2010-03-06
| 00:09 | brandonw | if you have a situation where you are using map and need to define an anonymous function, but you already have one in a higher scope, is it more idiomatic to define that first anonymous function as a private function in the same namespace, or as a function defined within the scope of the top level public function you are working on |
| 00:09 | brandonw | s/but you already have one/but you already have an anonymous function |
| 00:10 | brandonw | (just to be clear) |
| 00:15 | slyphon | uh, where do i get swank-clojure.jar |
| 00:15 | slyphon | i followed the directions here: http://github.com/technomancy/swank-clojure |
| 00:16 | rickmode | slyphon: that's the right place |
| 00:17 | slyphon | yeah, it says to copy swank-clojure.jar to the lib/ directory if you want to use the project feature |
| 00:17 | slyphon | but i'm retarded tonight and can't seem to find that jar file |
| 00:17 | rickmode | slphon: this helped me too: http://www.bestinclass.dk/index.php/2009/12/clojure-101-getting-clojure-slime-installed/ |
| 00:17 | slyphon | rickmode: thanks :) |
| 00:18 | rickmode | slyphon: I was referred that video by another person here in this channel within the last week ;) |
| 00:19 | slyphon | heh |
| 00:19 | slyphon | oh, bastard |
| 00:20 | rickmode | watch the video anyway - there's lots of good bits there |
| 00:20 | slyphon | cool |
| 00:20 | rickmode | it kinda explains the "why" of it all |
| 00:21 | slyphon | ah |
| 00:21 | slyphon | i've been SLIMEd before, but yeah, any info is useful |
| 00:22 | rickmode | i'm still not entirely clear on the differences between slime and swank... but I got my emacs, clojure and leiningen stuff working smoothly now |
| 00:23 | slyphon | yeah, the slime/swank difference was always a bit of a mystery, i think swank is a lisp-impl-specific server that answers requests from slime |
| 00:23 | slyphon | or, something like that |
| 00:24 | rickmode | ya - that's my sense too - UI vs. server stuff |
| 00:25 | slyphon | 10 years of muscle-memory is not easily undone |
| 00:26 | rickmode | tell me about it... but I'm coming from an IDE background (winblows and java IDEs) |
| 00:26 | slyphon | ah |
| 00:27 | rickmode | i still get baffled by the frames/buffers/windows stuff in emacs - it's got its own *way* |
| 00:27 | tomoj | slyphon: just use elpa |
| 00:27 | tomoj | it's pretty simple http://tomojack.com/2010/02/emacs-for-clojure-from-scratch/ |
| 00:27 | slyphon | tomoj: yeah, i did, i needed to copy the swank-clojure.jar into an existing project |
| 00:28 | tomoj | lein? |
| 00:28 | clojurebot | leiningen is a build tool designed not to set your hair on fire (http://github.com/technomancy/leiningen) |
| 00:28 | slyphon | didn't realize it put it in .swank-clojure |
| 00:28 | tomoj | if your project needs swank-clojure, I think it should be in the dependencies for the build tool |
| 00:28 | tomoj | it's pretty easy to work with a lein project through emacs |
| 00:29 | slyphon | well, i'm not really using a build tool right now |
| 00:29 | tomoj | why not? too much work? |
| 00:29 | slyphon | eh, i don't come from a java background, so it was a high barrier to entry for me to start thinking about leingen or maven or w/e |
| 00:29 | tomoj | I think I will be making a screencast soon with emacs+lein |
| 00:29 | tomoj | yeah |
| 00:29 | slyphon | i just put my jars in lib/ |
| 00:29 | tomoj | it can be a bit intimidating |
| 00:30 | slyphon | and wrote a Rakefile to spawn nailgun w/ vim-clojure |
| 00:30 | tomoj | but I think technomancy has really got it down so that it's easy to do |
| 00:30 | slyphon | technomancy is a sharp cookie |
| 00:30 | slyphon | [sic] |
| 00:30 | tomoj | oh, I kind of assumed you were using emacs :) |
| 00:30 | slyphon | yeah, i switched |
| 00:30 | slyphon | well |
| 00:31 | slyphon | that makes it sound really official |
| 00:31 | slyphon | i used emacs for about a year back in 2005 or so |
| 00:31 | slyphon | i wanted to see where it was at |
| 00:31 | tomoj | and now you're a vim dude? |
| 00:31 | slyphon | i've always been a vim dude :) |
| 00:32 | tomoj | well, there's a lein-nailgun plugin |
| 00:32 | tomoj | I dunno how well it works, though |
| 00:32 | slyphon | i was working on an all-emacs team though at the time, and i got tired of everyone always giving me shit about it |
| 00:32 | tomoj | haha |
| 00:32 | tomoj | I didn't know there were all-emacs teams who gave shit about it |
| 00:32 | slyphon | well |
| 00:32 | slyphon | heh |
| 00:32 | slyphon | it was twisted-python |
| 00:32 | tomoj | if my coworkers start trying to use vim for clojure, I'd give them shit |
| 00:32 | tomoj | but not aggressively :) |
| 00:32 | slyphon | and it was my first real gig, and i kind of wanted to fit in |
| 00:33 | tomoj | yeah makes sense |
| 00:33 | tomoj | well, maybe I can convert you with my screencast |
| 00:33 | slyphon | now that i'm not trying to use emacs for IRC and reading email, and all that other nonsense, i'm interested primarily because i remember SLIME being *really* impressive with sbcl |
| 00:33 | slyphon | plus, messing around with elisp is kind of fun |
| 00:34 | tomoj | yes, I've done a couple fun things with elisp |
| 00:34 | tomoj | I've made it so the editor can be aware of which namespace a source file defines |
| 00:34 | slyphon | ah, that's handy |
| 00:35 | tomoj | so you can get a skeleton ns decl from the path, or cause the repl to switch to the namespace in the current buffer, for example |
| 00:35 | tomoj | but I feel like there is much more you could do |
| 00:35 | tomoj | (and a lot I love already works with slime) |
| 00:35 | slyphon | context-awareness is a really powerful thing |
| 00:35 | tomoj | e.g. M-. will jump to the function definition, even if the clj source file is buried in a jar |
| 00:35 | slyphon | see, *that* stuff is really nifty |
| 00:35 | tomoj | slime macroexpansion stuff works, some documentation stuff |
| 00:36 | tomoj | plus, ...paredit |
| 00:36 | slyphon | yeah, srsly |
| 00:36 | tomoj | this is what my screencast will cover |
| 00:36 | slyphon | oh, excellent |
| 00:36 | slyphon | i look forward to watching it |
| 00:36 | tomoj | but I can't stop the pain of learning (or using) emacs |
| 00:36 | tomoj | :( |
| 00:36 | slyphon | hahaha |
| 00:36 | slyphon | well, i've got the basics down |
| 00:37 | slyphon | and having vim-mode (which is light-years better than Viper) definately helps |
| 00:37 | slyphon | i just need hjkl |
| 00:37 | tomoj | I suppose you don't need much more than the basics |
| 00:37 | slyphon | and :e :w |
| 00:37 | tomoj | unless you want to live in emacs |
| 00:37 | slyphon | the 'C-w j' bindings are good too |
| 00:38 | slyphon | that's one thing that always drove me insane about emacs, i didn't think the split-handling was nearly as straightforward |
| 00:38 | slyphon | "Why did that open up down there? |
| 00:38 | slyphon | " |
| 00:38 | tomoj | hmm |
| 00:38 | tomoj | with paredit? |
| 00:39 | slyphon | nah, just with standard emacs buffer stuff |
| 00:39 | tomoj | I don't notice anything that sounds like that with my setup |
| 00:39 | tomoj | but I don't really understand what you mean |
| 00:39 | slyphon | it could just be some insane combination of plugins i was using back in the day |
| 00:39 | slyphon | it seems much more intuitive now |
| 00:39 | slyphon | with this setup i have |
| 00:40 | tomoj | I think there was some plugin that made it so while you're editing, the parens are exploded c-style, but when you are done they are redone to be lisp-style (e.g. '))))))' at the end of a line) |
| 00:40 | tomoj | but I don't find the need for that |
| 00:41 | slyphon | argh, one thing i wish i could figure out is how to globally define the friggin font i want EVERYWHERE |
| 00:55 | tomoj | aquamacs is :/ |
| 00:55 | tomoj | I'm not really sure what to do about emacs on osx |
| 00:55 | tomoj | almost want to just run it in Terminal.app, but I hate Terminal.app |
| 00:56 | danlarkin | I live inside console emacs in iTerm |
| 00:57 | slyphon | i usually run everything in mrxvt |
| 00:57 | slyphon | well, anything that has to do with "work", but the meta key in X11 on OS-X is totally hosed, iirc |
| 01:00 | offby1 | newbie question -- just built my program with leiningen. When I try to run it with 'clj anagrams.jar "Your Mother Wears Army Boots"', I get 'Exception in thread "main" java.lang.Exception: No such namespace: ' (followed by a bunch of binary garbage). What might be wrong? |
| 01:01 | tomoj | danlarkin: will check iTerm out, thanks |
| 01:01 | tomoj | I hadn't considered trying to get rxvt to work on mac |
| 01:02 | slyphon | i require jmk-neep |
| 01:02 | slyphon | well, that too |
| 01:02 | slyphon | usually non-concurrently though |
| 03:03 | ennui | have a public Java class 'Bar' in package 'foo'. (import 'foo.Bar) succeeds. How can I now call its (public static) main method? |
| 03:04 | derefed | (foo.Bar/main) I think |
| 03:04 | derefed | that's just a guess though |
| 03:06 | tomoj | I imagine you will need to pass a string[] in, no? |
| 03:07 | tomoj | (foo.Bar/main (into-array String [])) |
| 03:11 | ennui | derefed, tomoj, thanks. the last one works perfectly. i was struggling with how to properly pass the string argument. |
| 03:11 | derefed | np =) |
| 04:13 | dmiles | http://libomv.pastebin.ca/1824907 <- am i runing the pastebin line 109 every "new" ? |
| 04:14 | nteon | so if i have ((a b) (c d) (e f)), how can i get (a b c d e f)? I want to use reduce but my brain is tired |
| 04:15 | gregh | ,(flatten '((a b) (c d) (e f))) |
| 04:15 | clojurebot | (a b c d e f) |
| 04:17 | nteon | gregh: ah! cool |
| 04:17 | nteon | thanks |
| 04:22 | nteon | ,(doc flatten) |
| 04:22 | clojurebot | "([x]); Takes any nested combination of sequential things (lists, vectors, etc.) and returns their contents as a single, flat sequence. (flatten nil) returns nil." |
| 04:22 | hiredman | ,`flatten |
| 04:22 | clojurebot | clojure.contrib.seq-utils/flatten |
| 04:25 | nteon | hiredman: yes, that was the command I was looking for, thanks :) |
| 04:35 | zab | Evening all. I'm trying to get Enlive working on App Engine. GAE docs say I can read files using Class.getResource() or ServletContext.getResource(). NFI how to do that from Clojure. Any pointers? |
| 04:45 | tomoj | zab: (.getResourceAsStream (clojure.lang.RT/baseLoader) "filename") |
| 04:46 | tomoj | maybe that's what you're looking for |
| 04:46 | zab | tomoj: thanks! |
| 05:41 | tomoj | what is the point of having (:load) in the ns macro? |
| 05:42 | tomoj | I mean, I understand what it does, I just don't understand why we need it |
| 05:42 | tomoj | so that we can split a namespace into multiple files easily? |
| 10:01 | ned | forgive my stupidity but whats the proper way to 'clone' a lein repo? |
| 10:01 | ned | i cant quite figure it out and i've searched relatively extensively |
| 10:02 | tcrayford | use git clone? |
| 10:02 | ned | sorry the repos hosted on clojars |
| 10:02 | ned | just git clone http://clojars.org/repo/name ? |
| 10:07 | ned | oh it would seem i have to create some boilerplate code, before lein deps. |
| 10:11 | tcrayford | yep |
| 10:11 | tcrayford | in project.clj |
| 10:16 | rsh | is there a good way to tell emacs to indent custom macros a certain way? particularly when dealing with macros that contain a parameter like (& body)? |
| 10:18 | vy | rsh: It should indent them right, that's how (with-foo ...) macros get indented also. |
| 10:23 | rsh | I see. with-foo macros seem to work fine, but any arbitrary name doesn't like (defmacro =bind [parms & body]) |
| 10:27 | vy | Hrm... That feature might be implemented in recent versions of SLIME, Clojure uses an obsolete version. |
| 10:27 | vy | But I'm pretty sure SLIME is capable of indenting by looking at function's arguments. |
| 11:33 | jkr | If I want to use clojure-mode from ELPA with my own clojure.jar in /opt/clojure (as opposed to downloading a new one), what should I set? |
| 11:33 | jkr | I used to be able to set swank-clojure-jar-path, but it seems to have changed in the months since I last used it. |
| 12:18 | joshua-choi | Performance-wise, which is better: conj!-ing many times onto a vector transient, or cons-ing many times onto a list? |
| 12:24 | rsh | what is the clojure equivalent to CL's atomp function? |
| 12:27 | boojum | joshua-choi: seems to be transient is faster |
| 12:28 | joshua-choi | Okay |
| 12:28 | joshua-choi | Thanks |
| 12:31 | boojum | joshua-choi: http://gist.github.com/323814 |
| 12:32 | joshua-choi | Great, thanks for running that. |
| 12:35 | krumholt_ | rsh, probably it's (not (coll? x)) |
| 12:36 | rsh | thanks |
| 12:37 | krumholt_ | rsh, but thats not quite it i think a function is not an atom in CL but it is not a coll in clojure either |
| 12:37 | dnolen | krumholt_ rsh: I believe seq? is more equivalent to atomp |
| 12:38 | krumholt_ | ,(seq? [1 2 3]) |
| 12:38 | clojurebot | false |
| 12:41 | krumholt_ | ,(seq? (list 1 2 3)) |
| 12:41 | clojurebot | true |
| 12:43 | krumholt_ | ,(map #(not (or (coll? %) (fn? %))) [1 "foo" 'foo (list 1 3) [1 3] (fn [x] x) {:a 1} #{1 2 3}]) |
| 12:43 | clojurebot | (true true true false false false false false) |
| 12:43 | krumholt_ | i think (not (or (coll? x) (fn? x))) should be atomp |
| 12:44 | rsh | is that kind of calculation not common in clojure? |
| 12:44 | krumholt_ | rsh, you mean for recursion? |
| 12:44 | rsh | ye |
| 12:44 | rsh | yes* |
| 12:45 | dnolen | krumholt_: interesting when I wrote clj-cont I brought it this very topic, but I don't recall why I decided seq? was better than coll? |
| 12:45 | dnolen | brought -> brought up |
| 12:45 | krumholt_ | rsh, i used it a lot in commonlisp almost never in clojure. i ceased to use symbols that much |
| 12:46 | krumholt_ | dnolen, i think seq? checks if it implement ISeq like lists. coll is almost everything |
| 12:47 | krumholt_ | dnolen, look at this beautiful class diagram :) http://github.com/Chouser/clojure-classes/blob/master/graph-w-legend.png |
| 12:48 | krumholt_ | dnolen, you can see very good what coll? checks for and what seq? checks for |
| 12:49 | dnolen | yeah, again I don't remember why I chose seq?. It was a for a code walker. |
| 12:50 | dnolen | code is from almost a year ago |
| 12:50 | dnolen | I remember intending to use coll? |
| 12:50 | dnolen | but bringing it up on IRC and changing my mind. |
| 12:52 | rsh | if recursing through a list representing a tree is uncommon in clojure, what are the alternatives |
| 12:53 | technomancy | rsh: I don't think it's particularly uncommon. have you used tree-seq? |
| 13:08 | rsh | technomancy: I have not heard of tree-seq, i will look into it more. We were discussing earlier the clojure equivalent of atomp and landed on (or #(coll? %) #(fn? %)) . Does this look right to you? |
| 13:09 | rsh | CL's atomp |
| 13:10 | offby1 | newbie question -- just built my program with leiningen. When I try to run it with 'clj anagrams.jar "Your Mother Wears Army Boots"', I get 'Exception in thread "main" java.lang.Exception: No such namespace: ' (followed by a bunch of binary garbage). What might be wrong? |
| 13:10 | technomancy | rsh: isn't atomp just "is this not a compound data type?" it's been a while since my last CL usage. |
| 13:10 | rsh | technomancy: yes |
| 13:11 | technomancy | offby1: does your project.clj specify the :main namespace? |
| 13:11 | rsh | technomany: clojure not having a this included seems odd to me, like it is implying you should not be doing this much |
| 13:11 | technomancy | offby1: the JVM has some odd rules wrt translating between dashed-namespaces and files_on_disk.clj |
| 13:11 | rsh | technomany: not having this* |
| 13:11 | technomancy | rsh: if you want to know if something is not a collection, you can just call (not (coll? x)) |
| 13:12 | offby1 | technomancy: lemme check |
| 13:12 | rsh | technomany: yeah, but a fn would return true there even though it shouldn't |
| 13:13 | offby1 | technomancy: it says nothing about :main |
| 13:13 | offby1 | I let leiningen create it, and didn't edit it further |
| 13:15 | technomancy | offby1: oh, I see. yeah, your clj script is probably expecting to be pointed at a .clj file instead of a jar |
| 13:16 | offby1 | oh? |
| 13:16 | technomancy | you'd want to run jar files using the "java" command-line launcher, but the rules about that are quirky too |
| 13:17 | technomancy | you need to either build it as a standalone jar, or include all the deps on the classpath |
| 13:17 | offby1 | technomancy: aha, you are correct, sir |
| 13:17 | offby1 | clj ./src/anagrams/core.clj yo ho ho => happiness |
| 13:17 | technomancy | and you'd also need to specify a :main namespace containing a -main function |
| 13:17 | technomancy | right... command-line launching is definitely a weak point of clojure |
| 13:19 | technomancy | offby1: you can also try the leiningen-run plugin |
| 13:19 | offby1 | oh hell -- next problem is my program runs out of heap space. Is there a way I can increase it with clj, or do I just need to cruft up a shell script that directly invokes java on the .jar? |
| 13:19 | offby1 | .oO("plugin"?) |
| 13:19 | technomancy | where'd you get your clj script? there's not really a standard one yet. |
| 13:20 | offby1 | ClojurX |
| 13:20 | technomancy | offby1: looks like the one in contrib takes a CLOJURE_OPTS env var |
| 13:20 | technomancy | CLOJURE_OPTS="-Xmx1g" clj [...] |
| 13:21 | technomancy | if yours is based on that you could try that |
| 13:21 | technomancy | JAVA_OPTS is also a common one |
| 13:21 | technomancy | (bbiab) |
| 13:23 | offby1 | tx |
| 13:27 | LauJensen | Evening team |
| 13:39 | joshua-choi | I think I asked this before, but I don't remember the answer. Is there a high-level function mystery so that (mystery f g) is equivalent to (fn [& args] (and (apply f args) (apply g args)))? |
| 13:41 | slyphon | joshua-choi: you could write that macro pretty easily |
| 13:42 | joshua-choi | Yeah. Just wondering if there was a standard one. |
| 13:46 | texodus | Anyone bored who wants to give me some feedback on a Clojure library I wrote to wrap Netty? http://github.com/texodus/saturnine |
| 16:25 | greybeard | Is there One True Swank/Slime Configuration document? Particularly in the use of swank-clojure-extra-classpaths? |
| 16:29 | greybeard | My code runs fine from the command line - but not in slime. It doesn't seem to find the jars as I get the "No suitable driver found for jdbc:mysql://..." |
| 16:43 | dnolen | greybeard: install via ELPA and you'll probably find it easier to get working estup |
| 16:43 | dnolen | estup -> setup |
| 16:45 | greybeard | I did use ELPA. But I prefer to have less magic and more understanding. |
| 16:48 | greybeard | In .emacs I have something like: (setq swank-clojure-extra-classpaths (list "/Users/edglas/jars/" "/opt/jars/")) |
| 17:16 | bsteuber | why is there a difference between concat and lazy-cat, though concat states the seq is lazy? |
| 17:18 | hiredman | one of them is a macro so it can make things lazier |
| 17:19 | bsteuber | hmm |
| 17:19 | hiredman | ,(first (lazy-cat [1] (println :x))) |
| 17:19 | clojurebot | 1 |
| 17:19 | hiredman | ,(first (lconcat [1] (println :x))) |
| 17:19 | clojurebot | java.lang.Exception: Unable to resolve symbol: lconcat in this context |
| 17:19 | hiredman | ,(first (concat [1] (println :x))) |
| 17:19 | clojurebot | 1 |
| 17:19 | clojurebot | :x |
| 17:20 | bsteuber | ,(take 4 (concat (repeat 1) [2])) |
| 17:20 | clojurebot | (1 1 1 1) |
| 17:21 | bsteuber | i see |
| 17:21 | bsteuber | the there is only a difference when side effects are involved? |
| 17:21 | bsteuber | or is there a pure example where they differ? |
| 17:22 | hiredman | if the things you are concating are all lazy it makes no difference |
| 17:22 | hiredman | well |
| 17:23 | hiredman | lazy-cat wraps things in an additional layer of laziness |
| 17:23 | bsteuber | so if I have a 5000000 element array as first argument, will concat traverse it eagerly? |
| 17:23 | hiredman | no |
| 17:23 | bsteuber | good |
| 17:23 | hiredman | concat doesn't traverse |
| 17:24 | bsteuber | ok, so it is usually "lazy enough" |
| 17:24 | hiredman | but if you have a function that generats a 50000 element list, and (concat (f) [1]) because concat is a function f will be evaluated |
| 17:24 | hiredman | er |
| 17:24 | hiredman | (f) |
| 17:25 | hiredman | the arguments to functions are evaluated and the result is passed |
| 17:25 | bsteuber | ok, that makes sense |
| 17:25 | hiredman | lazy-cat essential generates a thunk of each argument and that thunk is not called until it is needed |
| 17:26 | bsteuber | so I will just remember that not even function calls are evaluated |
| 17:26 | bsteuber | thank you :) |
| 17:28 | chouser | disconnecting to rearrange my network at home. please refrain from saying anything interesting until the logger is back online. |
| 17:28 | chouser | :-) |
| 17:36 | bsteuber | I'm wondering why replicate is there as well as repeat |
| 17:36 | chouser | bsteuber: historical reasons. I would expect replicate to go away |
| 17:38 | bsteuber | chouser: ic - so maybe it should be marked depricated? |
| 17:38 | chouser | sounds good to me |
| 17:40 | rhickey | ticket/patch welcome |
| 17:46 | bsteuber | oh, didn't know it is spelled deprecated |
| 17:47 | bsteuber | rhickey: ok, sounds like fun and a good reason to start learing about your assembla stuff |
| 17:47 | bsteuber | though it's just a patch of one line saying "DEPRECATED" :) |
| 17:54 | hiredman | ,(doc *asert*) |
| 17:54 | clojurebot | Gabh mo leithscéal? |
| 17:54 | hiredman | ,(doc *assert*) |
| 17:54 | clojurebot | "; " |
| 17:54 | tomoj | what's that for? |
| 17:55 | hiredman | good question |
| 17:55 | bsteuber | ,*assert* |
| 17:55 | clojurebot | true |
| 17:55 | hiredman | ,(doc assert) |
| 17:55 | clojurebot | "([x]); Evaluates expr and throws an exception if it does not evaluate to logical true." |
| 17:56 | tomoj | aha |
| 17:57 | tomoj | so bind *assert* to false to skip assertions |
| 17:58 | hiredman | (when *assert* (binding [*out* *err*] (println "X is depricated, use Y"))) |
| 17:58 | hiredman | pre |
| 18:00 | bsteuber | .(assert false) |
| 18:00 | bsteuber | ,(binding [*assert* false] (assert false)) |
| 18:00 | clojurebot | java.lang.Exception: Assert failed: false |
| 18:01 | bsteuber | hmm |
| 18:04 | hiredman | ,(binding [*assert* false] *assert*) |
| 18:04 | clojurebot | false |
| 18:07 | hiredman | ,(binding [*assert* false] (macroexpand '(assert false))) |
| 18:07 | clojurebot | nil |
| 18:08 | hiredman | it must be some kind of quirk of clojurebot's sandbox |
| 18:08 | hiredman | or not |
| 18:22 | bsteuber | hiredman: I think the binding is done at run-time while the macroexpand already expands at compile-time |
| 18:23 | bsteuber | so *assert* is still true at this point |
| 18:23 | hiredman | mm |
| 18:25 | bsteuber | so how can we rebind *assert* at compile-time? |
| 18:25 | hiredman | write or find a binding macro |
| 18:26 | bsteuber | not sure if it's that easy |
| 18:26 | hiredman | why not? |
| 18:28 | bsteuber | ,(defmacro with-production-mode [& body] (binding [*assert* false] `(do ~@body))) |
| 18:28 | clojurebot | DENIED |
| 18:31 | bsteuber | (with-production-mode (assert false)) |
| 18:31 | bsteuber | this fails |
| 18:32 | bsteuber | because the binding is only there when the `(do) is returned, but not when it's macroexpanded |
| 18:34 | bsteuber | we had to do sth like that |
| 18:34 | bsteuber | (defmacro with-production-mode [form] (binding [*assert* false] (macroexpand form))) |
| 18:34 | bsteuber | but explicit macroexpand is ugly |
| 18:34 | hiredman | ,(X [*assert* false] (assert false)) |
| 18:34 | clojurebot | nil |
| 18:35 | hiredman | but that is rather horrible |
| 18:35 | bsteuber | ,(doc X) |
| 18:35 | clojurebot | "([[name value] & code]); " |
| 18:40 | bsteuber | yeah, why not |
| 18:41 | bsteuber | in CL you would do this kind of thing with (eval-when (compile) ...) |
| 18:42 | bsteuber | but the CL stuff is doing side effects, so you can't transfer this to clojure easily |
| 18:42 | hiredman | (defmacro X’ [x & code] `(binding ~x (eval '~@code))) |
| 18:43 | hiredman | there really should be a do in there somewhere |
| 18:45 | bsteuber | how about eval and the environment? |
| 18:45 | bsteuber | it won't work in a let, right? |
| 18:45 | hiredman | right |
| 18:46 | hiredman | it's hardly perfect |
| 18:46 | hiredman | if you really want to you can always alter-var-root |
| 18:47 | bsteuber | somehow it stays true |
| 18:48 | bsteuber | maybe we're not supposed to alter the vars from core |
| 18:48 | hiredman | oh |
| 18:49 | hiredman | direct binding maybe |
| 18:49 | bsteuber | ,`macroexpand-all |
| 18:49 | clojurebot | clojure.walk/macroexpand-all |
| 18:50 | hiredman | and the repl must bind it |
| 18:50 | hiredman | ,(set! *assert* false) |
| 18:50 | clojurebot | java.lang.IllegalStateException: Can't change/establish root binding of: *assert* with set |
| 18:50 | hiredman | but clojurebot doesn't |
| 18:50 | bsteuber | (defmacro in-prod [& code] (binding [*assert* false] `(do ~@(macroexpand-all code)))) |
| 18:50 | bsteuber | ,(defmacro in-prod [& code] (binding [*assert* false] `(do ~@(macroexpand-all code)))) |
| 18:50 | clojurebot | DENIED |
| 18:51 | bsteuber | that does the job |
| 18:52 | bsteuber | though I don't know whether it is okay to macroexpand-all on arbitrary code |
| 18:52 | bsteuber | but I guess it's better that eval :) |
| 18:54 | bsteuber | why did clojurebot deny my in-prod macro? |
| 18:55 | hiredman | clojurebot doesn't allow defs |
| 18:56 | bsteuber | hm okay :) |
| 20:05 | ichernetsky | Hi. Is there possibility to gen-class some class that has some java annotation? |
| 20:20 | danlarkin | ichernetsky: I don't think clojure can do annotations yet |
| 20:20 | ichernetsky | danlarkin: Is it going to be added in the near future? |
| 20:21 | danlarkin | not to my knowledge |
| 22:48 | mikem | hi, if I have a function with arguments such as [first & rest], in the body of the function rest is a list. how can I pass that list to another function which accepts arguments the same way (ie: [firt & rest])? |
| 22:52 | nteon | ,(doc apply) |
| 22:52 | clojurebot | "([f args* argseq]); Applies fn f to the argument list formed by prepending args to argseq." |
| 22:52 | nteon | mikem: so (apply some-fn rest) |
| 22:56 | mikem | nteon: I'm actually trying to reply to this email on the mailing list: http://groups.google.com/group/clojure/browse_thread/thread/ea67c317dafb47df |
| 22:56 | mikem | in the last posting, the author wraps (struct-map) in his own function, passing in & init, but when inits is passed to struct-map it's a single list, not a variable number of arguments. |
| 23:00 | nteon | mikem: I would guess its because he's returning the struct 'sm', not the modified one returned from assoc |
| 23:02 | mikem | actually, it's the call to (struct-map ) which fails. the correct way to call struct-map is: (struct-map st :a 1 :b 2) ... inside his my-struct-map, it ends up being called (struct-map st (:a 1 :b 2)) |
| 23:02 | mikem | so I'd like to "expand" that sequence (:a 1 :b 2) before passit it to the real struct-map |
| 23:02 | mikem | s/passit/passing/ |
| 23:06 | nteon | mikem: (apply struct-map s inits) |
| 23:06 | nteon | is the particular change |
| 23:07 | nteon | ah, I misread that if statement |
| 23:07 | mikem | nteon: ok, that works |
| 23:07 | mikem | thanks |
| 23:08 | mikem | yeah, the if statement is correct |
| 23:08 | nteon | mikem: glad I could help |
| 23:31 | hiredman | http://gist.github.com/324145 my first chunked seq |
