#clojure log - Dec 14 2009

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

0:50 tolstoy: how do you get Ant to print out the classpath it thinks it has?

0:50 Oops! Nevermind! Getting a useful error message! Yay!

2:44 ordnungswidrig: g'morning, all

3:24 LauJensen: Morning team :)

3:55 Where are the results for the Wide Finder 2 runs ? For all languages?

3:58 Ah got it

4:14 Is Danny from Dannynet present here today?

4:30 G0SUB: hey

4:30 I am facing some problems in downloading compojure using lein + clojars

4:31 compojure depends on clojure-1.1.0-alpha

4:31 there is nothing called clojure-1.1.0-alpha

4:31 it's either 1.1.0-master or 1.1.0-new

4:32 hiredman: well, you might be able to go into lein's cache (~/.m2/ maybe?) and just change the name of a jar file

4:36 G0SUB: hiredman: hmm, but why does compojure depend on a non-existent jar?

4:39 hiredman: G0SUB: the name of the jars built was changed to include the name of the branch

4:40 happened a few days ago

4:40 G0SUB: hmm

4:41 hiredman: any idea when compojure will get fixed?

4:41 hiredman: nope

4:42 I think I might have heard something about a recent version of lein that somehow works around this, but I could be wrong

4:42 G0SUB: hmm

5:06 hiredman: where can I get the project.clj file used for compojure?

5:08 hiredman: no idea

5:08 G0SUB: ok

6:02 which is the best FTP client for Java?

6:14 cark: i'm using apache commons ftp

6:15 G0SUB: cark: any nice Clojure wrappers around it?

6:15 cark: not that i know

6:15 G0SUB: ok

6:15 cark: it wouldn't be that usefull anyways

6:17 G0SUB: cark: found one http://github.com/kyleburton/sandbox/blob/master/clojure-utils/kburton-clojure-utils/src/main/clj/com/github/kyleburton/sandbox/ftp.clj :)

6:18 cark: nice =)

7:17 fliebel: Woha! This morning I woke up thinking about an application as fluid going though a set of pipelines and filters from input to output. Like… streaming data instead of calling functions. I invented flow programming, looked it up on google, found this: http://en.wikipedia.org/wiki/Flow-based_programming This actually fits quite nicely in the lazy functional style of clojure, doesn't it?

7:21 actually more like this maybe: http://en.wikipedia.org/wiki/Dataflow_programming

7:27 karmazilla: I wonder how common it is to have this experience where you reinvent dataflow programming. I've tried it too.

7:32 lpetit: ,(doc promise)

7:32 clojurebot: "([]); Experimental. Returns a promise object that can be read with deref/@, and set, once only, with deliver. Calls to deref/@ prior to delivery will block. All subsequent derefs will return the same delivered value without blocking."

7:37 fliebel: karmazilla: I'm not sure about that… anyone else had the same?

7:38 lpetit: that is something like a lazy var, right? I'm not sure if I understand that...

7:39 lpetit: fliebel: it enables threads to synchronize on the availability of data. Sort of "dataflow" building block

7:39 fliebel: ,(doc deliver)

7:39 clojurebot: "([promise val]); Experimental. Delivers the supplied value to the promise, releasing any pending derefs. A subsequent call to deliver on a promise will throw an exception."

7:40 fliebel: lpetit: So I set up a whole train of functions to use that object, and the 'release' it with the input source?

7:42 lpetit: fliebel: something like that, but I'm not expert. Just remembered this having been mentioned in the ml. I guess if there are several functions waiting for the delivery of the value, they must be in different threads.

7:42 karmazilla: lpetit: Neat. I've been annoyed that Java doesn't have a standard Promise class.

7:43 fliebel: lpetit: yea, you'd need treads… or you'd need to be able to return something and then happily continue computations.

7:43 lpetit: fliebel: there is this :http://richhickey.github.com/clojure-contrib/dataflow-api.html , I don't know it much, though.

7:44 fliebel: lpetit: sounds cool

7:45 I've found another Wikipedia page about something like this: http://en.wikipedia.org/wiki/Functional_reactive_programming I think it might be more appropriate for a functional language…

7:55 lpetit: fliebel: yes, there so much interesting stuff to read. this one is close in my todo, also :)

7:57 fliebel: lpetit: agreed! :D

8:08 It seems like such a powerful paradigm, yet I can find out so little about it. Some obscure languages doing it and some packages adding it to existing languages, most likely not designed with this paradigm in mind.

8:13 lpetit: fliebel: there's an introductory book on haskell which as a comprehensive example using this (if I remember correctly), but I didn't manage to get through the book past this example, since: 1./ I had little time to read the book, and 2./ I rediscovered clojure in the mean time :-)

8:14 fliebel : there it is : http://www.haskell.org/soe/

8:16 fliebel: lpetit: thanks

8:32 LauJensen: I apologize for the OT question, but Google is let me down. Does anyone know of a time-tracker with similar functionality as org-mode, which just sits in a little linux desktop app instead of Emacs? (not Hamster)

8:37 ohpauleez: LauJensen: I know of a Java-based time tracker app I thought looked real nice, but I never ended up using

8:37 let me see if I can dig it up

8:37 LauJensen: ohpauleez: Makagiga? I'm looking at it right now

8:37 ohpauleez: LauJensen: http://rachota.sourceforge.net/en/index.html

8:37 LauJensen: ah ok, thanks

8:38 ohpauleez: totally welcome

8:50 LauJensen: Is there a way to pass the working directory as a command-line-arg to Java? java -jar /myproj/start.jar -wp /myproj ?

8:57 fliebel: I think you're supposed to cd there. If you only need the classes you can of course add it to cp.

8:57 LauJensen: k

9:10 fliebel: lpetit: I found a new language! http://www.mozart-oz.org/ Not nearly as cute and clean as Clojure and Python, but it looks powerful.

9:15 lpetit: fliebel: yes, it's the language the "CTM" book is based upon

9:15 fliebel: "Concepts, Techniques and Models" of programming. Great book, bought it, 940 pages, read ~ 50 at the moment :-)

9:16 fliebel: http://www.amazon.com/Concepts-Techniques-Models-Computer-Programming/dp/0262220695/ref=sr_1_1?ie=UTF8&s=books&qid=1260800058&sr=8-1

9:16 fliebel: cool

9:18 lpetit: fliebel: honestly, I find it a little bit hard to follow. But certainly worth reading it because it does not impose a certain vision (object oriented, purely functional, imperative, declarative, etc) on you, but rather seems to present all those alternatives and compares them

9:19 fliebel: by "compares", I don't mean with the goal of having some final word such as "which one is the best", but rather "which one works the best for which kind of situation"

9:20 fliebel: lpetit: so you know oz?

9:20 lpetit: fliebel: just by name, and because it's the language used for the examples in the CTM book.

9:23 fliebel: they use the notion of a "kernel language", with just enough primitives to allow write other primitives on top of it. They present the programming paradigms one by one, showing adding the missing primitives to the kernel language as they go, showing which notion is missing to be able to introduce each new paradigm.

9:23 fliebel: sometimes I which I had not a job and could just sit at home in front of my laptop and with the dozen or so books I've bought but not yet read :-)

9:24 LauJensen: lpetit: sounds like an interesting comparison :)

9:25 lpetit: LauJensen: talking about my desire to have more time to study, or the previous post describing CTM ? :-)

9:26 LauJensen: "You'd want Objects when you need complexity which is unmanageable and no real tool for handling time. On the other hand, go with Functional paradigms for the cases where you dont want bugs and need a smaller codebase."

9:26 Your 'book review'

9:28 lpetit: LauJensen: where did you see this citation ?

9:28 LauJensen: I was just imagining what 'compares in terms of what to use when and not which is better' looks like written out in a book

9:29 lpetit: LauJensen: when I eventually finish the book (hopefully in 10 months, more certainly in 10 years) I promise I give you the answer :-)

9:30 LauJensen: I'll look forward to it :)

9:57 Tommock: hello

9:57 can anyone figure out what is wrong with this http://pastebin.org/64723 ?

9:59 chouser: ((if ...)) will try to execute the return value of the 'if'

9:59 perhaps you want (if ...), this is only a single pair of parens

10:03 Tommock: chouser: oh so easy, thanks a lot!

10:04 chouser: np

10:04 Tommock: been trying to figure out what whas wrong for an hour or so now :)

10:04 chouser: out of curiousity, where'd the extra parens come from in the first place?

10:05 Tommock: my reasoning probably whent something like "the body of the function should be inclosed by something

10:05 chouser: ok

10:05 it jumped out at me because (( is pretty rare in Clojure code

10:05 Tommock: chouser: how is my indent style? looks ok?

10:05 not sure how things should look

10:07 chouser: The body of 'do' should be indented another level

10:07 that way the 'then' and 'else' parts of 'if' stand out as the only two things at the same level under 'if'.

10:08 other than that it's ok, though I'd recommend spaces instead of tabs -- 2 spaces per level is most common, though 4 is probably ok.

10:09 Tommock: chouser: my personal preference is tabs but thanks for the advice :D

10:10 well I

10:10 'm off then

10:10 chouser: bye

10:21 rhickey: http://clojure.org/funding

10:22 chouser: rhickey: might be good to have a donate button (or link to one) on that page.

10:22 rhickey: I added a link back to the home page

10:22 chouser: oh ok, I see that now.

10:22 rhickey: could put it right there...

10:25 LauJensen: rhickey: I think that a public indication of how much support you raise would be motivational for private contributors, how would you feel about that?

10:25 rhickey: ok, Donate button right on funding page now

10:26 LauJensen: I am asking people if they want to be listed on a Donors page, will set that up soon

10:27 LauJensen: rhickey: Not quite was I was thinking, but something like Wikipedias progressbars.

10:27 rhickey: LauJensen: no

10:27 lpetit: LauJensen: you mean a progress bar towards a full year of work from Rich funding ?

10:28 LauJensen: Yes something like that

10:28 I always find that stuff motivational

10:38 cemerick: rhickey: A donors page would be good. Mix that with the opportunity to list job openings from sponsors, perhaps.

10:42 rhickey: cemerick: you've mentioned that before - what's the idea there?

10:44 cemerick: rhickey: well, from the perspective of larger sponsors (whatever that means), being able to highlight that connection and therefore perhaps attract highly self-screened candidates (programmers already looking at the clojure.org site) would be an obvious advantage compared to existing recruitment paths

10:44 icey: Python uses "sponsor members" for corporate donors: http://python.org/psf/#id2

10:44 maybe something like that?

10:45 cemerick: for you and the broader clojure community, being able to show that there is an active job market lends to the language's credibility in all sorts of ways

10:46 rhickey: I think it would be totally reasonable for you to host a clojure-centric job board, charge some nominal amount to submit listings (which obviously helps with the funding), and then sponsors' listings are always at the top, or highlighted, or whatever.

10:47 icey: yeah, something like that, although the presentation there is...lacking

10:47 rhickey: cemerick: I'm all for job listings, I guess my initial pang is that limiting it to sponsors might be excluding. OTOH, once you are hiring Clojure devs, sponsoring is a good idea. This would be for Clojure-related jobs only, right?

10:48 icey: cemerick: well I'm not suggesting using their design ;) However, I do know that the donors page has turned into a good recruitment tool for companies that want to show that they support the language

10:48 chouser: well. Clojure and PHP of course.

10:48 rhickey: :)

10:48 angerman: phew ... I can continue to use clojure.

10:48 cemerick: rhickey: oh, I would never say that listings should be limited to sponsors. I would say that a sponsorship of $X or more would get the organization in question N "free" job listings for the following year or whatever.

10:48 angerman: ;;)

10:49 cemerick: rhickey: "clojure-centric" is a reasonable phrase, although the meaning is obviously flexible.

10:49 * rhickey imagines a world where people are competing for placement at the top of Clojure job listings

10:49 chouser: :-)

10:49 cemerick: heh

10:50 angerman: ~t-test (:ns incanter.stats)

10:50 clojurebot: I don't understand.

10:50 angerman: :/

10:51 cemerick: rhickey: it's obviously going to be a little thin for some time, so until there's a regular deal flow, as it were, just keeping a list of available jobs on a static page would be fine :-)

10:51 chouser: angerman: I'm curious about your comment about continuing to use clojure.

10:51 icey: You could always threaten people with adoption of a bizarre Franz-like licensing model for forward versions ;)

10:51 angerman: chouser: I'm student at TU Munich

10:51 cemerick: ouch

10:51 angerman: money is quite shallow

10:51 chouser: angerman: ah, ok. That was my guess.

10:52 angerman: My stupid questions reveal my identity I guess, he

10:53 chouser: angerman: oh, nonsense.

10:54 angerman: well, Google Talk on YouTube in 10 minutes. bye

10:55 the-kenny: Google Talk on Youtube?

10:55 maddis: perhaps a talk at Google live on YouTube?

10:55 angerman: yes. They host it in our biggest lecture hall at the Mathematics and Computerscience Bldg. Campus Garching.

10:56 * _fogus_ wonders how to do this without involving Paypal

10:57 the-kenny: Oh, Google holds a talk about Youtube? I read it "Google Talk (the Xmpp-Service) on Youtube"

10:57 angerman: yes

11:02 chouser: it might be fun to add "I paid for Clojure" to your signature on google group posts.

11:03 cemerick: ah, good idea

11:03 icey: have you considered forming a nonprofit (501(c)3) for the Clojure finances?

11:03 cemerick: if licensing won't do it, if entreaties won't do it, then shame will! :-)

11:03 esj: cemerick: you must be Catholic :p

11:03 cemerick: heh, hardly

11:04 but shame is a *wonderful* motivator.

11:04 internet anonymity makes it slightly harder to pull off, but it's worth a shot

11:09 rhickey: chouser: Yes, "I paid for Clojure" is a fine idea

11:09 shoover: As distinct from a "Clojure was my idea" marketing campaign

11:10 rhickey: :)

11:10 shoover: Only authors of books and langs on the Clojure reading list can put that

11:11 the-kenny: Maybe everyone who donates can get a handwritten closing-parenthesis from rhickey, xkcd does something like this with scans from their comics.

11:12 chouser: haha! a closing paren.

11:12 _fogus_: If you donate above a certain amount you can get a matching opening paren

11:13 Chousuke: what about the other data structures? :/

11:13 icey: I wonder how much effort it would take to create one of those laminated 1-page "cheatsheets" you see out there, and sell them from the website. The language is small enough that the pertinent information should fit

11:15 cemerick: little chotchkies are nice, but I'll wager making them would be a huge PITA for rhickey *shrug*

11:16 icey: cemerick: yeah, that's what i was concerned about

11:16 stuartsierra: You'd want to outsource all the printing/delivery to something like cafepress.

11:16 devlinsf: cemerick: So do what PBS does, and "sell" them for $50 each

11:17 stuartsierra: devlinsf: Yes!

11:17 icey: although, along the cafepress route, i'd certainly buy a clojure logo coffee cup. but that's probably just the caffiene addict in me speaking

11:17 devlinsf: icey: :)

11:17 chouser: there are already clojure T-shirts

11:17 icey: chouser: they don't hold my coffee well :(

11:18 chouser: to date they've brought in about enough money to pay for about 30mins of rhickey clojure development time. :-)

11:19 which granted is long enough to fully develop a feature like 'trampoline', but still...

11:19 the-kenny: icey: But you can drain them in coffee and absorb the coffeine with your skin

11:19 icey: I definitely agree that straight donations are the most efficient way to go; but there will always be a number of people who won't donate straight cash but will buy little toys or whatever. no idea why, but i know they're out there

11:20 _fogus_: chouser: Are those the "State: YDiW" shirts?

11:20 chouser: _fogus_: there are a few styles.

11:21 http://www.zazzle.com/clojure

11:21 "don't be so eager"

11:21 "lisp is not dead. The url is now clojure.org"

11:21 "I get more done when I'm lazy"

11:22 _fogus_: Ahhh. I got the state shirt a long time ago but didn't know there were more to choose from. Looks like I will never again need to wear non-Clojure clothing.

11:23 chouser: heh.

11:23 oh, you bought it from zazzle? How was the print quality? I haven't seen one yet.

11:24 _fogus_: Very nice actually. Good quality fabric also. I've had it since perhaps day 2 and it's still in great shape.

11:26 maravillas: certainly there must be a threads pun appropriate for a tshirt

11:26 hiredman: ping?

11:26 clojurebot: PONG!

11:27 hiredman: aww, come on

11:27 there we go

11:29 esj: is it idiomatic to put your tests in .clj where the target functions are defined ?

11:30 devlinsf: esj: how are you testing?

11:30 esj: clojure.test

11:30 devlinsf: My opinion is no

11:30 If you were using the meta, I'd say yes

11:31 You're talking about putting it all in the same ns, right?

11:31 esj: Thanks. I tend to agree. I want to put them in a seperate file and then (use it. I'm missing something though, because then I run into ns issues.

11:31 aaah, devlinsf, read my mind

11:32 devlinsf: Well, I do the following w/ my code

11:32 I have (ns a.long.ns ... and I define a test lib. Usually (ns test.a.long.ns

11:32 Also depends one the code

11:32 esj: ok

11:33 devlinsf: A lot of my stuff has a lib prefix in the ns

11:33 stuartsierra: esj: It's tempting to try to put tests and fns in the same file, I even tried to support it with clojure.test. But it ends up being messy in practice.

11:33 esj: then test.a.a.long.ns uses a.long.ns and badda bind ?

11:33 devlinsf: badda?

11:33 esj: s/badda bind/badda-bing/

11:33 devlinsf: Oh, badda nibg

11:33 bing

11:33 Yes

11:33 esj: :)

11:34 ok, thanks gents. This time I'm reall, REALLY, going to write the tests for my code...

11:34 devlinsf: check out contrib's structure

11:34 there's clojure.contrib.a-ns

11:34 esj: devlinsf: great idea, shoulda thought of that.

11:34 devlinsf: and usually clojure.contrib.test.a-ns

11:35 Some libs are different, though

11:35 esj: will do.

11:35 patrkris: I'm reviewing some of the articles that are critical of STM, and wondered if there is a fundamental difference between the transaction systems found in database systems that makes transactions viable in those and not in programming languages?

11:35 devlinsf: esj: Good luck

11:36 chouser: patrkris: STM is a very broad category of functionality. Don't assume that critisizm of one STM implementation applies to another.

11:37 esj: devlinsf: thanks.

11:37 arohner: patrkris: and even between programming languages, the functionality in one language STM is very different from another

11:38 most of the "STM is a failure" articles you see talk about cramming STM into Java / C#, with little to no change in syntax or semantics

11:39 patrkris: chouser, arohner: yes, but reading "STM: Why is it only a research toy?" gives the impression, that there is something fundamentally wrong with STM - but I see so many things in that article that doesn't apply to Clojure's implementation

11:39 arohner: STM in java with lots of mutable state everywhere and no change in language semantics is a very different beast that building STM in at the start to a functional language with immutable data

11:40 patrkris: arohner: yes you are right

11:40 arohner: the proof is still in the pudding. their articles really just say "*we* couldn't get STM to work". Look at clojure yourself, and decide if clojure got STM to work. :-)

11:41 patrkris: yes... their focus is also on performance, it appears. Doesn't matter whether your programs are correct, I suppose :)

11:41 chouser: for one, because of the immutable collections and other reference types, you can actually do quite a lot of even multithreaded work in Clojure without ever using the STM.

11:41 dabd: where is the macro 'defprotocol' defined?

11:42 chouser: dabd: in the 'new' branch

11:42 patrkris: chouser: yes, that is one of the things I am going to mention in what I'm writing

11:42 dabd: chouser: I'm trying to compile the enclojure plugin and it fails because it uses this macro

11:45 liwp: dabd: defprotocol is defined only in the 'new' branch of the clojure github repository, so if you're using the 'master' branch or the 1.0 version of clojure you won't have the defprotocol macro

11:46 dabd: you'll have to move to the 'new' branch to be able to use the compojure plugin if it depends on defprotocol

11:46 hiredman: liwp: enclojure

11:46 liwp: uhh, sorry

11:47 hiredman: it's odd that enclojure would depend on the new branch

11:47 chouser: that's what I was thinking.

11:47 dabd: maybe you should try a slightly older version of enclojure?

11:48 dabd: chouser: yes currently I'm unable to use enclojure because it fails to start a project REPL

11:48 so I was trying to build it from source to see what is happening

11:49 cemerick: rhickey: to what degree are you going to report on fundraising efforts? I know that's getting very close to your personal finances, but being able to say "X has been raised across N contributions averaging K per contribution" would be interesting/encouraging.

11:50 dabd: hiredman: you can see defprotocol being used here in the enclojure source http://github.com/EricThorsen/enclojure/blob/master/src/protocols/src/org/enclojure/protocols.clj

11:51 rhickey: cemerick: maybe X% of a year of funding? and a graph/chart?

11:52 cemerick: rhickey: sure. A normalized distribution of contribution amounts would be decent.

11:53 People could infer a lot from something like that, so I wouldn't blame you for not sharing to that degree.

11:53 icey: rhickey: the only real problem with stating a goal like that is that I think you'll end up less likely to recieve donations as you near your goal

11:54 although it's not bad for the time you're under 50% of your goal

11:54 rhickey: icey: true

11:54 cemerick: someone mentioned a progress bar before though, which can be motivating for some

11:55 Chousuke: what's the minimum donation that benefits you, by the way?

11:56 icey: hmmm... if you made the progress bar the goal for 2 years instead of one, it'd probably be less of an issue - even if you exceeded the goal for 1 year, it would give you the chance to start raising for the next year

11:56 and if you could raise 2 years worth of funding inside of a 1 year window, you could change the verbiage

11:57 * reify donated this morning

11:59 * chouser is still tussling with paypal.

12:01 icey: I am curious if today's push will pay for at least a month of development when all is said & done.

12:21 reify: time to ebay all my non-essential computer equipment =)

12:21 clojure, the "cloud", is all i need

12:23 _fogus_: Hacker News seems to think that Clojure was created by some guy named Rick.

12:31 maddis: i have a remote REPL running through leiningen ('lein swank') and connected to it through M-x slime-connect, but I can't figure out how to compile/run local code I have open in another buffer through C-c C-k (or the slime-compile* commands). and tips on that?

12:32 the-kenny: maddis: Maybe with M-x slime-connection-list?

12:32 It should be possible to make a connection the "default" one in that buffer.

12:41 maddis: hmm, on C-c C-k it tires to find the local file on the remote location, which is not there.

12:41 1 compiler notes: Unknown location: error: java.io.FileNotFoundException: /home/matthias/da/code/garec/src/garec.clj (No such file or directory). Compilation failed.

12:42 or am I completely wrong here and this all works different? :)

12:46 the-kenny: maddis: Hm.. then try M-x eval-region

12:46 uhm

12:47 slime-eval-region or slime-eval-buffer

12:49 maddis: ok, that works. thanks

13:00 the-kenny: What are the naming-conventions for the defproject from leiningen, if I want to push it to clojars>

13:01 patrkris: are there any case studies available of large applications with a high degree of concurrency using Clojure (and especially the STM) succesfully?

13:03 maddis: patrkris, perhaps this will help you: http://portal.acm.org/citation.cfm?id=1562870

13:04 patrkris: maddis: thanks

13:49 qed: where can we donate to clojure?

13:49 hiredman: clojure.org

13:49 qed: oh, right :)

13:49 thanks

13:50 KirinDave: I hope Rich will disclose how much he makes in a day from this campaign. ;)

13:54 qed: KirinDave: heh

14:09 headius: technomancy: ping

14:15 djork: if I spent as much on Clojure as my wife did on her hair rhickey would be better off

14:15 or at least his wife could get her hair done

14:16 LauJensen: Is Matt of Conjure fame here tonight?

14:20 technomancy: headius: ICMP_ECHO_REPLY

14:20 headius: hi there!

14:21 I was going to suggest we should bundle up a gem that's just clojure and release it

14:21 so other gems could depend on that one

14:21 SergeyDidenko: Hi, could anyone advise the workaround for the absent "byte-array"? - (byte-array [1 2 3])

14:21 headius: I want to do a pass at wrapping the collections with some Ruby stuff

14:21 technomancy: headius: oh yeah, sure. I don't think my gem ever made it out to any servers outside github, so there shouldn't be any overlap issues.

14:21 headius: ok

14:21 SergeyDidenko: Is "make-array" + "aset" the only way?

14:22 cemerick: SergeyDidenko: (into-array Byte/TYPE [1 2 3])

14:22 technomancy: I never quite got that exception-wrapping problem solved for transaction retries.

14:22 cemerick: ,(into-array Byte/TYPE [1 2 3])

14:22 clojurebot: java.lang.IllegalArgumentException: argument type mismatch

14:22 cemerick: hrm

14:23 yuck

14:23 ,(into-array Integer/TYPE [1 2 3])

14:23 clojurebot: #<int[] [I@1a8de42>

14:23 cemerick: ,(into-array Byte/TYPE (map byte [1 2 3]))

14:23 clojurebot: #<byte[] [B@2b9edc>

14:23 cemerick: SergeyDidenko: Looks like ^^ is the best you can get right now

14:24 chouser: ,byte-array

14:24 clojurebot: java.lang.Exception: Unable to resolve symbol: byte-array in this context

14:25 SergeyDidenko: cemerick, strange it does not work on my computer. Probably I need to get the latest sources

14:25 hiredman: SergeyDidenko: that should work

14:25 what exception are you seeing?

14:25 cemerick: ,(into-array Integer/TYPE (map byte [1 2 3]))

14:25 clojurebot: #<int[] [I@1994498>

14:25 SergeyDidenko: cemerick, do you use the recent Clojure version ?

14:26 cemerick: SergeyDidenko: I'm generally on 1.1-new, although that should work fine even in 1.0 (I think)

14:27 SergeyDidenko: (into-array Byte/TYPE [1 2 3]) throws java.lang.IllegalArgumentException: argument type mismatch (NO_SOURCE_FILE:0)

14:27 lghtng: I started out wanting to just contribute a few sparql wrappers as a learning project and now im sitting here reading Marvin Minsky's "Semantic Information Processing" c.1968: Is there a word for that?

14:27 danm_: SergeyDidenko: missed the map byte part

14:27 chouser: 1.1-alpha and 1.1-new should have byte-array now.

14:27 cemerick: ah, (into-array Byte/TYPE [1 2 3]) would work if a type.cast() call were made in RT.seqToTypedArray

14:27 SergeyDidenko: thank you guys :)

14:27 chouser: lghtng: enlightenment?

14:28 cemerick: although I suppose we wouldn't want the overhead

14:28 lghtng: works for me

14:34 KirinDave: I think I may have a functioning webapp using compojure. Whee.

14:34 What compojure needs, what every webapp framework needs, is a generator.

14:35 chouser: nonono

14:35 cemerick: ooh, this'll be fun :-)

14:35 KirinDave: ?

14:35 Chousuke: generator? :/

14:35 lghtng: transformer

14:35 KirinDave: Like, lein new.

14:36 Chousuke: aha.

14:36 KirinDave: Something that does mundane shit like drop a shell script that can help target public and things like that.

14:36 Chousuke: write a lein plugin!

14:36 cemerick: KirinDave: don't mind me, I've just always enjoyed the dynamic vs. static site debate.

14:36 the-kenny: Should be easy to write a leiningen plugin for this :)

14:36 Chousuke: nein new-compojure :P

14:36 KirinDave: Sure.

14:36 Chousuke: lein* too

14:38 hiredman: http://github.com/hiredman/appengine-helper

14:41 the-kenny: Google should allow Threads for clojure :(

14:41 stuartsierra: My take on datatypes/protocols: http://stuartsierra.com/2009/12/14/objects-are-not-adts

14:44 KirinDave: Man, the power of SSDs

14:44 it takes less time to start my app in -server on my laptop than on this big server.

14:44 maacl: Does anyone know of any Crane usgae examples ?

14:54 jgracin: I see that add-classpath has been deprecated. Is there some new way of modifying classpath at runtime?

14:56 the-kenny: jgracin: IIRC, changing the classpath just causes problems, I wouldn't do it.

14:56 danm_: all I've read is that you can't

14:57 so if you want to change the classpath, restart clojure

14:57 the-kenny: danm_: There *was* add-classpath in clojure, but it just caused problems, so it's deprecated now.

14:57 danm_: yeah

14:57 it never seemed to work for me anyway

14:57 stuartsierra: It's a problem with how classpaths are usually implemented in Java.

14:58 The root classpath is created when the JVM starts and cannot be changed.

14:59 hiredman: stuartsierra: would replacing the system classpath fix that?

14:59 er

14:59 stuartsierra: hiredman: You can't replace the system classpath without hacking the JVM, I think.

14:59 hiredman: system classloader

14:59 stuartsierra: You can replace the system classloader I think.

14:59 hiredman: -Djava.system.class.path

14:59 stuartsierra: Yes. But that's real JVM voodoo right there.

15:00 hiredman: http://gist.github.com/255766

15:01 danm_: I'd like to see people who had trouble with add-classpath try this out

15:01 rhickey: java.system.class.loader

15:01 stuartsierra: Like I said, voodoo.

15:01 hiredman: rhickey: right!

15:02 rhickey: would that fix the visibility problems from add-classpath?

15:02 jgracin: I have a standalone application which needs to contact a configuration registry system and find out which plugins to load.

15:02 rhickey: I think it would be generally useful for someone to make a solution based upon that, just for development

15:02 the big problem is that people will build runtime dependencies on it, and that can't be satisfied in all runtime environments

15:03 hiredman: :| like the above mentioned plugin system

15:03 rhickey: hiredman: yes, it would not have those problems, but it can't be used in all situations, i.e. when you don't control startup

15:03 hiredman: right

15:04 I was thinking along the lines of lien

15:04 lein?

15:04 stuartsierra: See? No one can spell it!

15:04 hiredman: :P

15:05 lghtng: tangential question: can the classpath be modified via emacs/slime?

15:05 hiredman: anyway, that gist above generates a cl.jar (that contains a classloader class) and prints out (cryptic) usage instructions

15:05 stuartsierra: lghtng: no, same problem

15:08 fanatico: ,seen technomancy

15:08 clojurebot: java.lang.Exception: Unable to resolve symbol: seen in this context

15:08 lghtng: .seen technomancy

15:08 hiredman: clojurebot: seen technomancy?

15:08 clojurebot: technomancy was last seen quiting IRC, 18 minutes ago

15:09 fanatico: thanks.

15:11 qed: Just sent my relatives an email asking they donate to clojure instead of buying me presents or gift certificates.

15:11 Hopefully they haven't gotten me anything yet. :)

15:15 lghtng: http://en.wikipedia.org/wiki/GNU_Classpath

15:16 hiredman: :|

15:22 maddis: qed, that is a really good idea! don't think my parents/relatives understand that, but it's worth a try :)

15:23 qed: my mom wont have a clue, but my dad actually asked me if clojure was like dylan

15:23 i was pretty surprised he knew about dylan :)

15:24 the-kenny: qed: Bob Dylan? :P

15:24 qed: haha

15:24 nah, you know, the programming language

15:25 he's an old mac nerd, worked w/ Kai Krause and Eric Wenger once upon a time

15:25 maddis: i thought of Dylan from 90210, the tv series ;P

15:25 qed: hahaha

15:26 the-kenny: My father asks me every two days if and how he can copy a folder into another folder.

15:26 qed: do you tell him `cp -R`

15:26 as in "you'll need `cp -R` if you ask me that again"

15:26 the-kenny: ...he is using windows

15:26 qed: get him on ubuntu and show him how to google

15:27 the-kenny: I told him if he want me to keep supporting him with computer related things, he should either buy a mac or let me install linux.

15:27 qed: i really think that there needs to be some google instruction for people

15:27 the-kenny: We agreed to do this when his current laptop dies

15:27 qed: nice

15:27 the-kenny: (Which should happen in the next months :))

15:28 He's also using an old pc with a pentium 2 in his office

15:28 windows 2000

15:33 qed: That is..old

15:33 the-kenny: Yes

15:33 qed: There are still some machines we support that run Win2k unfortunately

15:34 the-kenny: and full of viruses ;)

15:34 It's time to replace it.

15:39 I had a dream last night about a cool project in clojure and something with tagsoup.... But I forgot what it was :(

15:50 lghtng: would be interesting to see clojure running inside jnode or looking glass

15:50 or as an appliance in jnode

15:51 hiredman: maxine!

15:51 guestos vm!

15:51 qed: looking glass?

15:52 is that a way to view router configs?

15:53 lghtng: http://research.sun.com/projects/maxine/max.html

15:53 lg3d has been pronounced dead by its founder

15:54 the deb repositories dont work but there is still an .iso built on pclinuxos

15:57 lisppaste8: the-kenny pasted "lein deps problems" at http://paste.lisp.org/display/92100

15:57 the-kenny: Can someone tell me where's the problem there?

15:58 hiredman: stuff got renamed a few days ago yadda yadda yadda

15:58 qed: someone was talking about this the other day

15:58 yeah

15:58 i think chouser is the guy who renamed some stuff

15:58 cemerick: darn that chouser :-P

15:58 qed: the-kenny: your best bet is to just download them and put them in ~/.m2

15:59 hiredman: clojurebot: rename?

15:59 clojurebot: Huh?

15:59 hiredman: clojurebot: rename is <reply>clojure-alpha was renamed clojure-master, and this broke most of the projects on Clojars

15:59 clojurebot: Ik begrijp

16:00 qed: clojurebot: rename?

16:00 clojurebot: clojure-alpha was renamed clojure-master, and this broke most of the projects on Clojars

16:00 qed: clojurebot: @markov chouser

16:00 clojurebot: Who??

16:00 qed: heh, worth a shot...

16:02 the-kenny: hm.. how long will it take clojars to normalize again?

16:03 hiredman: I imagine it is up to the projects to update their config and push

16:04 the-kenny: Yeah, got that.. maybe someone should write about this in the mailing list?

16:04 * drewr investigates his line-seq bug

16:05 hiredman: has been

16:05 like a week ago, or something, when the rename came up

16:06 the-kenny: So [org.clojure/clojure "1.1.0-master-SNAPSHOT"] is the right thing to include in project.clj now?

16:06 hiredman: I suppose

16:10 the-kenny: _ato: Is that your compojure-jar on clojars?

16:17 qed: with file-seq, I get a coll of Files, im trying to break them up into a structure like xml

16:20 Intertricity: Is there a clojure equivalent to python's pickle?

16:22 maravillas: c.c.duck-streams/spit and /slurp maybe?

16:22 stuartsierra: No.

16:23 arohner: pr is closer, but still not comprehensive

16:23 stuartsierra: Intertricity: read/pr gets you 9/10 of the way there

16:23 Intertricity: stuartsierra, ahh thanks :)

16:24 qed: No matching method found: parse for class com.sun.org.apache.xerces.internal.jaxp.SAXParserImpl

16:26 hiredman: qed: read the docs on parse and pass it one of the specified classes

16:28 drewr: chouser: I'm happy to fix that *-seq issue

16:29 the buggy patch is from me :-/

16:33 qed: hiredman: I have something like (file-seq (java.io.File. "/path/to/dir")) -- parse looks like it takes Files, so why can't I parse that?

16:33 hiredman: ,(doc clojure.xml/parse)

16:33 clojurebot: "([s] [s startparse]); Parses and loads the source s, which can be a File, InputStream or String naming a URI. Returns a tree of the xml/element struct-map, which has the keys :tag, :attrs, and :content. and accessor fns tag, attrs, and content. Other parsers can be supplied by passing startparse, a fn taking a source and a ContentHandler and returning a parser"

16:34 hiredman: ,(doc file-seq)

16:34 clojurebot: "([dir]); A tree seq on java.io.Files"

16:34 qed: how would i turn what I have into an InputStream?

16:34 hiredman: hmmm

16:34 ~def file-seq

16:35 lghtng: ,(doc element)

16:35 clojurebot: No entiendo

16:35 qed: how could I turn it into an InputStream?

16:35 lghtng: ,(doc :tag)

16:35 clojurebot: java.lang.ClassCastException: clojure.lang.Keyword cannot be cast to clojure.lang.Symbol

16:35 lghtng: right

16:35 ,(doc InputStream)

16:35 hiredman: what does the output of file-seq look like?

16:35 clojurebot: Excuse me?

16:36 qed: (#<File /Users/defn/foo> #<File /Users/defn/foo/bar> #<File /Users/defn/foo/bar/bar.txt> #<File /Users/defn/foo/baz> #<File /Users/defn/foo/baz/baz.txt> #<File /Users/defn/foo/foo.txt>)

16:36 lghtng: ,(doc startparse)

16:36 clojurebot: Gabh mo leithscéal?

16:36 qed: (doc clojure.xml/startparse)

16:36 clojurebot: Titim gan éirí ort.

16:37 lghtng: ,(doc clojure.xml/startparse)

16:37 clojurebot: excusez-moi

16:37 qed: (doc clojure.xml/startparse-sax)

16:37 clojurebot: "([s ch]); "

16:37 hiredman: and how are you calling parse?

16:37 lghtng: ,(doc ddt)

16:37 clojurebot: Excuse me?

16:37 hiredman: lghtng: knock it off

16:37 lghtng: eep

16:38 qed: hiredman: i tried to just run (parse (file-seq...)), also tried to map it to see if that worked but got an exception about content and prolog

16:40 hiredman: a. file-seq does not return one of the types mentioned by xml/parse b. the other error most likely means you don't have a well formed xml file (the default parser is a stickler)

16:40 ignoring exceptions like that is a bad habit to get into

16:40 just the fact that you are getting a different exception is useful information

16:41 qed: i guess my question is, can i turn that file-seq output into an inputstream

16:41 hiredman: it won't matter if your xml is not well formed

16:41 which is most likely what that exception was trying to tell you

16:41 qed: ah, okay -- ill mess with it and try to figure out what's up with it

16:45 lghtng: does clojure do natural inheritance? like native OO?

16:45 Chousuke: Clojure is not very object-oriented.

16:46 lghtng: or does it rely on backdoor methods like CLOS did?

16:46 Chousuke: backdoor?

16:46 I'm not sure what you mean.

16:46 lghtng: we've got object orientation, only its all extentions

16:47 so inheritance is something you build, not included

16:47 Chousuke: the closest to a native OO construct in Clojure (not interop) would be defprotocol and deftype I guess.

16:47 the-kenny: _ato: ping

16:47 Chousuke: but even those emphasise *decoupling* objects from methods

16:47 chouser: also 'derive' and multimethods

16:48 hiredman: dispatch hierarchies are reified

16:48 lghtng: i know it can do all those things and alot more, im just wondering how much has already been done

16:48 chouser: multimethod dispatch has type inheritence built in.

16:49 lghtng: since this is all library building stuff going on here, there is a SHITLOAD of undocumented stuff thats available but unlisted

16:49 hiredman: not true

16:49 clojure.org

16:49 lghtng: we dont want any unnecessary duplication, right?

16:49 Chousuke: I'm not sure what you mean.

16:49 What are you looking for, exactly?

16:50 lghtng: just knowledge

16:50 hiredman: lghtng: I would avoid righting a oop library for clojure

16:50 I doubt anyone will use it

16:50 lghtng: i have no plan to, im just interested in the state of it's emergence

16:51 as it will emerge, just at what rate and how dominant it will be

16:51 chouser: what feature are you looking for that such a library would provide?

16:51 lghtng: its funny about that

16:51 Chousuke: I mean, if you really want to do the "traditional" kind of OOP, maybe Clojure isn't the best language. However, many things you would do using OOP in other languages are perfectly doable using non-OOP methods in Clojure.

16:51 lghtng: nothing in particular

16:51 not me, in general

16:51 hiredman: there is at least one oop library that has already been abandoned

16:52 ~rationale

16:52 clojurebot: rationale is http://clojure.org/rationale

16:52 Chousuke: OOP in clojure that is not completely whacked is essentially equivalent to having functions that take some "special" value as their first argument, anyway.

16:53 hiredman: "for Functional Programming"

16:53 clojurebot: oo?

16:53 clojurebot: OO is to programming what astrology is to astronomy

16:53 Chousuke: heh :P

16:53 hiredman: clojurebot: why oo sucks?

16:53 Chousuke: but I think protocols and deftype will satisfy most people's "OO" needs.

16:53 clojurebot: OO is to programming what astrology is to astronomy

16:53 lghtng: i've recommended clojure to developers as an alternative to java, so far the answers have ranged from 'what about side effects' to 'what about side effects'

16:53 hiredman: clojurebot: why oo sucks is <reply>http://www.sics.se/~joe/bluetail/vol1/v1_oo.html

16:53 clojurebot: Ik begrijp

16:54 hiredman: lghtng: you mean the questions

16:54 lghtng: 'ignore side effects at your peril' seems to be the moral that java programmers are offering

16:54 dnolen: lghtng: it might be helpful to look at some of the long threads discussing this stuff on the mailing list, there a couple actually.

16:54 Chousuke: lghtng: Clojure is not purely functional, so just write side-effecting functions.

16:54 drewr: does Sean Devlin hang out on here?

16:55 chouser: drewr: at times

16:55 lghtng: i should see if the guy will come here to speak with you himself

16:55 dnolen: lghtng: also don't understand the "what about side effects"? Clojure is firmly in the anti-side effects camp.

16:55 chouser: Did someone recommend ignoring side-effects? seems unlikely

16:55 hiredman: dnolen: but not purely functional

16:56 Chousuke: lghtng: clojure even provides time constructs for managing a real process (which has side-effects, in the form of changing state). :)

16:56 dnolen: hiredman: yes, of course.

16:57 lghtng: state transactional memory, right?

16:57 and if you want side effects you use 'do'

16:57 chouser: it's not nearly that magical

16:57 Chousuke: lghtng: software transactional memory

16:58 lghtng: ah, that

16:58 hiredman: lghtng: you don't have to use do

16:58 Chousuke: lghtng: it's just a bunch of mutable reference types with well-defined concurrency semantics.

16:58 hiredman: clojure seemlessly calls java

16:58 java is full of side effects

17:00 lghtng: a side effect is basically just something a function does that isnt part of its return value but affects something outside it's scope?

17:00 Chousuke: I guess so.

17:00 That's probably not a rigorous definition, but anyway :P

17:01 lghtng: it turns the knob 1 quarter turn to the right, igniting the flame on the burner

17:02 it places the teapot onto the burner

17:02 hiredman: http://en.wikipedia.org/wiki/Side_effect_%28computer_science%29

17:02 Chousuke: lghtng: rather, it activates the machinery that does those thins :)

17:02 things*

17:02 chouser: the future is a function of the past

17:02 hiredman: lghtng: if your world is map, then a function can return an updated map with the new position of the teapot

17:03 chouser: history is its domain

17:03 Chousuke: hiredman: but you also need a side-effecting procedure to activate the machinery.

17:03 chouser: right, it depends on if you're simulating the world or actually interacting with it.

17:03 Chousuke: hiredman: but as long as you isolate the side-effecting procedure from the "pure" world-map model, you're golden.

17:05 hiredman: Chousuke: he never mentioned machinery, so to assume he isn't just doing world modeling is a mistake

17:05 Chousuke: I'm pretty sure the pure model always exists, too, since software of course needs some values to work with. :)

17:07 hiredman: anyway, clojure is targeted squarly at java, and as such can do anything java can do (except annotations, at this point)

17:11 Chousuke: I guess in automation terms, the "pure model" would be your transfer functions, and the things that integrate them with the real world are the DA or AD transformers.

17:12 and the clojure stm models the fact that inputs are functions of t.

17:13 (hey, at least it's not a car analogy)

17:13 lghtng: which fits nicely with the concept that the program is the data

17:15 that clojure goes ahead and takes the next step of collapsing alot of byzantine verbiage is really helpful in learninr

17:15 it should be the exception that proves the rule, not the rules that prove the exception

17:17 that the author has to explicetly take alot of time in the videos to explain, well, this is a that and that is a this and they are pretty much the same thing and frankly irrelevant shows that there are so Original Ideas(tm) operating here, which is really cool and innovative-like

17:20 Chousuke: If you dig deep enough, I'm sure few of the ideas in Clojure are truly original

17:20 but it's the combination of them that makes it interesting :)

17:22 lghtng: sometime we all need to learn to ignore that missile-radar warning in our ears and go ahead and accept a new paradigm with grace

17:23 we are not all equally blessed in that regard, so it's fortunate for me that clojure appears to embrace that spirit

17:23 maybe i can finally learn how to program :D

17:24 Chousuke: You don't have to forget the old paradigms either. But chances are that once you learn new things, you realise that you *should* forget some things you learned previously

17:25 Like when learning about complex numbers... Addition and substraction still work pretty intuitively, but what, you suddenly can calculate the square root of a negative number? shocking!

17:26 lghtng: what previous knowledge must be forgotten there?

17:27 perhaps you mean the example of 1C+1C=1C where C=light-speed?

17:29 Chousuke: hm, I haven't seen that example.

17:32 but you might have learnt that it's not possible to take the square root of a negative number, and complex numbers give you a way to meaningfully do that, opening up new possibilities. Then, when you examine the properties of complex numbers, many assumptions you previously had about numbers will likely need to be revised :)

17:32 lghtng: granted.

17:47 hiredman: clojurebot: big o?

17:47 clojurebot: Huh?

17:47 hiredman: clojurebot: big o is <reply>http://twitter.com/cgoldberg/status/6340157473

17:47 clojurebot: Ik begrijp

17:58 qed: hiredman: haha

18:00 ohpauleez: haha, that's awesome

18:02 arohner: does defmacro support let-destructuring its arguments?

18:02 hiredman: yes

18:03 arohner: thanks

18:04 maddis: i am using this (http://wiki.github.com/technomancy/leiningen/emacs-integration) method from technomancy to connect to a remote repl. is there any way to open only a local port on the server side and tunnel everything through ssh? I am not very comfortable with opening my repl to everybody on the internet.

18:04 arohner: ssh -L localport:localhost:remote-repl-port user@remote

18:05 ohpauleez: that way or socat

18:06 arohner: that will connect to the remote box, and tunnel traffic from localhost:remote-repl port to localport on your box

18:06 maddis: ok, that's a ssh tunnel. but isn't the port on the server side still open to everyone?

18:06 arohner: no

18:06 maddis: with open I mean, everybody could connect?

18:06 arohner: it tunnels traffic from localhost:remote-repl-port to localport on your box

18:07 anyone on your box can connect to localport

18:07 if your firewall is set up properly, the repl port is closed to outsiders

18:08 then you ssh in via the SSH port, and tunnel is created on your side. the only remotely accessible open ports are 1) ssh on the remote box, 2) localport on your box. I assume the firewall on your local box is set up properly :-)

18:08 maddis: you mean the firewall on the server side? that is not the case here...that's why I ask this question :)

18:09 hmm, ok

18:09 i guess I have to thing about this again :)

18:09 arohner: this will work, even if your box allows no open connections, and the remote box only allows SSH connections

18:09 esbena: installing clojure-mode (using ELPA) results in: 'clojure-slime-config: Cannot open load file: swank-clojure-autoload' - known problem?


18:11 hiredman: I think you have to pass an argument to ssh to let connections not originating from localhost traverse the tunnel

18:14 mabes: is the fn rest lazy? I have a lazy collection which I want to map all of the elements except the first one...

18:15 arohner: mabes: rest will evaluate the first item, and not evaluate the rest of the list

18:15 mabes: arohner: ok, great, thanks

18:18 polypus: how to get a function's arity?

18:19 the-kenny: polypus: (:arglists (meta (var +)))

18:19 ,(:arglists (meta (var +)))

18:19 clojurebot: ([] [x] [x y] [x y & more])

18:20 polypus: ty, is this also for user defined functions? i thought not

18:21 the-kenny: Works with user-defined funcions too.

18:21 technomancy: esbena: yep; it's been deprecated. see the readme for swank-clojure.

18:21 maddis: arohner, i think you got me wrong. i am talking about the port 4005 opened by the swank server on the server side. when I start the server with 'lein swank', one could connect to that port simply with slime-connect in emacs

18:21 the-kenny: But you have to know the symbol, as far as I know, there is no way to get the arity of functions itself.

18:22 arohner: maddis: yes, swank-server will open port 4005 on the server side. If you don't want that to be accessible to anyone, firewall it off. Block everything on the box except port 22

18:22 polypus: the-kenny: right its metadata on the var really

18:23 arohner: maddis: the ssh command I wrote above will open a port 4006 on your box, which all traffic on your side will go through ssh to the remote box, and then to localhost:4005, from the perspective of the remote box

18:23 maddis: ok, so there is no option to let it accept only local connections. thanks :) will block it through the firewall

18:23 arohner, yes, I got the ssh part.

18:24 my issue was only the open port 4005 on thr server side. that's all

18:24 polypus: ,(:arglists (meta (fn [x] x)))

18:24 clojurebot: nil

18:25 hiredman: the arglist metadata is not on the function, but on the var

18:25 the-kenny: polypus: Functions don't have metadata. Only the vars

18:25 That's what I was trying to say

18:25 Sorry, English isn't my native language.

18:26 arohner: maddis: I haven't heard of an option to make swank only listen on localhost. It might be possible

18:26 technomancy: maddis: it should be added in swank-clojure, but it doesn't exist yet afaict.

18:26 arohner: I'm sure the patch would be easy

18:28 maddis: i was just wondering because I read this page http://common-lisp.net/project/slime/doc/html/Setting-up-the-lisp-image.html#Setting-up-the-lisp-image

18:28 see footnote 2

18:28 hiredman: *snort*

18:29 polypus: is that a thread local snort?

18:30 the-kenny: ,*snort*

18:30 clojurebot: java.lang.Exception: Unable to resolve symbol: *snort* in this context

18:51 jeremyevans: Is there a Clojure function similar to Ruby/Perl/Python pack/unpack, such that (unpack "V" "\0\1\0\0") returns 256

18:52 Or any other facility to pack/unpack integers to/from strings?

18:52 djork: no but it sounds like fun to write

18:53 seems like a clojurey kind thing to have

18:53 http://www.perlmonks.org/?node_id=211105

18:53 jeremyevans: djork: Thanks. Shouldn't be difficult, just didn't want to recreate the wheel

18:54 hiredman: ,(Integer/parseInt "0100" 16)

18:54 clojurebot: 256

18:54 jeremyevans: hiredman: That's "0100", not "\0\1\0

18:54 0"

18:55 hiredman: I understand, but what is the poing of the slashes to begin with?

18:55 ,(Integer/toString 256 16)

18:55 clojurebot: "100"

18:55 jeremyevans: hiredman: "\0" is ascii 0x0, "0" is ascii 0x30

18:56 hiredman: "0100" and "\0\1\0\0" are two very different things

18:57 ,(= "\0" "0")

18:57 clojurebot: false

18:58 hiredman: ping?

18:58 clojurebot: ping?

18:58 nuts

18:58 clojurebot: PONG!

18:58 PONG!

18:59 hiredman: clojurebot: why?

18:59 clojurebot: http://clojure.org/rationale

19:14 fawxtin: anyone running swank-clojure, doing a remote REPL with lastest slime?

19:15 dnolen: fawxtin: I'm not sure that you can use latest slime, there were some breaking changes a while back.

19:16 Raynes: I really wish they would change the name.

19:17 You can't tell someone you're running Slime without extra explanation.

19:17 fawxtin: dnolen: oh, I see, it seens to work with old versions (using the simple swank.swank/start-server ...)

19:18 but now when I eval "list" expression (with parenthesis), it hangs up

19:19 dnolen: fawxtin: I use slime from the git mirror, I note that I'm on commit 74a4b66bf85f55e878df2974ab3d713f87c1a42f.

19:19 fawxtin: dnolen: nice!

19:19 let me see

19:20 dnolen: git mirror -> github mirror

19:20 fawxtin: yeah, sure

19:27 Raynes: dnolen: I've seen some of your SO posts. Nice ones. ,3

19:27 <3*

19:28 dnolen: Raynes: SO posts?

19:30 fawxtin: dnolen: worked really fine, thanks

19:30 dnolen: fawxtin: great!

19:31 Raynes: dnolen: Stackoverflow

19:31 dnolen: fawxtin: there was some effort to try to keep the slime devs from breaking Clojure support, hopefully something comes of it. It was really great for a year or so.

19:31 Raynes: oh yes, thx.

19:35 fawxtin: dnolen: I see, it must be hard as hell to keep sync on everything

19:35 for slime devs

19:43 or not.. dnolen, that commit seens to be one of the lastest (if not *the*) to work

19:44 dnolen: fawxtin: yes, I used to always stay at slime tip, I believe it's only a 2 or 3 commits before it broke at most.

19:48 tomoj: what's the breakage look like?

19:49 fawxtin: the eval hangs up

19:50 Im brute forcing to see what commit broke it

19:50 eval I mean repl

19:53 found the bad guy: 22dba9f590a55977f41cb9a333012e9cb6619505

19:54 _ato: WTF? https://gist.github.com/ffc64fc2a18382dbfaf6

19:54 using a non-direct buffer and looping over the underlying array with aget is also slower than the infinite loop with a underflow catch

19:54 NIO is weird

19:58 and it's not that (.remaining) is O(n), even if I call it above the loop it still doesn't slow down the second one

19:59 maybe the JIT does more magical inlining in the second case or something

20:00 alexyk: I want to parse twits like "hey @man there's a #clojure blog at http://bot.ly/xyz" into: @names, #hashtags, URLs, and remaining words. Questions: how do folks tag token types, and is there a library to simplify this?

20:00 dnolen: _ato: you going to publish your blog post soon, or are you still fine tuning wf2-optimized? :)

20:04 alexyk: I I mean I can boringly scan and add characters to a StringBuilder... should I then place into a map, {:token blah :kind :url } or something?

20:05 is this how algebraic types are simulated? :)

20:06 _ato: dnolen: still fine tuning my extremely optimised version. I had an idea that might make it (shock!) competitive with C++, but NIO is proving troublesome. It can be *really* fast in certain cases, but it seems to be really easy to slow it down, like that weird example I just posted

20:06 dysinger: rhickey: are around ?

20:07 dnolen: _ato: wow, head to head with C++. I'm impressed you got it close to Ocaml, that's a feat. Those are respectable numbers already.

20:47 arohner: is there a way to specify that a #() should take any number of arguments, but only use the first one?

20:48 I have a #(foo %1) that is unhappy about being called with more than one argument

20:49 timothypratley: (fn [x & ignore] ...) :)

20:50 arohner: timothypratley: yeah ;-), I was hoping for a #() form

21:25 harrison: ,(floats (into-array [1 2 3]))

21:26 aldebrn: I need to do FFTs in Clojure (really need to :D), are there any gold-standard Java FFT libraries? or wrappers to FFTW?

21:34 harrison: ,(floats (into-array [1 2 3]))

21:34 clojurebot seems to be slacking

21:39 tomoj: clojurebot: botsnack

21:46 harrison: is there any way to make a float array from clojure? (not a Float array. i'm using java.nio.FloatBuffer, and it wants a float array)

21:46 chouser: (floats (float-array [1 2 3]))

21:47 you don't need 'floats'

21:47 tomoj: huh, how long has floats been there?

21:47 and ints, never noticed them

21:48 chouser: if your version of Clojure doesn't have float-array, you can use (into-array Float/TYPE [1 2 3])

21:49 harrison: chouser: excellent. float-array was all it needed. i wasn't aware that existed. also, floats is redundant there.

21:50 liebke: aldebrn: I've never used it, but checkout jtransforms (which is included in the Incanter distribution): http://sites.google.com/site/piotrwendykier/software/jtransforms

21:57 chouser: tomoj: a while now. booleans, bytes, chars and shorts are newer.

22:02 hamza: guys, i don't know if this is one of those FAQ questions but why does

22:02 , (seq? [1 2 3])

22:02 equals false?

22:03 laughingboy: Newbie Q: Is it possible to change the current working dir used by load-file?

22:03 tomoj: hamza: vectors aren't seqs

22:03 harrison: these nio buffers are a pain in the ass.

22:03 tomoj: ,(seq? (seq [1 2 3]))

22:04 clojurebot would say "true" if she were not dead

22:04 hamza: tomoj: is there a pred for testing vectos?

22:04 he he

22:04 *vectors

22:04 or should i use (seq? (seq ..))

22:05 tomoj: well, there's vector?

22:05 but is that what you really want to know? that what you were given was a vector?

22:05 clojurebot: false

22:05 true

22:06 tomoj: oh hello there clojurebot

22:06 hamza: lol, i assumed it vas vec? i am iterating a vector using recursion i want to use this for stopping contition

22:06 tomoj: then, no, I don't think you're looking for vector?

22:06 laughingboy: Newbie Q: Is it possible to change the current working dir used by load-file?

22:07 tomoj: well, what will the value be when you want to stop?

22:08 hamza: i am going rest rest so i can use nil? then right?

22:08 tomoj: if you're using rest, you're not getting a vector anyway

22:08 ,(vector? (rest [1 2 3]))

22:08 clojurebot: false

22:08 tomoj: unless I misunderstand you

22:09 hamza: laughingboy: (System/setProperty "user.dir" "mydirectoryPath") works in java but i never used it in clojure.

22:09 ohh ok there is my mistake..

22:09 tomoj: if you're using rest, you're getting seqs

22:09 laughingboy: hamza: thx, I will try

22:10 tomoj: the idiomatic stopping condition is (if (seq ...) ...)

22:10 hamza: ok so seq? then should work so i got another logic problem.

22:10 tomoj: if you use next instead you will get nil when you're done

22:11 hamza: ok thanks..

22:11 tomoj: might be better to use seq anyway, depends I guess

22:42 interferon: does clojure have support for annotations?

22:42 arohner: interferon: no

22:47 interferon: are there ways to work around that?

22:49 arohner: not that I know of

22:54 KirinDave1: Is there a way to react to a ref or atom changing?

22:54 arohner: ,(doc add-watcher)

22:55 oh, no clojurebot

22:55 anyways, add-watch and add-watcher

23:04 drewr: any slime users have issues with namespaces not being recognized when there's metadata present?

Logging service provided by n01se.net