| 20:41 | danielcompton | [4 7 9 10] |
| 20:41 | hyPiRion | danielcompton: there's no need for a reverse array |
| 20:41 | danielcompton | hyPiRion yeah, just realised that |
| 20:42 | hyPiRion | so I'd guess the idiomatic clojure algorithm creates the cumulative vector with reductions |
| 20:42 | danielcompton | So a successful test case is [1 2 3 3] |
| 20:42 | hyPiRion | ,(reductions + [1 2 3 4]) |
| 20:42 | clojurebot | (1 3 6 10) |
| 20:42 | danielcompton | Where 1+2 = 3 seen from the third element in the vector |
| 20:43 | danielcompton | ,(reduction + [1 2 3 3]) |
| 20:43 | clojurebot | #<CompilerException java.lang.RuntimeException: Unable to resolve symbol: reduction in this context, compiling:(NO_SOURCE_PATH:0:0)> |
| 20:43 | danielcompton | ,(reductions + [1 2 3 3]) |
| 20:43 | clojurebot | (1 3 6 9) |
| 20:44 | danielcompton | I don't quite see how this helps, do I take 6-3 = 9-6? |
| 20:44 | hyPiRion | danielcompton: the sum of the previous elements are simply the previous element (3). The sum of the remaining elements are the total sum (9) minus the cumulative sum so far (6) |
| 20:45 | danielcompton | hyPiRion: it all becomes clear |
| 20:45 | hyPiRion | nice |
| 20:45 | danielcompton | hyPiRion thanks |
| 20:45 | hyPiRion | np |
| 20:49 | augustl | are clojurescript atoms available as a stand-alone js lib? |
| 20:50 | augustl | using mori in a js app and noticing that I'm also missing atoms :) |
| 20:51 | danielcompton | hyPiRion I just went from 90 seconds with 10k elements to 0.5 seconds |
| 20:51 | hyPiRion | heh, the power of algorithms |
| 20:51 | danielcompton | who knew 50+ years of computer science had something to offer |
| 20:52 | hyPiRion | strange, eh |
| 20:53 | augustl | O(n) path matching is my pet peeve these days.. Almost every URL dispatcher I know about is O(n) over each defined route |
| 20:53 | Hodapp | danielcompton: the flip side is that, when you are a developer at a job that is using the State of the Art(tm), and you read a 50-year-old Comp Sci paper that is addressing a simple, practical solution to a problem that you realize is still basically unsolved within the company, it tends to be depressing. |
| 20:53 | technomancy | Bronsa: is it really that bad if you have false positives around terrible code? |
| 20:54 | technomancy | I mean, for slamhound I made a point of saying "we can provide benefits X, Y, and Z, but only if you have well-formed code that doesn't do anything too crazy" |
| 20:54 | danielcompton | Hodapp there is that too. Are you thinking of any paper in particular? |
| 20:54 | Hodapp | nope. |
| 20:54 | technomancy | so if you have a c.c/load call in your code, it's not going to be able to help you; you're on your own, but it doesn't seem like too much to ask. better than trying to be all things to all people. |
| 20:56 | hyPiRion | Hodapp: I don't think it's depressing. There's just so much out there it's impossible to know everything. It's a bit sad that people don't attempt to check if the problem's already solved, though |
| 20:56 | Hodapp | hyPiRion: that latter part was what depressed me. |
| 20:57 | hyPiRion | Hodapp: alright. Then I agree. Sad world =( |
| 20:57 | Hodapp | hyPiRion: It wasn't, "someone didn't just know out of the blue that they were dealing with a well-studied problem," but rather "someone lacked the humility to even think that maybe, maybe someone else had looked at this problem, even after being beaten over the head with the paper" |
| 20:58 | Hodapp | But, on the bright side, there are certainly some gems in those early papers, and I find it very interesting to see the foresight the authors had. |
| 20:58 | Bronsa | technomancy: I'd agree with you, but it's really annoying when you have to explain the users of a library you're trying to push that it's not your new library that's broken but the official-contrib-db-layer they've been using for 2 years that's doing something weird but not really illegal |
| 20:59 | Hodapp | I've been meaning to read, "Augmenting Human Intellect: A Conceptual Framework" from Doug Engelbart. |
| 21:18 | technomancy | Bronsa: yeah. gotta draw the line somewhere, but I suspect it's probably harder in a new tool that people are still forming opinions about. |
| 21:26 | kenrestivo | the commercial world of software was for a long time very divorced from the academic world. the culture of the bill gates'es who dropped out of college to become millionaires, etc. it seems that the two worlds are starting to merge again, which is a good thing. |
| 21:27 | Hodapp | kenrestivo: there was something of an anti-intellectual culture around the software development at the last job. |
| 21:28 | kenrestivo | indeed. it was once pervasive through much of the industry, as i recall it. |
| 21:28 | kenrestivo | maybe google and others are changing that culture. maybe it's just that the world is more complex, and the problems are harder to solve, and a solid academic background is not optional anymore? |
| 21:28 | Hodapp | we had a contractor - who was let go - who had a rather stubborn, vocal view that whatever I did in my spare time that pertained to software development was worthless, academic, and something that only a nerd would care about. |
| 21:29 | kenrestivo | Hodapp: hopefully you've landed in a more friendly company since then. |
| 21:30 | p_l | kenrestivo: actually, Google is getting rid of bias towards degrees in new hires :> |
| 21:30 | Hodapp | they got rid of him, but yes, I've landed elsewhere too |
| 21:30 | kenrestivo | p_l: the pendulum swinging a bit in the other direction, eh? |
| 21:31 | Hodapp | I don't particularly care if I get to use Clojure at $job, but I do care if it's a job where people turn their noses up at the mere idea of Clojure and take a sort of "Oh, it's one of THOSE languages that only neckbeard academics understand" stance. |
| 21:32 | p_l | kenrestivo: they identified a too strong focus on university and grades as detrimental |
| 21:34 | justin_smith | Hodapp: I take that kind of attitude as defensiveness / insecurity |
| 21:34 | p_l | kenrestivo: mind you, from my pov, software industry was full of "you must have degree or we won't even talk with you" |
| 21:34 | p_l | (at least in bigger places) |
| 21:34 | justin_smith | not to say I haven't seen, for example, similar attitude about Haskell here. |
| 21:35 | kenrestivo | p_l: i've noticed that in recent years. 20-30 years ago it was quite the opposite. |
| 21:35 | p_l | justin_smith: TBH, a certain *group* (not everyone) of people who associate with haskell tend to propagate that view |
| 21:35 | hellofunk | Hodapp just tell them "I want to use this Java library called Clojure." |
| 21:39 | Hodapp | hellofunk: sneaky :P |
| 21:39 | hyPiRion | "It's a dsl for java very suitable for our needs" |
| 21:42 | platz | 501 developers (http://thedevstop.wordpress.com/2012/04/19/reflections-on-the-501-developer-manifesto/) can be hostile to those that put in time outside of work hours, because it threatens their lifestyle |
| 21:43 | platz | if (b) codes all night, will I have to as well, to keep up? |
| 21:43 | Hodapp | hyPiRion: There was a huge inertia, not just in terms of attitudes but in terms of technical barriers, on most things there. As a result of using languages with crappy abstraction, and then trying to tack on Java support, we had 4 or 5 kind of half-assed ad-hoc minilanguages. These minilanguages were used, largely, as templates for automatic generation of code. |
| 21:44 | Hodapp | And in order to add basically anything, it had to go through all of these gigantic, inflexible layers where there simply was no abstraction. |
| 21:44 | platz | also relevant, pretty much everything written by michaelochurch |
| 21:45 | justin_smith | hellofunk: we did that, and it worked out great, until the client tried to run a static analysis tool because another vendor had deployed an app with a security hole, and discovered that what we deployed was simply impossible for a java sa tool to analyze |
| 21:45 | justin_smith | then the jig was up |
| 21:45 | justin_smith | and we also lost some trust from the client |
| 21:45 | TimMc | platz: I have a contract with my employer that buys them 40 hours of my attention each week. Anything on top of that is at my pleasure. Other employees may do as they wish. |
| 21:46 | TimMc | (They can also sometimes borrow extra time per week, but only with interest...) |
| 21:46 | Hodapp | justin_smith: was that to me? |
| 21:46 | justin_smith | Hodapp: it wasto hellofunk who was replying to you above |
| 21:47 | justin_smith | re: "java lib called Clojure" |
| 21:48 | justin_smith | regarding the proliferation of half-assed minilanguages, yeah that is the price people pay for lisp-phobia |
| 21:48 | Hodapp | it wasn't even lisp-phobia, it was a phobia to all things not C++. |
| 21:49 | Hodapp | one developer had created the "AutoCode Generator" and was really proud of it. |
| 21:49 | justin_smith | I am sure, since it was IH and all |
| 21:49 | Hodapp | and I couldn't help but see it as nothing but a set of crutches that let us hobble along on a broken language for a little bit longer |
| 21:49 | TimMc | platz: I do some work on my own time for an open source project that I also work on for my employer, but that's because it's important to me. |
| 21:50 | platz | calling C++ a broken language? hrmm |
| 21:50 | danielcompton | ,(partition 2 1 (reductions + [1 2 3 4 5 6 7 8 9 4])) |
| 21:50 | clojurebot | ((1 3) (3 6) (6 10) (10 15) (15 21) ...) |
| 21:51 | danielcompton | how do I get ,(partition 2 1 (reductions + [1 2 3 4 5 6 7 8 9 4])) and also tack on the last value of reductions, 49 in this case |
| 21:51 | Hodapp | platz: completely broken for the sorts of burdens being put on it. |
| 21:51 | danielcompton | Is it lazy-cat? |
| 21:51 | Hodapp | but with everyone having a strong emotional attachment to it. |
| 21:52 | technomancy | hard to believe there are strong emotions around C++ that aren't hatred and loathing |
| 21:52 | justin_smith | danielcompton: maybe partition-all? |
| 21:52 | danielcompton | justin_smith, I'm wanting it to look like ((1 3 49) (3 6 49) (6 10 49))... I don't think partition-all would give me that? |
| 21:53 | justin_smith | oh, misunderstood your request |
| 21:53 | hyPiRion | ,(let [sums (reductions + [1 2 3 4 5 6 7 8 9 4]) total (last sums)] (->> (partition 2 1 sums) (map (fn [[a b]] [a b total])))) |
| 21:53 | danielcompton | justin_smith It wasn't phrased particularly well, rereading it |
| 21:53 | clojurebot | ([1 3 49] [3 6 49] [6 10 49] [10 15 49] [15 21 49] ...) |
| 21:53 | hyPiRion | not sure if there's any better way |
| 21:54 | danielcompton | hyPiRion thanks, I was hoping there was a single golden clojure function that would solve it all for me |
| 21:54 | TimMc | technomancy: Except for wistfulness for the days when nobody did anything important enough on computers to have to worry about sploitz. |
| 21:55 | TimMc | Ah, those were the days... |
| 21:56 | hyPiRion | danielcompton: unfortunately that magic function is not always available, so better use what you got/know |
| 21:56 | danielcompton | hyPiRion yeah, got to work with what I have |
| 21:59 | fifosine | Can anyone vouch for or against using pedestal for single page applications? |
| 22:02 | justin_smith | fifosine: pedastal is kind of experimental, from what I have seen |
| 22:02 | justin_smith | it has some cool stuff, but it is a little "weird" |
| 22:02 | fifosine | do you have any recommendations for frameworks for building single page apps? |
| 22:02 | justin_smith | compojure / cljs / om should work pretty nicely |
| 22:03 | justin_smith | well, skip compojure if it is truly single page maybe, and just use httpkit+ring/ cljs / om |
| 22:04 | fifosine | justin_smith: Since I'm new to web dev with clojure, where do I start with incorporating those libraries together? There seems to be a lack of literature |
| 22:04 | justin_smith | lein new om-starter should give you all of that |
| 22:05 | justin_smith | or something close enough |
| 22:05 | justin_smith | I mean lein new om-starter project-name |
| 22:07 | justin_smith | trying that myself, the only thing missing is the http-kit ring backend, which is not hard to set up once you have the default template |
| 22:11 | fifosine | justin-smith: http-kit deals with the server logic/ |
| 22:11 | fifosine | ? |
| 22:11 | justin_smith | it implements the http server |
| 22:11 | justin_smith | which you interact with in your code via the ring library |
| 22:12 | justin_smith | (ring means you can switch it out with other servers trivially) |
| 22:12 | fifosine | would I need to incorporate these for a dev environment? |
| 22:12 | justin_smith | well, http-kit implements web sockets (which are often useful for single page apps) and jetty does not |
| 22:12 | justin_smith | so maybe, if you need the feature |
| 22:13 | justin_smith | otherwise, http-kit is probably the best server for production for performance reasons |
| 22:34 | danielcompton | fifosine I wouldn't recommend it unless you're working with Cognitect as well |
| 22:42 | platz | I wonder if it would be possible to use that crazy fast 'undertow' framework that keeps trouncing the techempower benchmarks with clojure |
| 22:44 | justin_smith | http://undertow.io/ this one? |
| 22:45 | danielcompton | platz why would it not be possible? |
| 22:45 | ddellacosta | platz: https://github.com/piranha/ring-undertow-adapter |
| 22:53 | platz | wow cool |
| 23:02 | technomancy | I'm trying to see how the upcoming lein release will perform as far as startup time; could some folks run this on their own machine and report results? http://p.hagelb.org/bench-startup |
| 23:10 | cbp | technomancy: https://www.refheap.com/86388 |
| 23:12 | benkay | i'm sure there's a function similar to any which checks a predicate over a coll, but instead checking a set of predicates over a value |
| 23:13 | benkay | forgive me for blanking... |
| 23:13 | cbp | (doc every-pred) |
| 23:13 | clojurebot | "([p] [p1 p2] [p1 p2 p3] [p1 p2 p3 & ps]); Takes a set of predicates and returns a function f that returns true if all of its composing predicates return a logical true value against all of its arguments, else it returns false. Note that f is short-circuiting in that it will stop execution on the first argument that triggers a logical false result against the original predicates." |
| 23:17 | benkay | sweet! thanks, cbp |
| 23:21 | Frozenlock | Is there a way to manually login a user in friend? Changing the ring session and fiddling with :cemerick.friend/identity ? |
| 23:21 | ddellacosta | Frozenlock: do you need it for testing or as part of the real app flow? |
| 23:21 | benkay | Frozenlock: in the context of looking at your app in realtime? |
| 23:22 | Frozenlock | real app |
| 23:22 | Frozenlock | I'm creating an account when the user pay with credit card. I'd be nice if he would be logged in automatically after that. |
| 23:24 | benkay | something like (assoc response :session {:logged-in true}) ? |
| 23:25 | benkay | Frozenlock: you may also want to dig in and learn about how sessions work in Ring. |
| 23:25 | Frozenlock | I know how it works, I'm asking because it feels a little primitive. :-/ |
| 23:26 | beppu | technomancy: https://gist.github.com/beppu/5cf0d636ca78359aefe6 |
| 23:26 | Frozenlock | I was hoping to have missed a friend function "log-user" or something like that... |
| 23:26 | technomancy | cbp, beppu: thanks |
| 23:27 | benkay | Frozenlock: have you read this? https://github.com/cemerick/friend-demo |
| 23:27 | Frozenlock | benkay: Not recently, might be worth to look again. Thanks |
| 23:28 | benkay | reading the source for that will answer like 90% of your friend questions |
| 23:28 | benkay | and raise 50 more... |
| 23:30 | Frozenlock | yeah... while friend might be the best for clojure right now, it feels like walking on eggs. |
| 23:30 | benkay | consider writing your own handler for logged-in-ed-ness |
| 23:32 | hellofunk | Frozenlock walking on eggs? do you mean fragile or unstable? Friend is used quite a bit in production. |
| 23:32 | Frozenlock | I think friend/merge-authentication is what I need. https://github.com/cemerick/friend/blob/master/src/cemerick/friend.clj#L65 |
| 23:33 | Frozenlock | hellofunk: Yes, that's what I mean. |
| 23:33 | Frozenlock | Not that it is, but it *feels* |
| 23:33 | benkay | hellofunk: lol running in production is *not* a guarantor of quality in my experience. |
| 23:33 | benkay | look at how much php there is in the wild for starters ;) |
| 23:33 | hellofunk | Frozenlock well Chas knows what he's doing and a lot of people have kicked around that library. |
| 23:34 | Frozenlock | I know. I wouldn't be using it otherwise. |
| 23:38 | seanaway | technomancy: re: Bronsa's comment about java.jdbc - it still has "bad" :arglists but it has a JIRA ticket to fix them so Eastwood can make sense of them |
| 23:50 | danielcompton | technomancy my bench results https://www.refheap.com/86389 |