#clojure log - Aug 05 2012

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

0:00 m0smith: oh I have another question about futures: Does using futures keep the program from terminiating, even if they are all completed?

0:02 that is, are they deamon threads?

0:02 (##source future)

0:02 ##(source future)

0:02 lazybot: java.lang.RuntimeException: Unable to resolve symbol: source in this context

0:08 m0smith: I found it: http://dev.clojure.org/jira/browse/CLJ-124

0:09 ok thanks all. night

0:11 mdeboard: adios

0:16 Are there ever actually any locals in the clojure stacktrace?

0:23 xeqi: mdeboard: http://dev.clojure.org/jira/browse/CLJ-860 looks like a way was added to keep them

0:24 mdeboard: oic

0:26 wonder if that's slated for 1.5 or what

0:27 xeqi: says resolved in 1.4

0:27 though I've never tried it

0:28 muhoo: slingshot can show locals in stacktrace, IIRC

0:30 mdeboard: xeqi: Not clear here on how I would even try it.

0:47 amalloy: you can't put locals into a stacktrace. disabling locals clearing just gets the compiler to stop nulling out values you're done with so that a debugger, or slingshot, can collect that information for you and put it somewhere useful

2:06 polypus: ~ping

2:06 clojurebot: PONG!

2:06 polypus: carmine Q here: http://gist.github.com/3262135

3:07 mdeboard: Why is this macro giving me a stack overflow error? https://gist.github.com/d85b7f024a431cbbdfdf

3:08 These two macroexpand exprs yield practically the same code https://gist.github.com/5263a49008da659e89ec

3:12 amalloy: mdeboard: proxy-foo unconditionally contains itself in its expansion

3:12 so that nested macro needs to be expanded, ad infinitum

3:12 mdeboard: amalloy: I've been reading through your blog post on this exact topic btw (macros in macros)

3:12 amalloy: hah

3:13 but you're not doing any nested macros

3:13 mdeboard: well, I guess I mean that like proxy-foo contains proxy, which is a macro itself

3:13 amalloy: yeah, but pretty much every macro contains another in that way, because let is a macro too

3:13 mdeboard: oh, right. defn, etc.

3:14 So, in this case with proxy-foo invoking itself, how do I go about expressing this?

3:14 amalloy: you probably need your macro to generate a function, and have withMeta call that function

3:14 mdeboard: I see

3:15 amalloy: btw, no need for this klass tomfoolery. clojure doesn't have java keywords

3:15 mdeboard: alternatively, is there a way to define a variable like `(def r PdfReader)' then `(proxy [r clojure.lang.IObj] ...)'

3:15 amalloy: no

3:15 mdeboard: ok

3:17 amalloy: anyway, there are some other problems in your macro (what leaps out at me is multiple-evaluation of moremeta, and evaluating moremeta and infile at macro-expansion time), which will all be pretty much fixed if you just have your macro emit a function, rather than trying to expand into a call to proxy

3:18 mdeboard: multiple eval of moremeta? Where?

3:19 amalloy: then it's more like (defmacro proxy-foo [class] `(fn make# [infile# & [moremeta#]] (proxy [~class IObj] [infile#] (withMeta [newmeta#] (make# infile# newmeta#)))))

3:20 mdeboard: the fundamental thing I'm trying to do here is not have a function to proxy every class I need so I can use the iobj interface, and that requires at least being able to use one function to proxy any class

3:20 I see

3:20 amalloy: mdeboard: you let metadata be (merge blah meta), but you do that at compile-time. then every time someone calls (meta the-proxy), you expand metadata. so if moremeta is {:x (println "1")}, that println happens every time someone calls meta

3:24 polypus: anybody here used carmine (redis) atomically helper? wondering if i'm doing it right or if i've found a bug? http://gist.github.com/3262135

3:25 mdeboard: amalloy: Waddya know, that works.

3:25 amalloy: You're ok

3:26 this is like 3 levels of abstraction higher than I'm prepared to go at 3am

3:27 amalloy: So basically in this gist, the macro below is generating the (proxy-meta) function above right?

3:27 amalloy: errr, what gist, what's above/below?

3:28 mdeboard: https://gist.github.com/4010a4cc804b8e83bb56

3:29 amalloy: yes

3:29 so the latter macro lets you write the former function with (def proxy-meta (proxy-foo PdfReader))

3:30 (though seriously, when it's no longer 3am, work on those names)

3:30 mdeboard: lol, you got it

3:54 muhoo: heh, a promise is an empty future.

3:57 foodoo: muhoo: sounds poetic, but sad

3:57 michaelr`: good morning

3:59 mdeboard: god I hate itext so much

4:00 Impossible to extend it via proxy

4:00 er, it being a particular class in the lib

5:16 alexyakushev: Is it possible to attach pre- and postconditions to a lambda?

5:57 TEttinger2: (https://github.com/technomancy/leiningen/blob/master/bin/lein.bat)

5:57 ,(https://github.com/technomancy/leiningen/blob/master/bin/lein.bat)

5:57 clojurebot: #<CompilerException java.lang.RuntimeException: No such namespace: https://github.com/technomancy/leiningen/blob/master/bin, compiling:(NO_SOURCE_PATH:0)>

5:57 TEttinger2: anyone know how to get lazybot to announce the titles of linked webpages?

5:58 &(https://github.com/technomancy/leiningen/blob/master/bin/lein.bat)

5:58 lazybot: java.lang.RuntimeException: No such namespace: https://github.com/technomancy/leiningen/blob/master/bin

5:58 TEttinger2: &https://github.com/technomancy/leiningen/blob/master/bin/lein.bat

5:58 lazybot: java.lang.RuntimeException: No such namespace: https://github.com/technomancy/leiningen/blob/master/bin

5:58 TEttinger2: hey edlothiol

5:58 just so you know, I am debugging my own lazybot over in #rgrd

5:59 https://github.com/technomancy/leiningen/blob/master/bin/lein.bat

6:00 and I imagine that a bot meant to do this passively wouldn't respond to a PM with a link

6:02 -> https://github.com/technomancy/leiningen/blob/master/bin/lein.bat

6:19 raek: TEttinger2: note that bots can usually only print titles for HTML documents

6:20 TEttinger2: raek, that is a github webpage right?

6:20 raek: oh, I see. the link _is_ to an html file...

6:28 Raynes: TEttinger2: Did you get it working?

6:28 $title https://github.com

6:28 lazybot: "GitHub · Social Coding"

6:28 Raynes: To make it do it automatically, you have to add some configuration.

6:28 Let me look it up.

6:28 TEttinger2: Raynes, I did not. however, I know about the config

6:29 I am trying to get the thing to actually run a selenium test

6:29 Raynes: I don't follow.

6:29 TEttinger2: yes

6:29 so if you paste in www.github.com

6:30 then my version of lazybot is supposed to load the page in firefox (complete security hole, I may get around to stopping that up), check a specific plugin's added span area, and report it back

6:30 what it is supposed to do is check for trackers on linked sites

6:31 google analytics, advertising networks, the whole deal

6:39 Raynes, as far as I can tell, there isn't an error with LazyBot. I just don't know how to work semperos's lib

6:39 https://github.com/semperos/clj-webdriver

6:40 I keep getting an unable to resolve symbol error on what should be a valid symbol

6:42 I am trying to edit the title plugin, if that matters

7:28 guh

7:29 java.lang.IllegalArgumentException: No value supplied for key: true

7:29 anyone know what that could be?

7:30 is that like a mismatched hash literal?

8:12 DaoWen: I've been looking through the docs, and I'm confused by these three macros: binding, with-bindings, with-redefs

8:12 when would I want to use one as opposed to the others?

8:15 clj_newb: Hi all, how do I merge sets?

8:16 DaoWen: clj_newb: (into set-a set-b)

8:16 clj_newb: thanks!

8:17 DaoWen: clj_newb: you can also use the functions in clojure.set: http://clojuredocs.org/clojure_core/clojure.set

8:18 clojure.set/union and clojure.set/intersection specifically, depending on what you mean by "merge"

8:45 clj_newb: Thanks very much DaoWen

8:46 DaoWen: you're welcome!

8:50 gfredericks: DaoWen: with-redefs is just for testing

8:50 and/or development

8:50 binding is standard and used whenever you want to give a var a thread-local value

8:51 with-bindings I wasn't even aware of, but it looks like it's useful for when you have runtime references to var objects rather than the normal compile-time case

8:52 which is probably rare

8:52 DaoWen: ok

8:52 so it sounds like with-redefs would never come up in production code

8:52 gfredericks: if it does you're abusing it

8:53 it changes globals across all threads

8:53 so it's not really safe

8:54 DaoWen: that makes sense

8:56 I'm still confused about with-bindings, but since you didn't even know it existed it sounds like it would only come up in really weird situations since it works with runtime references to vars

8:56 thanks

9:30 benedikt: Using emacs and lein, if I create a new project with lein, open it with emacs and start "clojure-jack-in" and try to evalute something with C-M-x, i get the error "No buffer named *nrepl-connection*". Google is of no help.

9:48 somewhat confused with including libraries with use. Is the convention to use (user 'library) or somehow use it in the ns macro (if so, what's the syntax?)

9:50 michaelr`: benedikt: both

9:50 you can do (use 'clojure.string)

9:51 or in (ns (:use [clojure.string))

9:51 benedikt: what is the syntax like if i embed it in the ns namespace?

9:51 ah

9:51 thanks

9:51 michaelr`: 'welcome

10:13 gfredericks: ,(rationalize (rand))

10:13 clojurebot: 6761690242602623/10000000000000000

10:13 gfredericks: :)

10:15 michaelr`: ,(rationalize (rand))

10:15 clojurebot: 4030645784873027/125000000000000000

10:16 TEttinger2: Raynes, do you remember what customizations lazybot requires to have auto-announce of the titles of links?

10:17 I have :title {:automatic? true} in my config.clj

10:27 ,(apply str (shuffle ["▼" "◀" "▶" "◢" "◣" "◥" "◤" "△" "▽" "◿" "◺" "◹" "◸" "▴" "▾" "◂" "▸" "▵" "▿" "◃" "▹" "◁" "▷" "◅" "▻" "◬" "⟁" "⧋" "⧊" "⊿" "∆" "∇" "◭" "◮" "⧩" "⧨" "⌔" "⟐" "◇" "◆" "◈" "⬖" "⬗" "⬘" "⬙" "⬠" "⬡" "⎔" "⋄" "◊" "⧫" "⬢" "⬣" "▰" "▪" "◼" "▮" "◾" "▗" "▖" "■" "∎" "▃" "▄" "▅" "▆" "▇"

10:27 clojurebot: #<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: EOF while reading>

10:28 TEttinger2: ,(apply str (shuffle ["▼" "◀" "▶" "◢" "◣" "◥" "◤" "△" "▽" "◿" "◺" "◹" "◸" "▴" "▾" "◂" "▸" "▵" "▿" "◃" "▹" "◁" "▷" "◅" "▻" "◬" "⟁" "⧋" "⧊" "⊿" "∆" "∇" "◭" "◮" "⧩" "⧨" "⌔" "⟐" "◇" "◆" "◈" "⬖" "⬗" "⬘" "⬙" "⬠" "⬡" "⎔" "⋄" "◊" "⧫" "⬢" "⬣" "▰" "▪" "◼" "▮" "◾" "▗" "▖" "■" "∎" "▃" "▄" "▅" "▆" "▇"

10:28 clojurebot: #<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: EOF while reading>

10:29 TEttinger2: ,(apply str (shuffle ["▼" "◀" "▶" "◢" "◣" "◥" "◤" "△" "▽" "◿" "▂" "▁" "▬" "▔" "▫" "▯" "▭" "▱" "◽" "□" "◻" "▢" ]))

10:29 clojurebot: "??????????????????????????????????????????????????????????????????"

10:30 TEttinger2: (take 10 (cycle (range 0 3)))

10:31 clojure-newcomer: hi guys, I've got two sets, which appear to be equal, I'm getting a different value for a contains? on them both… which got ya is catching me out ?

10:31 TEttinger2: ,(apply str (shuffle (take 100 (cycle (range 0 10)))))

10:31 clojurebot: "2305923435368826349813166532448267361445182423703409270509171857772184654509951599678082198760701906"

10:49 raek: clojure-newcomer: are the keys integers?

10:50 clojure-newcomer: raek: I'm trying the following : (contains? st2 {:db/id 17592186045427})

10:50 raek: where st2 is : #{{:db/id 17592186045427}}

10:50 this one works

10:51 but not st which to all intents and purposes is the same

10:51 one is a result from diatomic, the other is an explicit set definition

10:52 raek: what does (class? ...) return for the number values?

10:53 clojure-newcomer: raek: java.lang.Long

10:54 raek: for both ones?

10:55 clojure-newcomer: raek: it seems to

10:55 raek: (= st st2) yields false

10:55 raek: does these two maps yield true when passed to =?

10:56 what is st?

10:56 clojure-newcomer: both are the sets I mentioned earlier

10:56 antares_: clojure-newcomer: can you post a REPL session example?

10:56 raek: clojure-newcomer: I only see st2

10:56 clojure-newcomer: raek: st is the set returned from querying a diatomic result

10:57 raek: sorry, I was not clear on that

10:57 raek: what does it look like if you print it?

10:57 antares_: it is generally a good idea to use clojure.set functions instead of contains? because you won't have to worry about ref equality vs value equality

10:57 clojure-newcomer: raek: st and sty look identical when println'd

10:57 raek: *st2

10:58 raek: antares_: what? I think they use the same model of equality.

11:00 clojure-newcomer: so if you run (do (prn st) (prn st2) (prn (= st st2))) you get two identical lines, followed by false?

11:01 clojure-newcomer: here is a REPL session demonstrating my fail: http://pastebin.com/MKWeZ0p4

11:02 raek: yes I do

11:04 raek: clojure-newcomer: the first line has a different number that the rest

11:05 17592186045425 instead of 17592186045427

11:05 (last digit)

11:05 antares_: it is hard to tell what's inside entity but those timestamp-looking ids are not the same

11:06 ,(= "17592186045425" "17592186045427")

11:06 clojurebot: false

11:06 clojure-newcomer: raek: yeah, thats ok thats the paren entity, I am pulling a property off that

11:06 raek: that string comparison might yield true in PHP though... :-)

11:07 hrm. strange.

11:09 clojure-newcomer: if its any help, clojure: 1.4.0, datomic: 0.8.3343

11:09 raek: clojure-newcomer: what does (for [x [st st2]] [(class x) (class (first x)) (class (first (keys (first x)))) (class (first (vals (first x))))]) return?

11:09 clojure-newcomer: raek: it returns : ([clojure.lang.PersistentHashSet datomic.query.EntityMap clojure.lang.Keyword java.lang.Long] [clojure.lang.PersistentHashSet clojure.lang.PersistentHashMap clojure.lang.Keyword java.lang.Long])

11:10 michaelr`: antares_: Isn't (String. payload) in msg-handler in subscribe supposed to convert the byte array to string?

11:11 clojure-newcomer: raek: so the first has EntityMap, the second has PersistentHashSet

11:11 raek: to me it looks like equality is broken for the datomic.query.EntityMap type

11:11 or maybe this is by design and you should do things differently

11:11 I recommend asking the datomic folks about this

11:12 michaelr`: antares_: nm, it does :)

11:12 clojure-newcomer: raek: thanks for the help, I'll go try to get help

11:12 from the diatomic guys

11:13 *datomic… stupid auto correct

11:14 raek: I would advice against using the (String. byte-array) constructor

11:14 use (String. byte-array encoding) instead

11:14 michaelr`: oh

11:15 raek: otherwise it tries to guess the encoding?

11:15 raek: michaelr`: no, otherwise it will use the "default one"

11:16 which is not a very good idea

11:16 michaelr`: oh

11:16 raek: this encoding depends on the locale of the user

11:16 and is wrong on OS X computers

11:16 (java is told it is Mac Roman, but OS X uses UTF-8)

11:17 use "UTF-8" unless you have a good reason

11:17 michaelr`: ok

11:17 thanks

11:17 antares_: leiningen 2 will make JVM use UTF-8 starting with preview5 or so

11:18 via a -Dproperty

11:18 raek: that's good to hear

11:43 michaelr`: why pprint doesn't show anything when run inside a `lein run`?

11:44 i know it'

11:44 s

11:44 something with the stdout redirection

11:44 better question is how do i fix it? :)

11:47 ,(doc pprint)

11:47 clojurebot: I don't understand.

12:23 duck1123: michaelr`: are you able to see normal println messages

12:59 davidd___: I wrote a function that looks really ugly. can someone tell me how I can refactor it: https://www.refheap.com/paste/4055

13:06 tmciver: davidd___: It looks good but I think that you can replace the the if-let and ifs with when-let and when since they will return nil if their test expressions are falsey. That way you don't have to return false explicitly.

13:08 gfredericks: davidd___: s/save-user/save-user?/

13:21 davidd___: tmciver, gfredericks: thanks!

13:49 muhoo: hmm, cljsbuild seems to compile everything into .lein-cljsbuild-compiler-0, but then if you run the cljs repl, it re-compiles everything all over again, and puts it in .lein-cljsbuild-repl .

13:52 so there are two compilers going in parallel on what may be the same code?

14:01 also, one more cljs question since nobody's listening: how do i force cljs repl to give me a cljs-usable object, or a string, from stuff like "[object HTMLParagraphElement]"

14:02 i.e. this is not the expected behavior: https://www.refheap.com/paste/4057

14:03 oh, duh, it's javascript, i can use js interop. nm.

14:11 pandeiro: muhoo: what would you like to do?

14:13 are you trying to get (.-innerHTML el) or (.-innerText el) by chance?

14:13 or are you wanting to do the reverse of what crate does? ie HTMLDivElement -> [:div "something"] ?

14:13 muhoo: pandeiro: that's the one :-)

14:14 pandeiro: the reverse?

14:14 muhoo: well, the crate docs say this: https://github.com/ibdknox/crate

14:14 pandeiro: man that i have no idea but when you figure it out, please pass me that code :)

14:14 muhoo: but what i was getting in the repl was this: https://www.refheap.com/paste/4057

14:14 pandeiro: but is it necessarily coming from the HTML object?

14:14 muhoo: so, nrepl was, for me, returning the object, but in the docs, it looks like it was returning .-innerHTML on it

14:15 pandeiro: DOM object i mean?

14:15 muhoo: i'm not sure. this is cljs :-)

14:16 meenal: Hi, is there a StringBuffer like datastructure in clojure?

14:16 pandeiro: I don't think you can go from a DOM Object (ie <[object HTMLParagraphElement]>) to a crate/hiccup vector very easily

14:16 i think you'd need to serialize it to HTML string (.-innerHTML) and then parse

14:16 basically a new lib for that

14:16 muhoo: pandeiro: no i'm just looking at having readable in the repl

14:17 #<[object HTMLParagraphElement]> is not very informative

14:17 but, i guess the repl used to print out the HTML of the dom element. but now it doesn't, or maybe i'm doing something wrong

14:17 pandeiro: maybe somehow crate could save a string representation of the vector data structure when it makes the DOMElement

14:18 no it never printed out the entire html string, not that i remember

14:18 hiccup would do that

14:18 muhoo: but, that's what i'm seeing in that Readme.md

14:18 pandeiro: not crate, which creates DOMObjects

14:18 ah, different fn?

14:18 muhoo: https://github.com/ibdknox/crate/blob/master/README.md

14:19 not sure. again, i'm just feeling my way around, and when i see output that doesn't match the readme, i wonder what i might be doing wrong

14:20 this istuff is very cool though.

14:20 heh, not istuff, i don't like iStuff

14:21 pandeiro: nah if you look at crate.core, html is creating actual DOM objects

14:21 i think the README is wrong

14:21 maybe ibdknox will chime in

14:21 muhoo: right, as i'd expect. ok maybe that's it then.

14:22 pandeiro: wonder if you could somehow save the (pr-str) representation of the element before creating it, and keep that around (where? dunno) for inspection purposes

14:22 muhoo: metadata?

14:22 clojurebot: ? "Returns the metadata of obj, returns nil if there is no metadata."

14:22 pandeiro: right, in HTML?

14:22 muhoo: if it's an object, hmm.. how does cljs store metadata, i wonder

14:22 pandeiro: like, attached the DOM object? don't think it's possible?

14:23 i think that's all in clojure's data structure code, not applicable to a HTMLParagraphElement etc

14:23 jQuery can attach data to DOM nodes

14:24 i've never looked at how that's implemented though

14:24 i guess there's no reason it couldn't just be added to the DOM object itself

14:25 muhoo: No protocol method IWithMeta.-with-meta defined for type object: [object HTMLParagraphElement]>

14:25 pandeiro: maybe you could extend it

14:26 muhoo: looks that way

14:26 pandeiro: or just use JS interop a la (set! (.-crate el) crate-string)

14:27 but yeah i don't think that's a robust solution.. you'd need to change the toString() method for the REPL to output that crate key's values

14:27 muhoo: i'd rather it be clojureish, like with-meta. i'll put that on the loist of things to play with.

14:28 pandeiro: yeah should be possible to extend those DOM objects to IWithMeta right?

14:28 muhoo: just add the protocol i think.

14:28 i saw a talk by lynagh or chouser or someone on how to extend js with protocols

14:28 pandeiro: muhoo: jayq.core has a good examples, he made the jQuery object seqable

14:29 muhoo: cool, thanks!

14:29 antares_: Spyglass 1.0 is out: http://clojurememcached.info

14:29 muhoo: i came into cljs wanting to do productive stuff, but i got seduced by the dark side and now i just want to play around with how it works

14:31 meenal: fbuffe

14:31 piranha_: ehm... I'm trying to get defprotocol/defrecord working on ClojureScript and failing at this point. If I define them in different modules, cjsc generates different method names and `satisfies?` says `false`. If I define them (both protocol and record) in a single module, I can use protocol methods on record, because names are the same, but then `satisfies?` is still saying false. Any comments?

14:33 antares_: piranha_: start with showing your code

14:33 piranha_: antares_: https://github.com/piranha/cj-locations/blob/master/src/google.cljs#L9

14:33 here are they in different modules, protocol is obviously in map.cljs

14:35 muhoo: piranha: google maps wrapper in cljs? very cool!

14:35 antares_: I am not familiar with CLJS, in JVM Clojure protocols definitely can be defined in a different namespace from the type you extend them to

14:36 piranha: muhoo: well, this is only for a limited thing, I'm just starting with clojure/cljs

14:36 antares_: however, I typically use defprotocol + extend-protocol, not the extended defrecord form

14:36 piranha: well, I tried to go this way

14:36 muhoo: piranha: maybe it's the order in which they are included?

14:36 piranha: it generates code in same way...

14:36 muhoo: hmhm, well, the thing is that it looks like there is some support code which wasn't generated or something like that

14:36 muhoo: might want to look at the generated js and see the order of the require() calls

14:37 antares_: piranha: have you tried separating defrecord and protocol extension code (with extend-protocol, for example)?

14:37 piranha: well, satisfies? checks that Map knows about this Google it seems

14:37 antares_: yes, I did - result was the same

14:38 well, not exactly, since I couldn't use just names of properties (gmap/coder/info), but in other way same effect

14:39 eh... probably it would be better to write to a clojure mailing list, since not everybody is online on irc all the time :)

14:39 antares_: you :use [locations.map :only [Map init locate set-city]]. That's a little strange.

14:39 :only should be listing functions, not protocol names

14:39 although it should not matter, the point is to get the namespace loaded & compiled

14:39 and :use accomplishes that

14:40 piranha: hmmm

14:40 well, I wanted to import Map to check with satisfies?

14:40 I could check it with require as well, give me a second

14:48 antares_: it's the same...

14:49 antares_: ok, probably CLJS protocols have something to them that I do not know

14:49 piranha: seems like a bug to me

14:49 I'll write to mailing list then

14:54 clj_newb: Hi, in this persistent array map {false [1 2], true [4]} how do I get the keys?

14:54 gfredericks: ,(keys {false [1 2], true [4]}

14:54 clojurebot: #<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: EOF while reading>

14:54 gfredericks: ,(keys {false [1 2], true [4]})

14:54 clojurebot: (true false)

14:54 antares_: ,(keys {:a 1 :b 2})

14:54 clojurebot: (:a :b)

14:55 antares_: ,(values {:a 1 :b 2})

14:55 clojurebot: #<CompilerException java.lang.RuntimeException: Unable to resolve symbol: values in this context, compiling:(NO_SOURCE_PATH:0)>

14:55 clj_newb: (keys {false [1 2], true [4]}) yields (false true), but (:false {false [1 2], true [4]}) yields nil

14:55 antares_: ,(vals {:a 1 :b 2})

14:55 clojurebot: (1 2)

14:55 gfredericks: clj_newb: false and :false are different

14:55 ,(get {false [1 2], true [4]} false)

14:55 clojurebot: [1 2]

14:56 gfredericks: ,(map type [false :false])

14:56 clojurebot: (java.lang.Boolean clojure.lang.Keyword)

14:56 clj_newb: aah, I got confuse when (false {false [1 2], true [4]}) failed

14:56 thanks both

14:57 gfredericks: clj_newb: keywords (like :false) are special in that you can use them as functions for accessing maps

14:57 for any other type of key you can use get or you can use the map itself as a function

14:57 antares_: they are not "special", they just can be used as functions on maps

14:57 to look themselves up in the argument

14:57 gfredericks: antares_: what's not special about that?

14:57 antares_: (map :age [{:name "Joe" :age 30} {:name "Marie" :age 29}])

14:57 ,(map :age [{:name "Joe" :age 30} {:name "Marie" :age 29}])

14:57 clojurebot: (30 29)

14:58 antares_: gfredericks: several other things in Clojure can be used as functions, e.g. sets

14:59 gfredericks: antares_: I don't mean they're special because they're functions, I mean (like I said) that they're special because they're functions for accessing themselves in maps

14:59 * gfredericks decides not to care so much

15:10 muhoo: piranha__: i don't fully trust the compiler yet, so i'd recommend checking the generated .js and make sure it has your changes

15:11 piranha__: muhoo: yes, it does

15:11 jrheard: hi folks - i'm having a little trouble figuring something out: i have a simple validation function and want to let you pass in a particular kind of exception to throw if validation fails

15:12 the compiler doesn't like it when i have a bare Exception. sitting around though

15:12 piranha: muhoo: it's just they don't make much sense to me... i.e. satisifies? checks for something what is never done (Google record assigned as a property to Map protocol, for example)

15:12 jrheard: e.g. this fails:

15:12 (ensure-in x resource-types 'ClassCastException.)

15:12 with and without the quote

15:12 am i doing something really stupid :/

15:14 are java constructors not first-class functions or is my mental map of the situation messed up?

15:17 antares_: jrheard: they are not functions

15:17 jrheard: huh!

15:17 antares_: they are invoked via special forms

15:17 jrheard: aha

15:17 i'm working through "clojure programming" but haven't gotten to the interop chapter yet :/

15:17 do i have any options here? i kind of like this idea, it seems nice

15:18 antares_: jrheard: it's not hard: (ClassName. arg1 arg2 arg3)

15:18 jrheard: so i should pass along an instance instead of just what i was hoping was a reference to a function?

15:18 e.g. (ensure-in x resource-types (ClassCastException.))

15:19 antares_: jrheard: you want to instantiate an object from a class you get in a local?

15:19 jrheard: yes indeed

15:19 specifically that

15:20 muhoo: jrheard: the interop chapter in clojure programming is very good. once you get there, it will all make sense :-)

15:21 jrheard: good to hear :) this book has been really fantastic so far

15:21 muhoo: i'd recommend skipping ahead and skimming if this is in your critical path now.

15:21 jrheard: i'll do that, it is

15:21 muhoo: it's a great book for skipping around if you need to

15:23 antares_: jrheard: that's possible but a bit tricky. I will try to put together a reasonable example (out of curiosity)

15:23 jrheard: basically, Java classes have the newInstance method you can use but there are nuances

15:24 you can also do things like ,(Class/forName "java.lang.Long")

15:24 jrheard: whooa

15:24 antares_: this is how all those ORMs and stuff dynamically load DB drivers, for example

15:24 w/o depending on them all at once

15:25 jrheard: i would love to see that example whenever you get around to it :) i'm @jrheard in case i'm not online

15:25 thanks for the guidance antares_ muhoo!

15:29 gfredericks: jrheard: also you could instead of passing in classes pass in functions that will return an exception

15:29 so #(ClassCastException.)

15:29 jrheard: yeah, that makes sense

15:29 i'll just do that!

15:29 thanks :)

15:47 babysnakes: hi, I'm new here :) I want to create a website in hebrew (RTL) and I thought of ways to extract all the strings into other files. In my very limited java knowledge I know there is a good support for i18n, but I don't really need to support multi language. I just remember that they have nice resource handling. Is there a better way to do it (something like rails yaml files)?

15:47 muhoo: that's kinda funny. ($ :a) returns an object, but there are no links. and (count ($ :a)) is 0

15:48 leaky abstraction, perhance

15:50 babysnakes: one of the guys here (borkdude)? wrote a hebrew tutorial app in clojure

15:51 babysnakes: http://borkent-app2.appspot.com, i think

15:52 babysnakes: muhoo: thx, what I'm looking for have nothing to do directly with hebrew :) I just want to extract the RTL text from the regular clojure files and I was wondering what's a good way to do it.

16:10 TimMc: Presumably babysnakes wants to extract it because it screws up a normal text editor.

16:23 otfrom: evening all

16:38 weavejester: otfrom: Evenin'

17:01 Frozenlock: It's possible to sort in congomongo using :sort, but how does one can change the sort order?

17:02 cshell: does mongo have a sort construct?

17:02 Frozenlock: yes http://www.mongodb.org/display/DOCS/Sorting+and+Natural+Order

17:04 cshell: looks like you can

17:05 :sort {:key 1}

17:05 or -1

17:05 Frozenlock: Yeah, I've tried that :P

17:07 I'm unable to change the sort order... I've tried :sort {:key -1}, [:key -1], [[:key -1]]...

17:09 cshell: you're not using fetch-one are you

17:09 ?

17:10 Frozenlock: I am, why?

17:11 cshell: www.litbuildingdb.com/search

17:11 oops

17:11 not that

17:11 http://bit.ly/PxGe60

17:11 that

17:13 looks like you have to use fetch to do it

17:13 Frozenlock: eh what... If I understand correctly, if I want to sort, I will get all the results?

17:13 cshell: no, you'll get a cursor

17:14 Frozenlock: So if I do (first (fetch...)) I will NOT download the entire collection?

17:15 cshell: not sure, you'll have to experiment, but there's also :limit

17:15 Frozenlock: Oh right!

17:15 Forgot about this one.

17:16 Thanks for the answer btw, I would never have guessed that fetch-one was different from fetch!

17:16 cshell: haha either would I

18:12 akhudek: working with google's closure is often very frustrating. :-(

18:26 pandeiro: akhudek: tell me about it

18:27 and the partially unfinished abstractions in clojure.browser.* don't make it much easier

18:28 i am trying to make the browser repl work from a content script in a chrome extension and debugging goog.net.xpc.CrossPageChannel etc, no fun

18:29 all the protocols in clojure.browser.net are just an additional layer of indirection that makes it harder to understand what's happening

18:51 akhudek: pandeiro: I briefly looked at the clojure.browser.net namespace and ended up using jayq to do RPC instead

18:51 my latest frustrations have been with the closure editor

18:52 some of the editor plugins just don't do what they advertise and there is no information or complaints on their discussion group

18:52 It's quite confusing. I can only imagine that not many people use it outside of google.

18:57 pandeiro: akhudek: yes and i get the impression even much of google has moved on?

18:57 definitely google docs isn't using it

18:57 then there's angular js or whatever it is

18:57 i was also disappointed with the editor for lack of documentation

18:58 not lack of docs per se, but the lack of an open source ecosystem, blog posts, irc, that sort of thing

19:16 cshell: has anyone had any problems with ring's sessions not being shared across heroku dynos?

19:16 weavejester: cshell: What session store are you using?

19:17 cshell: If it's the in-memory store, well, it's an in-memory store.

19:17 cshell: I'm using noir

19:17 that makes sense

19:17 is there an easy way to switch it?

19:17 weavejester: cshell: The in-memory store is the default. You'll probably want to change it. The cookie store is the easiest option, but you're limited on space.

19:18 cshell: I don't know how you'd change it in Noir. I assume there is a way.

19:18 cshell: what's the limit?

19:19 weavejester: cshell: I believe IE has a 4k limit on the size of cookies, which limits how much data you can store using the cookie session store.

19:20 cshell: so if you put it all in the cookie, you'd be vulnerable to replay/man in the middle attacks if they were able to obtain the cookie?

19:21 akhudek: cshell: I don't think you can't rely on the cookie information being secure. There are also various database backed session store libraries though.

19:22 cshell: okay, cool, thanks guys

19:29 ah, it looks like the cookie store can be encrypted

19:29 nice thinking Ring authors :)

19:36 weavejester: cshell: The cookie store is encrypted and signed with a HMAC. It's not really an option, as you wouldn't want to store any session data in plaintext.

19:37 cshell: Yeah, I noticed that it's there automatically - pretty nice!

19:40 hmm, still doesn't seem to be working when deployed to heroku on multi dynos - it seems to get the session on the server side only some of the time

19:46 akhudek: cshell: try deleting your browsers existing cookie

19:47 cshell: I did, it's still intermittent :(

19:47 on a single deployment it doesn't have the problem

19:51 hmm, i wonder if enlive is caching it on one and not the other

19:51 or in different states

19:51 akhudek: enlive shouldn't be caching anything other than templates

19:53 cshell: and even then it doesn't cache all templates, right?

19:56 akhudek: I think it caches anything defined with deftemplate or defsnippet

19:56 For dynamic parsing you should be using html-resource with emit*.

19:57 cshell: yeah, it looks like the 'at' form is assigning cacheable

20:32 omg

20:32 i bet it's because teh key i'm sending in is different on each server

20:33 since i'm not passing it, it's generating separate ones

20:50 lambda-stu: i'm trying to use a jar I've copied into my lib folder with lein 2.0.0-preview7, but it's not showing up with lein classpath. any ideas?

20:51 cshell: yes

20:52 lein 2 doesn't use the lib directory anymore

20:52 it uses your local maven repository

20:52 you might be able to enable some sort of local directory

20:53 xeqi: lambda-stu: lein2 uses deps straight from ~/.m2. You can use lein-localrepo to put it there

20:53 though you break repeatability

20:53 ~repeatability

20:53 clojurebot: repeatability is crucial for builds, see https://github.com/technomancy/leiningen/wiki/Repeatability

20:55 lambda-stu: thanks! i'll take a look

22:30 sridatta: hey is anyone here familiar with the "dieter" middleware for ring?

22:32 weavejester: dieter's a little odd...

22:32 sridatta: what makes you say that?

22:34 weavejester: Well, Dieter is middleware that implements a Ruby-inspired middleware system on top of itself.

22:36 There doesn't appear to be any advantage to this extra layer.

22:37 sridatta: ah I see. I was mainly trying to use it to compile LESS stylesheets on the fly in my web app. do you know any better way of doing something like this?

22:40 weavejester: sridatta: Well, Dieter is more complete than any other interpretation I know of. I wrote a library, clj-less, that would just compile a LESS file from a source, and I was planning on writing Ring middleware for LESS compilation and for general caching.

22:41 sridatta: But I've yet to get round to do that. Dieter has a lot of useful code, it's just wrapped up in a really odd structure :/

22:41 sridatta: hehe yeah I recall coming across your clj-less repo while searching, actually. was wondering why your handle seemed familiar...

22:42 Dieter seems to be broken for me though. it throws parse errors on Less files that the lessc compiler successfully compile :(

22:42 Peregrine: A fairly large number of the more prolific clojure library creators hang out here.

22:43 So a lot of names will appear familiar

22:43 rbxbx: Hey all, I was just following along with https://github.com/swannodette/logic-tutorial and got to this example:

22:43 weavejester: sridatta: See if clj-less does any better. Unfortunately I haven't written up the README, but I believe the library is on Clojars.

22:43 rbxbx: (run* [q]

22:43 (conde

22:43 ((fun q))

22:43 ((likes q 'Mary))))

22:44 why the double parens around the forms inside of conde?

22:44 Does (fun q) return a function which we're then calling?

22:44 (same with likes)

22:44 Peregrine: I'd open up a repl and experiment with removing those parens

22:45 and see what you get.

22:45 rbxbx: ArityException Wrong number of args (1) passed to likes

22:46 it does look like likes and fun return functions though... so yes. Repl'ing up good.

22:46 I would think you'd want to pass the functions unevaluated to conde though, hm.

22:46 weavejester: rbxbx: That example is a little confusing. It would be better if it used a vector, rather than a list.

22:47 rbxbx: So (conde [(fun q)] [(likes q 'Mary)])

22:48 It's basically a logical OR

22:48 rbxbx: weavejester gotcha. That works as well. Still not sure I grok it fully though.

22:48 as in, I understand what it's doing, but I don't quite get the invocation

22:48 weavejester: So: (conde [goal1 goal2] [goal3 goal4]) is (goal1 AND goal2) OR (goal3 AND goal4)

22:49 rbxbx: ah

22:49 That clarifies. Thanks weavejester :)

22:49 weavejester: In the same way that (run* q goal1 goal2 goal3) is goal1 AND goal2 AND goal3

22:49 rbxbx: *nod*

22:50 joshua__: $findfn [1 2 3] [2 3] [1]

22:50 lazybot: []

22:50 joshua__: $findfn [2 3] [1 2 3] [1]

22:50 lazybot: []

23:05 tos9: a/5

23:16 rbarraud: Who's watching the NASA MSL Curiosity probe coverage?

23:16 jhulten: rbarraud: you have a good link?

23:17 rbarraud: HD Internet, yep, seems good.... Dunno whether it's Akamaized here in NZ or not

23:18 Jhulten where u?

23:56 Roxxi: Can anyone tell me where standard out goes when using Slime/swank with emacs?

Logging service provided by n01se.net