#clojure log - Jan 06 2012

The Joy of Clojure
Main Clojure site
Google Group
IRC
List of all logged dates

0:50 jodaro: wow

0:50 lamina is really nice

1:42 alexbaranosky: do any of you use lazytest?

1:43 do you keep your tests in the same namespaces as your production code? I say that, because it seems like if you modify a production file, it won't run the test that goes with it, unless that test is in the same ns...

2:07 replaca: technomancy_: you there?

2:09 TimMc: looks cool!

2:09 I like the ns comment best

2:14 alexbaranosky: is there any annotation support in Clojure? I want to try using PowerMock, but I think it requires using some annotations

2:25 hiredman: alexbaranosky: yes, deftype has some kind of annotation support, not sure if anyone has used it

2:26 alexbaranosky: hiredman, thanks, I'm playing with adding support for mocking static methods, but the JAva library I was thinking of building off of requires using annotations. More research needed

5:39 true_droid: what's the most efficient way to append elements to the end of a sequence?

5:39 Say, I have a function which takes a sequence and modifies it in some way. But the order of elements has to be preserved.

5:39 Multiple alternatives come to mind:

5:39 1) start with an empty vector [], conj elements into it and then return (seq my-vec)

5:39 2) start with an empty list (), conj elements into it and then return (reverse my-list)

5:39 3) like the previous one, but use (into () my-list) in the end

5:39 4) start with a list (), append each consecutive element to it with (concat my-list (list element))

5:42 philandstuff: not 4), that's for sure

5:42 afaik each concat is linear in terms of my-list

5:42 morphling: true_droid: concat is lazy, so that is good, but do you really have to add items one by one?

5:42 philandstuff: it is?

5:42 ,(doc concat)

5:43 clojurebot: "([] [x] [x y] [x y & zs]); Returns a lazy seq representing the concatenation of the elements in the supplied colls."

5:43 philandstuff: morphling: ooh ta

5:44 true_droid: hm, I haven't thought about it

5:44 so, it depends on the use-case; whether I want the whole sequence at once or not

5:44 in each case efficiency is defined in a different way

5:46 philandstuff: morphling: so if you build up a sequence using concat, when you call first on the resulting sequence won't that be a linear operation, because it has to go through n subsequences?

5:47 true_droid: morphling: if I'm writing a compress function which removes consecutive duplicates from a sequence, is there a way to build a new list more than one element at a time?

5:49 as far as I get it, seq doesn't construct a new collection, it's merely an interface

5:49 does the same go for 'reverse'? does it reuse the existing collection?

5:54 morphling: philandstuff: yes, but you don't have to traverse the entire first sequence

5:54 philandstuff: yes

5:54 morphling: philandstuff: so it's linear in the amount of concat calls, but not the length of the first sequence

5:54 philandstuff: yeah

5:54 in this situation, the two were equal, but it ain't necessarily so

5:55 so 4) is still a bad idea, but not for my original reason

5:55 morphling: yes

5:55 true_droid: you might want to use lazy-seq (or perhaps for)

5:59 true_droid: ok, thanks, I'll look into these

5:59 *those

6:01 morphling: this is about the only case where I really miss lazy evaluation.. in haskell you'd just use reduce

8:34 AWizzArd: http://i.imgur.com/pAy4z.png

8:52 bsteuber: AWizzArd: rofl

9:04 AWizzArd: (:

9:46 mcrittenden: given these two factorial defns, why would the first one throw a stack overflow and the 2nd one works fine when (factorial 10000N) is called? https://gist.github.com/1570886

9:47 tsdh: mcrittenden: It consumes stack because it has to remember n in every recursion step.

9:48 mcrittenden: The recur version is in fact compiled to some iterative algorithm without recursion.

9:48 rabbler: The second uses recur, which gets compiled into what is basically equivelant to a "for loop"

9:48 Iterative is a much better response, thanks tsdh.

9:48 pandeiro: When I try to instantiate a Java class, I get a "No matching ctor found for class ...", so I am using clojure.reflect/reflect to examine the class and it seems to have an eponymous method -- that signifies a constructor in Java, right?

9:48 mcrittenden: interesting, thanks for the explanation tsdh and rabbler

9:49 tsdh: mcrittenden: Google for clojure recur, and you'll find a bunch of more elaborate explanations.

9:49 mcrittenden: tsdh yeah doing that now :)

9:57 pandeiro: Ah so Java throws the no ctor found just because the arity is off... I like clojure's error message better

10:38 AWizzArd: $seen jchoi

10:38 lazybot: I have never seen jchoi.

10:51 mcrittenden: what is a ref? http://clojure.github.com/clojure/clojure.core-api.html#clojure.core/ref is that the same thing as a managed reference? this is a concept I'm unfamiliar coming from python

10:53 TimMc: mcrittenden: http://clojure.org/refs

10:54 mcrittenden: TimMc yeah found that but it's still a little fuzzy for me. so the idea is just that you don't have to worry about different threads having a different value for it, whereas with vars you do?

11:00 TimMc: never mind, after reading through that page a couple more times, I think I mostly get it

11:02 this helped as well: http://en.wikipedia.org/wiki/Software_transactional_memory

11:13 pjstadig: a "managed reference" could be an atom, a var, or a ref

11:13 a ref is a very specific kind of "managed reference", which is a bit confusing

11:14 mcrittenden: pjstadig: yeah that tripped me up.

11:16 TimMc: mcrittenden: Once you understand 'ensure, you're probably good.

11:58 clj_newb: why is there defn- but neither def- nor defmacro- ?

12:20 TimMc: clj_newb: More need for defn-, I guess.

12:21 You can add ^:private as needed for the others.

12:23 cgray: and it's pretty easy to write the macro for def- and defmacro- just by looking at defn-

12:30 pandeiro: how does one handle a map key with a space in it?

12:31 hiredman: ,({"foo bar" 1} "foo bar")

12:31 clojurebot: 1

12:32 pandeiro: ,({:character count 1} ":character count")

12:32 clojurebot: #<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: Map literal must contain an even number of forms>

12:32 `fogus: ,(keyword "a b")

12:32 clojurebot: :a b

12:33 hiredman: `fogus: :( don't go doing that, people might think it's a good idea

12:33 `fogus: ,((keyword "a b") {(keyword "a b") :eville})

12:33 clojurebot: :eville

12:33 `fogus: It is a great idea!

12:34 hiredman: :(

12:34 `fogus: Although I prefer newlines and Klingon chars in my keys

12:34 pandeiro: yeah not my idea, you work with the libs you got...

12:38 DeusExPikachu: is there anything in enlive (or anything else) that converts a string to a url with the characters escaped?

12:39 ckirkendall: maybe: http://docs.oracle.com/javase/1.5.0/docs/api/java/net/URLEncoder.html

12:41 (URLEncoder/encode url)

12:42 DeusExPikachu: ckirkendall: seems to work, thanks

12:43 ckirkendall: other than the obvious of posting to the IRC whats the best way to get the word out on a new library?

12:43 DeusExPikachu: ckirkendall: the mailing list I believe is the official place

12:43 ckirkendall: google-groups?

12:44 DeusExPikachu: ckirkendall: yes, the mailing list, which is the list on google-groups

12:45 ckirkendall: DeusExPikachu: I will put a post out there

12:46 In case anyone is interested I created and enlive like dom manipulation library for clojurescript. It has all the standard transforms plus event handling and effects.

12:47 Here is the demo site: http://ckirkendall.github.com/enfocus-site/

12:49 cgray: very cool

12:53 `fogus: ckirkendall: Have you looked at domina?

12:54 ckirkendall: `fogus: no

12:55 `fogus: https://github.com/levand/domina

12:55 ckirkendall: `fogus: similar concept but the semantics are bit diffrent

12:55 `fogus: On first glance your lib seems more mature

12:56 ckirkendall: `fogus: I was able to handle all of the standard enlive transformation

12:56 `fogus: I then had to add all the stuff to deal with a dynamic dome

12:56 dom

12:56 `fogus: it looks nice.

12:57 ckirkendall: `fogus: I am working on adding a good abstraction for extracting data from the dom

12:57 `fogus: thanks

13:00 pandeiro: is there a :use-macros in clojurescript?

13:01 technomancy: replaca_: what's up?

13:02 arohner: does anyone remember seeing a new library for mongoDB? I seem to remember a new lib that claims to be better than congomongo, but I can't remember the name

13:03 ckirkendall: `fogus: bundled with something like the hiccup implementation in pinot for dom creation it makes a powerful tool for creating UI widgets.

13:04 benares_98: arohner: do you mean adia?

13:05 pandeiro: ckirkendall: have you seen hiccups?

13:05 arohner: benares_98: I don't think so. what I'm thinking of is purely a mongo library, no web stuff

13:06 ckirkendall: pandeiro: I have now nice!

13:06 jodaro: arohner: there are a couple others on github

13:06 arohner: jodaro: yes, but what are their names? or how do I find them? :-)

13:06 jodaro: monger

13:06 karras

13:07 arohner: jodaro: monger, that's what I was thinking of! thanks

13:07 jodaro: arohner: http://bit.ly/wFeyGX

13:07 pandeiro: ckirkendall: there was a project that was sort of backbone-inspired too but it's been subsumed by something else

13:07 jodaro: i just did a search for mongo with language:clojure

13:08 arohner: jodaro: thanks. I was using google, and not being very successful

13:24 zilti: Netsplit?

13:31 solussd: why does this return a vector of two items? (re-find #"([r\-][w\-][x\-]){3}" "rwxr-xr-x")

13:32 arohner: solussd: it returns the whole match, and then the parenthesized group

13:32 solussd: ok, is there a way to say, in the regex, that I want that group 3 times without having it be a 'match group' ?

13:33 arohner: yes

13:33 (?:)

13:33 http://docs.oracle.com/javase/1.4.2/docs/api/java/util/regex/Pattern.html

13:36 mr_rm: can someone tell me what the '-' means at the front of a function name, like (defn -main ...)?

13:37 solussd: in a genclass it is a method

13:37 clj_newb: TimMc: thanks for the tip on ^:private (re: defn- vs def- defmacro-)

13:37 solussd: (the default prefix is '-', yu can change it though)

13:39 mr_rm: solussd: i saw someone use it on a generic function that was not using using genclass. they just defined a standalone function with the '-' in front of the name and i was wondering if hit had any special meaning in that context

13:39 https://github.com/yogthos/Noir-Eclipse-Template-Project/blob/2a91fcd8b131137e51362f6ffb3b92c4d56b349e/src/my_website/server.clj

13:41 solussd: mr_rm afaik it doesn't mean anything special, but maybe it does in the context of the noir.server library

13:42 mr_rm: solussd: ok thanks

13:45 technomancy: I love how the "not marked as dynamic" warnings don't use fully-qualified names.

13:50 yazirian: that was just irritating me too, not ten minutes ago

13:50 clj_newb: ,(type identity)

13:50 clojurebot: clojure.core$identity

13:50 clj_newb: ,(type identity) ; why does this not return function, and how can I check is a var is a function?

13:50 clojurebot: clojure.core$identity

13:51 clj_newb: ,(meta identity)

13:51 clojurebot: nil

13:52 jodaro: clj_newb: try fn?

13:52 clj_newb: jodaro: nice; thanks

13:54 metajack: clj_newb: you might also want ifn? if you want things like :foo and {} to return true

13:55 clj_newb: i can imagine :foo is a function which extracts a label

13:55 but how is {} a function?

13:55 metajack: maps in clojure are functions that take keys and return values

13:56 TimMc: clj_newb: In Clojure, functions are implemented as class-per-function. identity is an instance of the identity class.

13:56 metajack: ,({:foo :bar} :foo)

13:56 clojurebot: :bar

13:56 TimMc: &(class identity)

13:56 lazybot: ⇒ clojure.core$identity

13:56 TimMc: &(ancestors (class identity))

13:56 lazybot: ⇒ #{clojure.lang.Fn clojure.lang.IMeta java.lang.Object java.util.Comparator clojure.lang.IFn java.util.concurrent.Callable java.io.Serializable java.lang.Runnable clojure.lang.AFn clojure.lang.IObj clojure.lang.AFunction}

13:57 clj_newb: TimMc: class-per-function as in "each clojure function has a special java class created just for the function" ?

13:58 jodaro: heh

13:59 everytime i think i've seen the radness i end up seeing more

13:59 technomancy: ,(throw (RadnessOverflowException.))

13:59 clojurebot: #<CompilerException java.lang.IllegalArgumentException: Unable to resolve classname: RadnessOverflowException, compiling:(NO_SOURCE_PATH:0)>

14:00 jodaro: like a never ending horizon of radness

14:01 clj_newb: one more dumb question: is there away in clojure, given two types, to ask " can this type be automatically converted ino that type" ?

14:02 ,(type 1) (type 1.0) (type 1919191919191919191919191919191919)

14:02 clojurebot: java.lang.Long

14:02 clj_newb: ,[(type 1) (type 1.0) (type 1919191919191919191919191919191919)]

14:02 clojurebot: [java.lang.Long java.lang.Double clojure.lang.BigInt]

14:02 hiredman: depends what you mean by automatically converted

14:02 clj_newb: I need someting where (f? Long Double) (f? Long BigInt) both return true

14:03 hiredman: ,(doc number?)

14:03 clojurebot: "([x]); Returns true if x is a Number"

14:03 hiredman: ,(number? 1)

14:03 clojurebot: true

14:03 clj_newb: hiredman: I'm buidlging a type system on top of Clojure via macros

14:03 I don't want to do equality checks on types, I want to do checks of the type "A can be cast into B automatically"

14:04 hiredman: clj_newb: hardly seems like a project for a clj_newb

14:04 but have fun

14:04 clj_newb: hiredman: what is the right questions to ask?

14:05 ah; I get it

14:05 I don't care about types

14:05 I care about what protocols it satisfies.

14:06 hiredman: there are a lot of questions, the first is "do I really know enough to implement a type system if I say things like 'automatically converted'"

14:07 clj_newb: hiredman: probably not

14:07 hiredman: so maybe some reading is in order

14:07 clj_newb: hiredman: what should I read?

14:08 hiredman: dunno, something about type systems, or dig into this "automatic" conversion and see what is really going on

14:10 clj_newb: hiredman: seeing how clojure does this "autoamtic" conversion is interesting; how do I grep through the source code for something ill-defined like this?

14:10 furthermore, does this "automatic" conversion ahppen for anything besides numbers?

14:18 TimMc: clj_newb: Correct. Each function, namespace, deftype, defrecord, proxy... each gets a class.

14:20 clj_newb: You're probably building a precondition system, not a type system.

14:22 hiredman: clj_newb: my main point is it's not automatic, what makes you thing it is?

14:22 technomancy: is he talking about boxing?

14:23 hiredman: no

14:23 TimMc: I think he's talking about the reader.

14:24 $findfn Long Double BigInt true

14:24 lazybot: java.lang.RuntimeException: Unable to resolve symbol: BigInt in this context, compiling:(NO_SOURCE_PATH:0)

14:24 TimMc: $findfn Long Double clojure.lang.BigInt true ; there you go!

14:24 lazybot: [clojure.core/not= clojure.core/distinct?]

14:36 cgray: is there a way to close all open files? i made a mistake and i don't want to have to reboot clojure

14:37 amalloy: (System/exit 1), cgray

14:38 (this kills the jvm)

14:49 zilti: Is there a library to make a clojure web app multi-lingual or would I have to implement such a thing myself?

14:51 TimMc: zilti: You mean a localization thingy?

14:51 zilti: TimMc: Yes

14:58 mcrittenden: anyone have any tips for learning to write more idiomatic clojure? I feel like I'm getting a decent grip on the language and I've worked my way through koans and half of 4clojure but e.g. I never would have thought to write flatten like this: https://github.com/clojure/clojure/blob/master/src/clj/clojure/core.clj#L6272

14:58 is it just experience?

15:01 technomancy: mcrittenden: you pretty much have to read lots of code

15:03 cemerick: mcrittenden: I wouldn't beat yourself up after N months with the lanugage because you wouldn't consider an impl that the creator of the language wrote. :-)

15:03 TimMc: Argh, clojure/core.clj links are like rickrolling for me.

15:03 cemerick: TimMc: Noted.

15:03 TimMc: Firefox 3.6 goes gray for like half a minute. >_<

15:04 cemerick: 3.6?

15:04 mcrittenden: cemerick: oh I'm not, I feel like I'm making good progress, I'm just wondering where to go after doing koans and 4clojure to continue to progress as quickly as possible

15:04 cemerick: All enterprisey over there,huh?

15:04 TimMc: cemerick: Linux Mint. :-(

15:04 AeroNotix: is there not a native clojure file IO library?

15:04 cemerick: only about 28 of 'em.

15:04 technomancy: ff is one of two things I still install manually on debian =\

15:04 cemerick: mcrittenden: Build a real app, then build 10 more.

15:04 technomancy: can't handle JITless 3.6

15:05 TimMc: mcrittenden: clojure/core.clj is also full of nasty hacks because of bootstrapping. Watch out in the upper reaches.

15:05 mcrittenden: cemerick: haha, I hear you, I'm just a bit worried that I'll continue building stuff in my non-idiomatic way

15:05 technomancy: mcrittenden: I recommend browsing http://clojuresphere.herokuapp.com

15:05 cemerick: mcrittenden: Man, that's just life. :-)

15:06 mcrittenden: technomancy: nice, bookmarked, thanks!

15:07 TimMc: mcrittenden: flatten isn't totally idiomatic either -- I'd have gone with remove instead of filter and complement.

15:07 zilti: Why remove?

15:07 technomancy: so... what would you do if you had a full graph of all clojure projects anyway?

15:08 zilti: Ah, I was looking for that clojuresphere link yesterday but didn't remember its name, thanks!

15:08 TimMc: Because it's the same thing, only shorter and easier to read.

15:08 cemerick: technomancy: you don't have to be so coy ;-)

15:08 TimMc: technomancy: I don't see lein-jit in there! o\__/o

15:09 technomancy: cemerick: I'm just getting started!

15:09 cemerick: my first idea is a "If I change my API in such and such a way, how will this affect the downstream consumers" plugin.

15:09 but

15:10 certain kinds of refactorings can be automated

15:10 and if they can be automated within your project, why not extend that downstream?

15:10 since you can open pull requests through the github API

15:11 (This is actually what Steve Yegge works on with the internal Google codebase apparently)

15:11 TimMc: Won't somebody please answer this page? http://clojuresphere.herokuapp.com/ring

15:11 It keeps saying "ring/ring, ring/ring"...

15:11 technomancy: BANANAPHONE

15:13 jowag: Hi, I have my state in an atom, and my function which updates this state also performs some side effects (io). How should I correctly update my state? (swap! state update-fn) can call update-fn more than once, so it is not a good solution...

15:13 pandeiro: zilti: you asked about localization... i'm implementing it in the app i'm working on... IMO not something you need a lib for, really. Just store it all in a strings.clj and create a fn that takes a key and a lang (or reads an atom somewhere)

15:13 zilti: pandeiro: Yes, might be the best solution. Thanks!

15:14 TimMc: pandeiro: Isn't it hard to read?

15:15 mcrittenden: how can (fn [x] x) be rewritten in anon func shorthand? #(%) doesn't work. or, in other words, how can I make (partition-by (fn [x] x) x) cleaner (as a solution to http://www.4clojure.com/problem/31)

15:15 TimMc: mcrittenden: identity :-/

15:15 cgray: ,(#(identity %) 1)

15:16 clojurebot: 1

15:16 TimMc: &(identity 1)

15:16 lazybot: ⇒ 1

15:16 mcrittenden: thanks.

15:16 TimMc: jowag: If you use a ref, you can send the io to an agent from inside the dosync.

15:16 zilti: I just offended the whole ##javascript channel :)

15:17 pandeiro: TimMc: isn't what hard to read, the other languages? :)

15:17 zilti: "Even the shittiest scripting languages can do http" Me: "like JS?"

15:19 TimMc: pandeiro: $(:blah-key) instead of "Blah blah blah:

15:21 pandeiro: TimMc: you mean, in the application code? I'm just following the pattern that Android apps use, of separating out strings into a separate file... then when I need a string I do (loc :signup-welcome-message) or something

15:22 jowag: TimMc: thanks, nice solution. So I have to switch to refs...

15:29 TimMc: pandeiro: I don't know, it can get weird when you have marked-up text. I've seen some l10n libs use format strings, actually.

15:37 jowag: TimMc: I'm afraid agents won't work in my case. The new value of state depends on the result of the io operations. So I have to know the result of the io operation before updating the state

15:39 jsabeaudry: What is the threading model for a noir web server? One thread / request?

15:39 (using noir web server loosely here, i mean for the noir web stack)

16:14 Somelauw: ,(String/format "%f" 5.4)

16:14 jcromartie: jsabeaudry: yes,

16:14 clojurebot: #<ClassCastException java.lang.ClassCastException: java.lang.Double cannot be cast to [Ljava.lang.Object;>

16:14 jcromartie: ,(format "%f" 5.4)

16:14 clojurebot: "5.400000"

16:14 jcromartie: String/format expects an array of arguments

16:14 jsabeaudry: jcromartie: Great, thanks!

16:14 adiabatic: in https://gist.github.com/1572428 I'm using (seq coll) as (= () coll), but I get a ClassCastException (can't cast PersistentList$EmptyList to IFn). How should I test for empty listness if destructuring assignment isn't an option?

16:14 jcromartie: objects, to be specific

16:15 ,(source format)

16:15 clojurebot: Source not found

16:15 Somelauw: jcromartie: ah, okay, I though String/format was variable

16:15 jcromartie: oh, well, anyway the source of "format" will shed more light on that

16:15 it is

16:15 but vararg Java methods don't work like that in Clojure

16:15 they require an array

16:16 Somelauw: So clojure doesn't support javastyle varargs

16:16 TimMc: Somelauw: Type... is just Java syntax for Type[]

16:16 jcromartie: here's your exact situation :) http://groups.google.com/group/clojure/browse_thread/thread/64b20acf76e2b687

16:16 Somelauw: Clojure itself does

16:16 kumarshantanu: Hi, is (require '[clojure.string :as str]) supposed to work in ClojureScript REPL in same way it works in the Clojure REPL?

16:16 jcromartie: Somelauw: (defn [arg1 arg2 & arg3-n) …)

16:16 TimMc: Somelauw: When Type... is compiled to bytecode, it looks exactly like Type[].

16:16 jcromartie: er, that was awful sorry

16:17 TimMc: I guess you could argue for Clojure being able to 'apply on those.

16:17 jcromartie: (defn some-vararg-fn [arg1 arg2 & more] … more is a seq here)

16:17 Somelauw: okay, thanks

16:17 jcromartie: TimMc: that would be really nice

16:19 TimMc: although that would definitely require reflection

16:21 TimMc: jcromartie: Not if it were explicit.

16:22 amalloy: adiabatic: you don't want the parens around accumulator

16:22 jcromartie: TimMc: at that point, though, it would be getting as verbose as any other approach

16:22 amalloy: gist is helpfully rendering it in bold-red to tell you you're calling it as a function

16:24 devinus: is there a binding to the last result in a swank REPL?

16:24 jcromartie: TimMc: although it would work if the semantics were a bit different; it would only be sort-of like Clojure's "apply"

16:24 devinus: *1

16:24 devinus: thanks

16:24 adiabatic: amalloy: why do I get that exception being thrown for that screwup, though?

16:24 TimMc: devinus: *1 *2 *3 *e (exception)

16:24 amalloy: &(1)

16:24 lazybot: java.lang.ClassCastException: java.lang.Long cannot be cast to clojure.lang.IFn

16:24 devinus: thanks, i knew the book mentioend it somewhere

16:25 ,*1

16:25 clojurebot: #<Unbound Unbound: #'clojure.core/*1>

16:25 TimMc: jcromartie: I guess it's really like the opposite of apply.

16:25 jcromartie: yeah :)

16:27 adiabatic: amalloy: and thanks a bunch for the help over the past few days

16:28 jcromartie: TimMc: I see a rather messy reflection-laden macro as the only solution… so I'll advocate for into-array

16:28 but it will continue to be a stumbling block for people trying to do interop

16:31 devinus: despite the obvious security implications, is there a way to eval a string in clojure?

16:31 TimMc: devinus: (eval (read-string ...)))

16:32 devinus: and look up *read-eval* refore you use read or read-string

16:32 devinus: TimMc: hrm…okay, another question. is there a way to get the tokens clojure parses from a string?

16:32 jcromartie: devinus: the thing to remember is that Clojure (and other Lisps) evaluate data, not strings of code

16:33 devinus: so reading is the step that turns strings into data

16:33 R E P L

16:33 Read -> Eval -> Print loop

16:33 devinus: jcromartie: right right, i get that

16:33 TimMc: I N C E P T I O N

16:33 jcromartie: ,(read-string "[1 2 {:x 3}]")

16:33 clojurebot: [1 2 {:x 3}]

16:33 devinus: i'm just trying ot get a deeper understanding of how clojure represents reader macros

16:33 TimMc: It's data all the way down.

16:33 clojurebot: Cool story bro.

16:33 devinus: e.g.

16:33 does "; foo" === "(comment " foo") ?

16:34 TimMc: nope

16:34 jcromartie: you can find out :)

16:34 ,(read-string "; comments?")

16:34 clojurebot: #<RuntimeException java.lang.RuntimeException: EOF while reading>

16:34 jcromartie: ,(read-string "[1 2 ; comments?\n 3]")

16:34 clojurebot: [1 2 3]

16:35 devinus: that's odd

16:35 TimMc: &[1 2 (comment 3)]

16:35 lazybot: ⇒ [1 2 nil]

16:35 devinus: hrm

16:36 okay, so definitely not the same

16:36 TimMc: &[1 2 #_ 3 4]

16:36 lazybot: ⇒ [1 2 4]

16:36 devinus: so how is ; a reader macro

16:36 TimMc: devinus: It isn't.

16:36 jcromartie: well the reader isn't that open

16:36 devinus: TimMc: what's happening there?

16:36 jcromartie: as far as users are concerned there aren't reader macros (as that would imply you could define them)

16:37 TimMc: devinus: When the reader sees #_ it ignores the next form.

16:37 devinus: oh interesting

16:37 cgray: jcromartie: there was a package recently announced on the list that implements reader macros

16:37 devinus: wonder whne that's useful

16:37 TimMc: #_ (comment ...) is a safe wa to comment out multiple forms without leaving a nil behind

16:37 It's a good combo.

16:37 devinus: interesting

16:37 jcromartie: cgray: a package or a patch?

16:37 cgray: jcromartie: a package

16:38 devinus: must be why str also ignores nil

16:38 jcromartie: cgray: interesting

16:38 cgray: jcromartie: https://github.com/klutometis/reader-macros/blob/master/src/reader_macros/core.clj

16:38 TimMc: devinus: I doubt that's related.

16:38 devinus: well

16:38 relaated in that

16:38 clojure does a pretty good job of keeping brittle code from being a problem

16:39 although sometimes it's inconsistent it feels like

16:39 TimMc: devinus: ##(let [x 4 (comment y 5)] x) <-- oops

16:39 lazybot: java.lang.IllegalArgumentException: let requires an even number of forms in binding vector

16:39 devinus: offtopic, but what's the diff between clojurebot and lazybot?

16:40 TimMc: Philosophical differences, feature sets, maintainers...

16:40 security :-P

16:40 amalloy: they also have different names

16:40 devinus: haha

16:40 TimMc: trigger characters, thank goodness

16:40 devinus: i wonder what philosophical differences

16:41 jcromartie: when I am in other programming channels, I miss having a bot that evaluates code

16:41 devinus: holy crap github just updated their interface!

16:41 amalloy: again?

16:42 jcromartie: devinus: is it Friday again?

16:42 devinus: :P

16:44 does anybody know what the status is on having a blessed standalone clj ?

16:45 technomancy: thank goodness; the grey backgrounds in the readme were way too low-contrast

16:45 devinus: there's an open ticket for it which seems to have had no progress for many months

16:46 devinus: i'd really like to see somebody address it

16:46 i was able to traverse the waters

16:46 but i see it as a huge barrier to entry for noobs

16:50 technomancy: I'd encourage you to offer to help out, but I'm too cynical to think it would work.

16:51 devinus: technomancy: why don't you think it would work?

16:51 wei_: how do you test whether an item is an atom (in the scheme sense), e.g. not a list-like thing? 

16:51 humblebot: lein repl actually worked pretty well for me, especially with rlwrap installed.

16:52 technomancy: devinus: in my experience, most patches get ignored.

16:52 TimMc: wei_: you can ask if it is a` coll?

16:52 wei_: nevermind, I found it-- coll? is the inverse

16:53 TimMc: thanks, just found that too

16:53 devinus: technomancy: that's…disconcerting

16:54 llasram: technomancy: Then who writes code that actually get included...?

16:54 devinus: technomancy: i was under the impression clojure as a language was more open than e.g. ruby, with it being on github. it seemed to me to be a much more living, moving project

16:55 technomancy: ruby is on github

16:55 devinus: didn't know that

16:55 technomancy: apparently they actually get to use pull requests: https://github.com/ruby/ruby/pulls

16:58 devinus: technomancy: clojure.core doesn't accept pull requests?

16:59 technomancy: maybe someone who's less negative could talk about it.

16:59 zilti: Could it be that swank has a problem with leiningen projects including java sources?

17:00 simonadameit: devinus: according to http://clojure.org/contributing you have to download and sign a copyright asignment agreement and send it via real mail

17:00 devinus: technomancy: -.-

17:00 zilti: simonadameit: That's almost paranoid

17:01 amalloy: i bet seancorfield could find something nice to say about the clojure.core attitude to accepting patches

17:01 technomancy: positive thinking: the great thing about Clojure is how much you can get done without making changes to clojure itself

17:02 llasram: zilti: It's not unheard of. The FSF does basically the same thing. Not doing it means that Linux is stuck with it's modified GPLv2 for all time

17:02 s,it's,its,

17:02 technomancy: llasram: the FSF allows patches under a certain size though.

17:02 plus they're a much more high-profile target for legal action

17:02 llasram: Ah, right, I'd forgotten that

17:03 duck1123: Ok, I can't seem to find this. What do I need to do to get around a function being private? I have a need for ring.middleware.params/parse-params

17:03 devinus: i'm not a lawyer, but wouldn't a CA mean Rich could one day change the license to e.g. GPLv3?

17:03 llasram: And true. But once Clojure is the most popular language in the known galaxy... :-)

17:03 jowag: this reminds me a discussion between technomancy and lau jensen a year ago :)

17:03 technomancy: jowag: that's why I'm trying to keep my mouth shut

17:03 jowag: I know :)

17:04 simonadameit: devinus: yes… although someone else could continue clojure under the existing licence, if that were to happen

17:05 devinus: altleast so I think :)

17:05 devinus: that's probably not a bad idea then

17:05 well, not the GPLv3, but being able to change the license :P

17:06 duck1123: it's already happened in the early days. The CA reduces headache if changing the liscense ever becomes a good idea

17:06 jkkramer: duck1123: (def parse-params #'ring.middleware.params/parse-params) is one way (caveat coder)

17:07 duck1123: jkkramer: could I just reference the var in the fn position? trying

17:07 hiredman: technomancy: the lein script on the 1.x branch seems to be broken, missing an exec, and the exit code is always 1 on non-windows machines

17:07 jkkramer: duck1123: yes, that too

17:07 hiredman: it looks like the exec was removed around 12/30

17:08 technomancy: hiredman: the exec had to be removed to support trampolining around plugins

17:08 hiredman: well, the exit value is incorrect now

17:08 technomancy: yeah, I broke the CI because it depends on getting the uberjar from github downloads, and github only had 1.6.3

17:08 hiredman: it is now the result of the test call at the bottom

17:11 technomancy: ok, lemme fix the exit code

17:11 hiredman: I have a fix

17:11 I could just push it?

17:11 technomancy: sure

17:12 but src/leiningen/test.clj is compiling for you?

17:15 hiredman: technomancy: I believe so, I am using my lein check out for running tests

17:15 pjstadig: contributor agreements are used by corporate controlled open source projects to consolidate copyright and squash competition

17:15 hiredman: technomancy: on 1.x, haven't checked master

17:15 pjstadig: also this http://www.oscon.com/oscon2011/public/schedule/detail/19242

17:17 source for my competition squashing statement: http://www.infoq.com/articles/ieee-controlling-and-steering-open-source-projects

17:18 also there's nothing in the clojure CA that prevents using pull requests, which (IMO) are a far superior way of accepting and reviewing changes

17:20 but i guess technomancy said he wanted someone less negative :)

17:21 replaca: technomancy: slamhound is crashing nastily when I try to run it on autodoc: https://gist.github.com/1572651

17:22 technomancy: this isn't really blocking me though, since I can construct the ns to pprint by hand (or just by cutting and pasting)

17:22 technomancy: replaca: whoa; wild.

17:22 hiredman: or just read the ns form from the beginning of a file?

17:23 technomancy: if it's not blocking can you open an issue?

17:23 cemerick: Is it CA pain Fridays? :-P

17:24 pjstadig: cemerick: hehe

17:24 TimMc: cemerick: I kind of get the CA. The patches instead of pull requests thing.... that boggles me.

17:24 technomancy: cemerick: no, CA complaining is every other Tuesday; Fridays is ignored patches day.

17:24 replaca: technomancy: yup, will do

17:24 cemerick: technomancy: darn, I need to write this down!

17:25 technomancy: http://wondermark.com/506/

17:26 djh__: naive question, when the Raspberry Pi comes out - will it be able to run Clojure/a JavaVM? It's an ARM based mini computer

17:27 devinus: TimMc: i don't even remember submitting patches

17:27 pjstadig: djh__: i suppose one could compile openjdk for it

17:27 devinus: pull requests are just so damn awesome

17:27 technomancy: hiredman: are you looking at the namespaces thing?

17:27 http://travis-ci.org/technomancy/leiningen/builds/487755

17:27 if you're busy I can field it

17:28 hiredman: oh

17:28 djh__: pjstadig - thanks, I suppose it's just going to be a case of "wait and see"

17:28 hiredman: no, I know what I did wrong

17:28 technomancy: hiredman: if you hop in #leiningen you can get yelled at by the travis-ci bot with me =)

17:28 devinus: technomancy: no rush, but when do you think 1.4.0 is gonna make an appearance?

17:28 technomancy: devinus: of swank?

17:29 devinus: technomancy: yeah, sorry of swank

17:29 zilti: Could it be that swank has a problem with leiningen projects including java sources?

17:29 Oh sorry, didn't want to post that line it was my fault about 10 mins ago -.-

17:29 technomancy: devinus: it depends on the underlying cdt library. right now it's using a snapshot and its maintainer has been MIA

17:29 devinus: technomancy: that's unfortunate, that's the reason i wanted to see it

17:30 zilti: MIA?

17:30 technomancy: missing in action

17:30 devinus: the snapshot seems to be working ok

17:31 hiredman: technomancy: "growled at"

17:31 devinus: technomancy: ugh, no offence i'm just trying to cut down on the unstable stuff i use so i think ill wait

17:31 zilti: What I btw really miss is an "Emacs plugin" that makes Emacs understand clojure. :)

17:31 devinus: offense*

17:31 solussd: what is the best way to test if something is a list or vector?

17:31 devinus: offence* lol

17:32 had it right the first time

17:32 solussd: sequential?

17:32 cgray: is there a good tutorial on extend-type et al?

17:32 zilti: solussd: There are functions for exactly that purpose

17:32 TimMc: devinus: Depends which side of the Atlantic you are on.

17:32 solussd: ,(map sequential? [] '() {})

17:32 clojurebot: ()

17:32 TimMc: solussd: list? is slightly brokenish.

17:32 solussd: ,(map sequential? [[] '() {}])

17:32 clojurebot: (true true false)

17:32 TimMc: depending on how you view lists vs. seqs

17:33 zilti: ,(vector? [])

17:33 clojurebot: true

17:33 solussd: looks like I want sequential?

17:33 zilti: ,(vector? '())

17:33 clojurebot: false

17:33 TimMc: solussd: Check out the Venn diagram: http://www.brainonfire.net/files/seqs-and-colls/main.html

17:33 devinus: TimMc: reminds me of when i spelled behavior as behaviour for like 8 months because of erlang

17:33 solussd: thanks!

17:33 technomancy: nothing wrong with behaviour

17:34 zilti: behaviour is british, behavior is american english

17:43 solussd: why doesn't defn work with both a doctoring and pre/post conditions? if I remove the doctoring, the pre/post are respected, otherwise they're ignored.

17:43 hiredman: doctoring?

17:43 solussd: hiredman: osx lion autocorrect

17:43 *docstring

17:43 amalloy: &(doc defn)

17:43 lazybot: ⇒ "Macro ([name doc-string? attr-map? [params*] body] [name doc-string? attr-map? ([params*] body) + attr-map?]); Same as (def name (fn [params* ] exprs*)) or (def name (fn ([params* ] exprs*)+)) with any doc-string or attrs added to the var metadata"

17:44 amalloy: probably you have them in the wrong order

17:44 solussd: am alloy, I'm assuming pre/post go in attr-map- but how can I refer to params before I define them?

17:44 *amalloy^

17:45 TimMc: amalloy: They go before the body, I think.

17:45 ugh

17:45 solussd: ^

17:45 solussd: but that doesn't work, it ignores them if there is a docstring

17:46 using 1.3, fyi

17:46 TimMc: WORKSFORME (defn foo "hi" [] {:post [false]} 6)

17:47 solussd: Put the doc string before the args.

17:51 solussd: ok. that works, but I swear that's what I was doing…. I'll assume the repl is deterministic and chalk it up to an error on my part. ;)

17:52 amalloy: hah. i wish more people would do that

18:22 zilti: Is this my fault or a bug? "Exception in thread \"main\" java.lang.RuntimeException: Unable to resolve symbol: pst-elem-str in this context, compiling:(swank/core.clj:128)"

18:22 technomancy: zilti: you need a newer clj-stacktrace; see the swank readme's "troubleshooting" section

18:23 zilti: Ok

18:24 Do I have to set a certain namespace for the .lein/init.clj?

18:25 technomancy: no, that typically gets loaded in user

18:27 zilti: Is there something else I have to do besides installing the clj-stacktrace and pasting the text into init.clj? It didn't change the stacktrace

18:29 technomancy: could be an older version embedded in some other plugin, maybe lein-difftest

18:30 zilti: My installed plugins are clj-stacktrace-0.2.4.jar lein-clojurescript-1.0.1-SNAPSHOT.jar lein-javac-1.2.1-SNAPSHOT.jar lein-noir-1.2.1.jar lein-ring-0.5.3.jar swank-clojure-1.3.4.jar

18:31 I guess I don't need lein-javac since it's inside lein too

18:39 Hmm it still doesn't work

18:56 adiabatic: ,(keep-indexed println ["spam" "eggs" "sausage" "bacon"])

18:56 clojurebot: (0 spam

18:56 1 eggs

18:56 2 sausage

18:56 3 bacon

18:56 )

19:41 srid: any heroku people here?

19:41 pdk: (* 60 .75)

19:41 ,(* 60 .75)

19:41 clojurebot: #<CompilerException java.lang.RuntimeException: Unable to resolve symbol: .75 in this context, compiling:(NO_SOURCE_PATH:0)>

19:41 pdk: ,(* 60 0.75)

19:41 clojurebot: 45.0

19:47 pdk: ,(* 50 0.75)

19:47 clojurebot: 37.5

20:00 zmaril: Next week I am teaching clojure to a bunch of high school kids. Would it be kosher if I brought them on here and introduced them to #clojure?

20:01 Not like one by one or anything

20:01 etosch: cool!

20:01 zmaril: But point out some of the people here, clojurebot, how to ask good questions, etc. etc.

20:08 wingie: is scala used by more people than clojure?

21:24 clj_newb: is there something like doc, but shows the :pre and :post conditions?

21:24 I'm basically hijacking them for type checking purposes, and I want them to show up when I retrieve the doc of a function

21:29 ,(type (type 2))

21:30 clojurebot: java.lang.Class

21:30 clj_newb: ,(type (type (fn [])))

21:30 clojurebot: java.lang.Class

21:30 clj_newb: is (type (type x)) <--- always Class ?

21:31 AWizzArd: clj_newb: no

21:31 ,(type (type nil))

21:31 clojurebot: nil

21:31 clj_newb: interesting

21:32 ,(type (type ()))

21:32 clojurebot: java.lang.Class

21:32 clj_newb: ,(type (type #false))

21:32 clojurebot: #<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: java.lang.ClassNotFoundException: false>

21:32 clj_newb: ,(type (type f))

21:32 clojurebot: #<CompilerException java.lang.RuntimeException: Unable to resolve symbol: f in this context, compiling:(NO_SOURCE_PATH:0)>

21:32 clj_newb: AWizzArd: this is insightful; thanks for the counter example

21:33 chipdude: ,(#Object)

21:33 clojurebot: #<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: java.lang.ClassNotFoundException: Object>

21:33 chipdude: ,(#java.lang.Object)

21:33 clojurebot: #<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: Unreadable constructor form starting with "#java.lang.Object)">

21:33 chipdude: OK, never mind

21:34 clj_newb: this is somewhat off topic: anyone else feel like a genius whenever they get a defmacro to work?

21:37 AWizzArd: clj_newb: I don’t think that there is a built-in way to show the :pre or :post conditions.

21:37 TimMc: clj_newb: I usually don't feel like a genius when I finally get a piece of code to work; instead, it is more like I finally emerged victorious from a protracted fight.

21:38 amalloy: &#java.lang.Object[] ;; chipdude

21:38 lazybot: ⇒ #<Object java.lang.Object@1e0a141>

21:38 AWizzArd: But one thing you could do is to define a macro over defn which preserves them, and adds them to the doc string.

21:39 Such a macro could even save the full code of your function and make it accessible.

21:39 TimMc: $google serializable-fn

21:39 lazybot: [technomancy/serializable-fn - GitHub] https://github.com/technomancy/serializable-fn

21:39 TimMc: AWizzArd, clj_newb ^

21:40 AWizzArd: I did that around 2005 or 2006 in CL already.

21:42 clj_newb: AWizzArd , TimMc : interesting ;thanks

21:45 as I build macros around defn

21:45 I feeel like I"mrewriting the laws of physics

21:45 TimMc: Hey, wasn't someone looking for a type checking function the other day? 'cast appears to be the thing.

21:46 clj_newb: that was me I believe

21:46 ,(doc cast)

21:46 clojurebot: "([c x]); Throws a ClassCastException if x is not a c, else returns x."

21:47 clj_newb: ,[ (cast (type 1.0) (type 1)), (cast (type 19191919191919) (type 1))]

21:47 clojurebot: #<ClassCastException java.lang.ClassCastException>

21:48 clj_newb: ,[ (cast (type 1.0) 1), (cast (type 19191919191919) 1)]

21:48 clojurebot: #<ClassCastException java.lang.ClassCastException>

21:49 TimMc: That's not a helpful message....

21:50 clj_newb: 1 is a Long, 1.0 is a Double. You can't cast between those, although you can coerce.

21:52 clj_newb: ,(doc coerce)

21:52 clojurebot: Gabh mo leithscéal?

21:55 arohner: has anyone had success with a midje (provided) in another thread?

22:02 alexbaranosky: arohner, what's the question you have?

22:03 I might be able to help you out

22:06 arohner: alexbaranosky: I'm having trouble using a (provided) against code that's called in a future

22:06 alexbaranosky: can I see a gist of your code?

22:07 arohner: i.e. (defn foo [] (future (bar))) (fact "test foo" (foo) => truthy (provided (bar) => truthy)

22:08 yeah, real gist in a second

22:09 alexbaranosky: thanks, it might need some real thought

22:09 Magnars: Given a symbol :myfn - any way of calling myfn as defined in the current namespace?

22:11 arohner: Magnars: yes, look at resolve and apply

22:11 Magnars: arohner: excellent, resolve was what I was looking for - thanks!

22:11 arohner: alexbaranosky: nvm, my simple test case works just fine. I must be doing something wrong in the real version

22:12 alexbaranosky: cool, may I ask what you're working on - is it on GitHub?

22:12 arohner: no, private project

22:12 circleci.com

22:14 oh, my future was racing. after derefing the future so it blocks, I get the expected result.

22:15 alexbaranosky: you're working on circlci.com ? neat

22:15 do you use Midje much? I love to hear feedback both good and bad

22:16 arohner: yeah, I'm one of the founders. I'm pretty happy with midje. The docs quantity and quality are good, but the organization is a little hard to navigate

22:17 and sometimes invalid syntax in the macros sometimes causes infinite hangs

22:18 alexbaranosky: arohner, I've donea LOT of work this past month working on validations of the macros, so you get error messages instead of mysterious failures

22:18 which macros in particular? I did a lot on the background and against-background

22:18 arohner: great! I'm still on 1.2.0

22:18 that's probably it. I'll investigate more the next time it happens

22:20 alexbaranosky: I've also a dded doc strings to the public api macros, so at least siome level of documentation should be available at your fingertips

22:21 arohner, I also definately consider issues people care about more when choosing what to work on, so feel free to make yourself heard if you care about some feature

22:22 arohner: alexbaranosky: great, thanks

22:25 TimMc: [Announcer voice] "Alex Baranosky: His future... was racing"

22:29 alexbaranosky: what's up TimMc

22:29 TimMc: goin' to bed

22:29 I used up all my Clojure yesterday.

22:30 alexbaranosky: it happens :\

22:30 TimMc: alexbaranosky: lein-jit, check it out. It beat me up, but I finally won.

22:31 alexbaranosky: will do!

22:33 jodaro: ran out of parentheses?

22:44 clj_newb: here's some more: (repeat "()")

22:55 amalloy: &(nth (iterate list ()) 50) ;; gotta keep em balanced

22:55 lazybot: ⇒ ((((((((((((((((((((((((((((((((((((((((((((((((((()))))))))))))))))))))))))))))))))))))))))))))))))))

23:05 pdk: ,(* 40 0.75)

23:05 clojurebot: 30.0

23:06 jodaro: i get mine imported from antarctica

23:06 made by penguins

23:09 * technomancy insists on only shade-grown, fair-trade parens

23:10 pdk: ,(* 56.78 0.75)

23:10 clojurebot: 42.585

23:12 pdk: ,(+ 35.81 7.50)

23:12 clojurebot: 43.31

23:13 jodaro: pdk: calculating tips for the week?

23:18 humblebot: ,(+ 1 2 3 4 5 6 7 8 9 10)

23:18 clojurebot: 55

23:18 humblebot: Cool!

23:18 Yes, I'm new.

23:21 TimMc: technomancy: Those are bananas, you've been coding with bananas.

23:23 technomancy: as long as paredit supports 'em; I don't want to know

Logging service provided by n01se.net