00:01 (join) mye_
00:07 (quit) cdidd: Ping timeout: 252 seconds
00:12 (quit) jeapostrophe: Ping timeout: 272 seconds
00:20 cky: eli: Re earlier convo: http://www.youtube.com/watch?v=9bZkp7q19f0
00:21 cky: eli: When I suggested to dyoo that he perform it in his presentation, he was very disturbed by the idea. :-P
00:31 mye: what is the correct terminology for "toplevel definitions", ie. the defintions in a module that are not nested. I always thought toplevel meant the repl
00:32 mye: is there even a distinction? I managed to confuse myself about this again
00:33 (join) jesyspa
00:34 (quit) mithos28: Quit: mithos28
00:34 mye: toplevel defines are shared between several requires of the module right? (i.e. the module is "cached")
00:35 (join) mithos28
00:39 mye: I guess it's time to reread http://docs.racket-lang.org/reference/syntax-model.html :-/
00:43 (quit) mye_: Quit: mye_
00:45 (quit) mithos28: Quit: mithos28
00:50 asumu: mye: I think people will often informally call those top-level definitions.
00:50 asumu: Even if they are technically module top-level.
00:50 asumu: Or module-level rather.
00:54 (join) mithos28
01:04 (quit) mithos28: Quit: mithos28
01:04 (join) mithos28
01:08 (quit) mceier: Quit: leaving
01:13 asumu: Yay, I fixed stchang and my snappy FFI binding. C is weird.
01:15 (quit) ssbr_: Ping timeout: 245 seconds
01:27 (quit) hash_table: Ping timeout: 246 seconds
01:41 (quit) jesyspa: Quit: leaving
01:45 (join) jesyspa
01:52 (join) hash_table
01:53 eli: cky: :)
02:00 (quit) hash_table: Ping timeout: 245 seconds
02:09 (join) em
02:14 (join) mceier
02:23 (join) vkz
02:30 (join) gciolli
02:33 (join) hkBst
02:38 (quit) hkBst: Ping timeout: 245 seconds
02:41 (join) hkBst
02:44 (quit) jackhammer2022: Quit: Textual IRC Client: http://www.textualapp.com/
02:56 (quit) mithos28: Quit: mithos28
03:00 (join) mithos28
03:00 (quit) mithos28: Client Quit
03:11 (quit) em: Ping timeout: 240 seconds
03:39 (quit) hkBst: Remote host closed the connection
03:40 (join) hkBst
04:02 (join) em
04:18 (quit) gciolli: Quit: Leaving.
04:23 (quit) hkBst: Read error: Connection reset by peer
04:23 (join) bitonic
04:24 (join) hkBst
04:29 (join) masm
04:55 (quit) hkBst: Read error: Connection reset by peer
04:55 (join) hkBst
04:58 (quit) vkz: Quit: vkz
05:02 (join) mye_
05:07 (join) cdidd
05:18 (join) soegaard_
05:35 (quit) mye_: Quit: mye_
05:37 (quit) cdidd: Remote host closed the connection
05:40 (quit) bitonic: Ping timeout: 260 seconds
05:48 (quit) hkBst: Ping timeout: 245 seconds
05:49 (join) hkBst
05:55 (join) noelw
06:00 (quit) noelw: Ping timeout: 245 seconds
06:30 (join) RacketCommitBot
06:30 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://git.io/G-nH6Q
06:30 RacketCommitBot: [racket/master] fix JIT bug in struct pred/get/set corner case - Matthew Flatt
06:30 RacketCommitBot: [racket/master] JIT-inline structure allocation - Matthew Flatt
06:30 (part) RacketCommitBot
06:32 mye: uhmm stupid question ahead: how do I connect an input port to an output port?
06:33 mye: I knew it, copy-port does that
06:33 mye: secrect sauce!
06:34 mye: *secret
06:38 mye: ahem, actually I want to connect an output port to an input port...
06:38 mye: there's no function for that?
06:46 (quit) hkBst: Ping timeout: 246 seconds
06:46 (join) hkBst
07:16 (join) bitonic
07:24 (join) vkz
07:38 (join) noelw
07:57 (quit) Demosthenes: Ping timeout: 252 seconds
07:59 (join) Demosthenes
08:03 (join) mizu_no_oto
08:05 (quit) Demosthenes: Ping timeout: 260 seconds
08:06 (join) Demosthenes
08:06 (join) vkz_
08:08 (join) jao
08:08 (quit) jao: Changing host
08:08 (join) jao
08:09 (quit) vkz: Ping timeout: 256 seconds
08:09 (nick) vkz_ -> vkz
08:09 (quit) bitonic: Ping timeout: 245 seconds
08:12 (join) carleastlund
08:21 (join) jeapostrophe
08:21 (quit) jeapostrophe: Changing host
08:21 (join) jeapostrophe
08:23 (join) noelw_
08:24 (quit) noelw: Ping timeout: 245 seconds
08:24 (nick) noelw_ -> noelw
08:25 (join) RacketCommitBot
08:25 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/0XKFqQ
08:25 RacketCommitBot: [racket/master] avoid compiler warning - Matthew Flatt
08:25 (part) RacketCommitBot
08:27 (quit) Demosthenes: Ping timeout: 260 seconds
08:29 (join) Demosthenes
08:31 (join) dnolen
08:37 (quit) Demosthenes: Ping timeout: 276 seconds
08:38 (join) Demosthenes
08:42 (join) RacketCommitBot
08:42 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/5OUgcQ
08:42 RacketCommitBot: [racket/master] show a dialog box saying that aspell/ispell cannot be - Robby Findler
08:42 (part) RacketCommitBot
08:59 (join) hash_table
09:00 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
09:14 (quit) hkBst: Remote host closed the connection
09:15 (join) xhyang`
09:15 (join) hkBst
09:16 (part) xhyang`: "ERC Version 5.3 (IRC client for Emacs)"
09:18 (quit) soegaard_: Quit: soegaard_
09:19 (join) uselpa
09:20 (join) Netfeed
09:20 (part) Netfeed: "WeeChat 0.3.0"
09:26 (join) ssbr_
09:27 (quit) uselpa: Remote host closed the connection
09:28 (join) bitonic
09:30 (join) noelw
09:30 (quit) hkBst: Ping timeout: 240 seconds
09:30 (join) soegaard
09:33 (join) hkBst
09:33 (quit) bitonic: Ping timeout: 246 seconds
09:35 (quit) jeapostrophe: Read error: Operation timed out
09:40 (join) momo-reina
09:43 (join) mizu_no_oto
09:43 (join) bitonic
09:44 (join) uselpa
09:48 (quit) hash_table: Ping timeout: 260 seconds
09:49 (join) neilv
09:49 (quit) bitonic: Read error: Operation timed out
09:51 (quit) hkBst: Ping timeout: 246 seconds
09:52 (join) hkBst
09:53 (quit) momo-reina: Remote host closed the connection
10:02 (nick) Shozan -> SHODAN
10:07 (quit) uselpa: Remote host closed the connection
10:33 (join) bitonic
10:37 (join) gciolli
10:45 (join) hash_table
10:49 (quit) dnolen: Ping timeout: 246 seconds
10:50 (join) jon7
10:57 (quit) mceier: Quit: leaving
11:03 (join) preyalone
11:03 preyalone: How do I define a main function in Racket?
11:03 preyalone: Surprisingly hard to Google.
11:03 jrslepak: you don't
11:04 asumu: preyalone: (define (main) ...)
11:04 jrslepak: oh, really?
11:04 asumu: Well, the --main flag to racket will call the function bound as main.
11:04 jrslepak: ah ok
11:04 asumu: You could also use a main submodule.
11:04 preyalone: I see a command line option in racket: -m | --main args automatically calls the main function with the relevant arguments.
11:04 jrslepak: what happens to top-level code if you do that?
11:04 preyalone: So I need to define (main) so that it accepts an argument vector.
11:05 jrslepak: you mean command line args?
11:06 mizu_no_oto: I remember people mentioning contracts at racketcon a few days ago.  Out of curiosity, could you check something like the monad or functor laws via contracts?  If not, is there a quick check like library that would let you do that?
11:08 jrslepak: you could check that things you actually do at runtime aren't examples of violations of those laws
11:08 preyalone: mizu_no_oto: I ported QuickCheck to Chicken Scheme (https://github.com/mcandre/cluckcheck). It shouldn't be that hard to port to Racket.
11:08 preyalone: jrslepak: Yes. I would like to, for instance, grab the equivalent of C's argv[0], the script name, if possible.
11:09 jrslepak: one of the "Start Quickly" examples on the racket-lang.org frontpage demonstrates how to access them
11:09 (quit) hkBst: Quit: Konversation terminated!
11:09 jrslepak: it's the dice example
11:10 jrslepak: `command-line' syntax lets you bind names to command line arguments
11:10 preyalone: jrslepak: exact url?
11:10 jrslepak: I'm ... not sure how to link directly to one of the examples
11:11 preyalone: ah, i see
11:11 preyalone: I tried (current-command-line-arguments), but it omits the .rkt filename.
11:11 jrslepak: oh hmm
11:12 preyalone: I'm trying to find Racket code that emulates Python's if __name__=="__main__": main()
11:12 jrslepak: what does that do in Python?
11:13 preyalone: In scripting languages like Python, Ruby, and Perl, you can bundle an API library and a CLI program using the library, in a single piece of source code.
11:14 preyalone: That Python snippet detects whether the script is being loaded as a singleton on the command line, e.g. "python myscript.py"
11:14 preyalone: The other case is that the script is being loaded as a module by another Python script: import myscript...
11:15 preyalone: With this little bit of Python magic, I can write an API and CLI in one file without having to write a separate "cli.py" or such file.
11:16 jrslepak: so I guess in Racket, the equivalent would be determining whether a module is being loaded as the actual program or by some other module?
11:16 preyalone: Yes.
11:17 jrslepak does not know the module system well enough for this
11:17 preyalone: I thought racket -m would do the trick, calling (main) a la Java, but -m really expects a separate main.rkt file.
11:17 preyalone: So I'll have to explore for special Racket tricks to do this.
11:19 preyalone: I've managed to do this in Chicken (https://github.com/mcandre/scriptname/blob/master/scriptname.scm).
11:21 stamourv: preyalone: Have you tried using a `main' submodule?
11:21 stamourv: e.g. `(module+ main ...)'
11:21 preyalone: stamourv: Could you provide a short pastebin for what that would look like?
11:22 preyalone: I'm not sure how to use submodules yet.
11:22 stamourv: Well, it's basically what I wrote above. You put the code you want to run where I put the `...'.
11:22 mye: (module+ main (display "hello, world"))
11:23 preyalone: okay..
11:23 stamourv: Inside the submodule, you have access to everything in the enclosing file.
11:23 preyalone: "module+: allowed only in a module body"
11:23 preyalone: How do I wrap this inside a module declaration?
11:24 stamourv: Use `#lang racket' at the top of the file.
11:24 stamourv: (Or the #lang language you want.)
11:25 (join) jeapostrophe
11:25 (quit) jeapostrophe: Changing host
11:25 (join) jeapostrophe
11:26 preyalone: stamourv: Hmm. Could you pastebin a Hello World example, including the command to run it?
11:26 preyalone: I've got a basic Hello World, but it only runs when I shebang and ./ it. If I call "racket -r hello.rkt" manually, nothing happens.
11:26 preyalone: How would you run Racket programs from the command line in Windows?
11:27 stamourv: I don't use Windows, but I guess that "racket.exe hello.rkt" would work.
11:27 stamourv: (or whatever the name of the racket executable is)
11:27 preyalone: https://gist.github.com/3893090
11:27 stamourv: You don't need "-r".
11:28 (quit) snorble_: *.net *.split
11:28 preyalone: stamourv: Thanks.
11:30 stamourv: "-r" uses `load', which is a bad idea, and doesn't work with modules.
11:30 preyalone: Okay, I'm getting an "unbound identifier" error in the code. https://gist.github.com/3893090
11:31 stamourv: `module+' defines submodules, use `define' for variables and functions.
11:31 preyalone: Not sure if idiomatic, but this works: https://gist.github.com/3893090
11:32 preyalone: How would I A) load scriptedmain and B) reference the meaning-of-life function in the module?
11:32 stamourv: You don't need the `meaning-of-life' submodule.
11:32 stamourv: The `main' submodule can access definitions from the enclosing module.
11:33 preyalone: updated: https://gist.github.com/3893090
11:33 stamourv: You just need to define `meaning-of-life', either at the module top-level, or inside the `main' submodule. Either would work.
11:33 (quit) jrslepak: Quit: What happened to Systems A through E?
11:33 stamourv: If you define it inside the `main' submodule, it won't be visible outside of it.
11:33 preyalone: How would I load the code from another Racket program and call the meaning-of-life function?
11:33 stamourv: Yep, that version looks good.
11:34 stamourv: You can use `provide' to export identifiers and `require' to import a module's exports.
11:34 preyalone: thanks! could you give a little more detail?
11:34 stamourv: http://doc.racket-lang.org/guide/modules.html
11:34 stamourv: rudybot: doc provide
11:34 rudybot: stamourv: your sandbox is ready
11:34 rudybot: stamourv: http://docs.racket-lang.org/reference/require.html#(form._((lib._racket%2Fprivate%2Fbase..rkt)._provide))
11:34 stamourv: rudybot: doc require
11:34 rudybot: stamourv: http://docs.racket-lang.org/reference/require.html#(form._((lib._racket%2Fprivate%2Fbase..rkt)._require))
11:35 stamourv: Each module contains definitions. If you want some of these definitions to be visible in other modules, you need to `provide' these definitions, and the other module need to `require' your module.
11:37 preyalone: stamourv: thanks for the doc link. not bad.
11:37 preyalone: i'm getting an "unknown module" error. https://gist.github.com/3893090
11:38 preyalone: ah, if i use (require "scriptedmain.rkt"), everything works!
11:38 stamourv: Yep.
11:38 stamourv: Symbol syntax is for "collects", or libraries.
11:39 stamourv: Like `racket/match', or `xml'.
11:41 (quit) noelw: Quit: noelw
11:42 preyalone: I'm a little bit confused by all the #lang options. I wish Racket offered a more standard environment, perhaps defaulting to #lang racket.
11:42 (join) asd_
11:42 asd_: ok
11:43 (nick) asd_ -> Guest38252
11:43 (quit) mizu_no_oto: Quit: ["Textual IRC Client: www.textualapp.com"]
11:45 (quit) preyalone: Quit: Page closed
11:45 (join) Aune
11:45 (join) aune_
11:46 (part) Aune
11:47 (quit) aune_: Client Quit
11:47 (join) Aune
11:50 (join) vkz
11:51 (quit) jeapostrophe: Ping timeout: 252 seconds
12:00 (join) jeapostrophe
12:00 (quit) jeapostrophe: Changing host
12:00 (join) jeapostrophe
12:03 asumu: BTW: "main" functions work fine with modules and the racket executable.
12:03 asumu: Contents of "foo.rkt": #lang racket (provide main) (define (main) (printf "hello~n"))
12:03 asumu: racket -tm foo.rkt
12:04 asumu: (note that you need the -t)
12:04 (join) RacketCommitBot
12:04 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/dA0cLQ
12:04 RacketCommitBot: [racket/master] Don't provide for/vector: helpers. - Vincent St-Amour
12:04 (part) RacketCommitBot
12:08 (join) mceier
12:08 (join) uselpa
12:18 (join) cdidd
12:21 (join) MayDaniel
12:22 (quit) uselpa: Remote host closed the connection
12:22 (join) mithos28
12:24 (join) anRch
12:30 (quit) ssbr_: Ping timeout: 245 seconds
12:33 (join) jrslepak
12:34 (join) h11r
12:41 (quit) vkz: Ping timeout: 272 seconds
12:41 (join) vkz
12:42 h11r: Hi. I've got a quick question: Is it a good or a bad idea to use custodians to automatically close system pipes when calling and piping shell commands together with subprocess bzw. process/ports? The idea is to make a new custodian, remember the old current one and replace it with the new one. Then do the shell  stuff and afterwards clean up by swapping back the custodian and calling custodian-shutdown-all on the new one?
12:43 mithos28: are you using parameterize on the custodians?
12:44 h11r: I was planning to use custodians.
12:44 mithos28: were going going to use parameterize to do the swaping?
12:44 mithos28: were you going*
12:45 h11r: you mean the 'current-custodian' parameter?
12:45 mithos28: It seems like a fine way to handle the exceptional cases, but I would still think you would want to do the closing manually for the cases where nothing goes wrong
12:46 mithos28: (parameterize ((current-custodian (new-custodian))) …)
12:46 mithos28: I'm not sure if new-custodian is the right syntax though
12:47 h11r: oh. No, i wasn't aware of this possibility. But that might be a good idea. In the short test I did i simply called (current-custodian <the-new-one>).
12:48 (join) vkz_
12:48 h11r: you have to call (make-custodian) but that isn't the point though :)
12:48 mithos28: If you are not using parameterize, I'm not sure what you are gaining from this
12:48 mithos28: I assume you want automatic cleanup
12:50 (quit) vkz: Ping timeout: 252 seconds
12:50 (nick) vkz_ -> vkz
12:51 h11r: yes, exactly. The problem is that, if I get the documentation right, the calls to subprocess create up to 3 system pipes per call that have to be closed manually. Therefore the idea to use custodians since they close file-type ports when they are collected.
12:51 mithos28: Which of those ports are you using?
12:51 h11r: But parameterize should do exactly what I was doing manually.
12:53 samth: h11r: this tutorial: http://docs.racket-lang.org/more/#%28part._.Terminating_.Connections%29 has a good intro to custodians
12:53 h11r: mithos28: I don't know yet :D
12:55 h11r: samth: I'll going to read it before asking further questions. Thanks.
12:58 (quit) anRch: Quit: anRch
13:02 (join) nilyaK
13:04 h11r: mithos28: The solution with parameterize seems not to work. http://pastebin.com/6u09csdc
13:05 (quit) bitonic: Ping timeout: 244 seconds
13:06 (quit) jon7: Ping timeout: 246 seconds
13:07 (join) nathanpc
13:08 (quit) mithos28: Quit: mithos28
13:13 (join) mithos28
13:14 h11r: Oh man, I didn't say nothing. It works if you actually call custodian-shutdown-all on the new custodian. Sorry! Maybe I should stop trying those things for today ...
13:14 mithos28: And custodian-shutdown-all should be called when the new custodian is garbage collected
13:16 mithos28: I believe, you should test that
13:17 h11r: mithos28: Yes, I recognized it. And thanks for your help.
13:18 h11r: samth: The tutorial seems to be just what I looked for. So thank you very much, too.
13:19 h11r: So I'll read and play a little with the custodian mechanic before bothering you any further.
13:19 h11r: Bye.
13:22 nathanpc: Does anyone have a simple TCP Socket example?
13:22 nathanpc: I was looking at the documentation but I could find one. http://docs.racket-lang.org/reference/tcp.html
13:23 mithos28: the third example on racket-lang.org
13:23 mithos28: its an echo server
13:24 nathanpc: hmm
13:24 (quit) h11r: Ping timeout: 276 seconds
13:24 (join) mizu_no_oto
13:24 nathanpc: oh
13:24 nathanpc: thanks very much mithos28
13:24 nathanpc: I'm starting with Racket (coming from Javascript)
13:25 nathanpc: and I wanted to have some fun with a Lisp dialect
13:25 asumu: nathanpc: good choice :)
13:25 nathanpc: so I choose Racket because I could use it on my Raspberry Pi (which is still on an aircraft coming to Brazil) :)
13:25 nathanpc: hehe
13:26 nathanpc: I was working with Node.js before
13:27 nathanpc: I was going to get into Lisp, but when I was searching I found that Racket is extremely more powerful than Lisp, so here I am :)
13:27 stamourv: nathanpc: Welcome! :)
13:27 nathanpc: and since I love to build sockets stuff to learn about a new language, I decided that a extremely simple chat server/client is a good way to start
13:27 nathanpc: thanks stamourv
13:27 stamourv: If you have any questions, let us know!
13:28 asumu: nathanpc: BTW, that reminds me. samth started writing detailed explanations of front page examples. Here is the echo server: https://github.com/samth/racket-examples/blob/master/echo.scrbl
13:28 nathanpc: ok :)
13:28 stamourv: And rudybot here is always happy to help.
13:28 stamourv: rudybot: You're happy to help, right?
13:28 rudybot: stamourv: dynamic binding is as dangerous (and occasionally useful, I admit) as unhygienic macros.  I got no problem with providing it as an option, but I'm not happy with it being the default.
13:28 nathanpc: oh nice
13:28 stamourv: rudybot: (+ 2 3 4)
13:28 rudybot: stamourv: ; Value: 9
13:29 stamourv: rudybot: doc +
13:29 rudybot: stamourv: http://docs.racket-lang.org/reference/generic-numbers.html#(def._((quote._~23~25kernel)._%2B))
13:29 nathanpc: wow nice
13:29 stamourv: rudybot: source
13:29 rudybot: stamourv: git clone git://github.com/offby1/rudybot.git
13:29 ianjneu: rudybot: sauce
13:29 rudybot: ianjneu: ido-mode is awesome sauce
13:29 nathanpc: what I'm loving about Racket is how simple and yet powerful it is, and how I can do a lot more with fewer lines
13:29 jrslepak_neu: rudybot: rudybot
13:29 rudybot: jrslepak_neu: apropos jordan
13:29 jrslepak_neu: k
13:33 (join) RacketCommitBot
13:33 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/QcJlKg
13:33 RacketCommitBot: [racket/master] adjust background compilation log messages so they only appear - Robby Findler
13:33 (part) RacketCommitBot
13:35 (quit) mithos28: Quit: mithos28
13:36 asumu: BTW: I've already mentioned this, but if you're building a PLaneT package, mcfly-tools is awesome.
13:36 neilv: whew. i was afraid i was the only one using it
13:37 asumu: neilv: https://github.com/stchang/snappy/blob/master/main.rkt
13:37 (join) ssbr_
13:37 neilv: cool
13:38 neilv: you preferred not to put the documentation in with the implementing code?
13:39 asumu: I haven't tried doing it that way yet, so I'm not sure.
13:40 neilv: i find it helps out as libraries grow in size. but the drawback is that it increases the bulk of text in your code files
13:42 neilv: recent example use of inline documentation: http://planet.racket-lang.org/package-source/neil/vlc.plt/1/4/vlc.rkt
13:43 neilv: if i used "@"-syntax, would be easier to distinguish "doc" forms from code, but i haven't figured out how to format "@"-syntax easily
13:47 (quit) nilyaK: Quit: Leaving.
13:53 (quit) jeapostrophe: Ping timeout: 252 seconds
13:56 (join) uselpa
13:59 (join) snorble_
13:59 (quit) uselpa: Remote host closed the connection
14:02 asumu: http://www.reddit.com/r/Racket/ <-- I changed the name of the subreddit a bit
14:03 (join) Blkt
14:04 neilv: asumu: if you can edit the sidebar of that subreddit, might want to take the "Dev" email list out. the distinction between the lists is confusing
14:05 (join) nilyaK
14:06 asumu: Fair enough. Probably most people who go on reddit was users.
14:07 asumu: And done.
14:10 (join) anRch
14:15 (quit) mye: Ping timeout: 265 seconds
14:25 (quit) acarrico: Ping timeout: 246 seconds
14:25 (join) RacketCommitBot
14:25 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/XTlxIw
14:25 RacketCommitBot: [racket/master] adjust logger to use a name instead of an environment variable - Robby Findler
14:25 (part) RacketCommitBot
14:30 (join) nilyaK1
14:32 (quit) nilyaK: Ping timeout: 244 seconds
14:32 (nick) nilyaK1 -> nilyaK
14:41 (join) acarrico
14:45 (join) t-rex
14:45 (quit) t-rex: Client Quit
14:46 (join) t-rex
14:47 (nick) t-rex -> Guest48778
15:02 (join) masm1
15:02 (join) jeapostrophe
15:02 (quit) jeapostrophe: Changing host
15:02 (join) jeapostrophe
15:03 (quit) masm: Ping timeout: 246 seconds
15:04 (join) RacketCommitBot
15:04 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/YJstFw
15:04 RacketCommitBot: [racket/master] Minor pickinesses. - Eli Barzilay
15:04 (part) RacketCommitBot
15:05 (quit) carleastlund: Quit: carleastlund
15:10 Cryovat: I was looking at asumu's slides
15:10 Cryovat: "Anything is only a *require* away" is a pretty good tagline
15:11 asumu: Cryovat: thanks
15:12 asumu: I considered several others, but in the end wanted something that implied everything is a library.
15:14 stamourv: Even languages! ;)
15:14 Cryovat: I can't wait until this makes it into TR
15:15 Cryovat: It feels like you're marrying the best things from Scheme with the best things from Haskell :)
15:17 stamourv: Why stop at Haskell? We're also taking cool stuff from Clojure, Scala, etc. ;)
15:17 Cryovat: It keeps blowing my mind how these advanced features can be implemented as libraries
15:17 (quit) anRch: Quit: anRch
15:18 stamourv: To me, that's really what makes Racket special: it can be anything you want.
15:19 (join) mithos28
15:20 Cryovat: I'm very glad you branched off from Scheme
15:21 Cryovat: R7RS is looking interesting, but it doesn't feel like it has the same momentum
15:24 (quit) mizu_no_oto: Quit: ["Textual IRC Client: www.textualapp.com"]
15:29 (quit) mithos28: Quit: mithos28
15:33 (join) carleastlund
15:34 (join) vkz
15:44 (quit) ssbr_: Ping timeout: 252 seconds
15:44 (join) gridaphobe
15:47 (quit) masm1: Ping timeout: 260 seconds
15:53 nathanpc: What's the best way to handle a thread globally on the code? Because I want to be able to kill the thread if a response from the client isn't the one the server is waiting for.
15:55 stamourv: nathanpc: Have you looked at custodians?
15:55 stamourv: rudybot: doc make-custodian
15:55 rudybot: stamourv: http://docs.racket-lang.org/reference/custodians.html#(def._((quote._~23~25kernel)._make-custodian))
15:55 nathanpc: hmm no
16:02 (join) masm
16:03 Guest48778: besides the racket guide, are there any racket books?
16:04 ianjneu: Realm of Racket is soon to hit the bookshelves
16:05 asumu: You might consider HtDP to be a Racket book, though it's more of a "how to program"
16:05 ianjneu: and it's not about racket idioms, but rather more beginner level stuff that's in most any functional language.
16:06 Cryovat: What kind of book is RoR?
16:06 Cryovat: Is it like Land of Lisp?
16:06 asumu: That's my understanding. I don't work on it though.
16:06 (join) ltbarcly
16:08 Guest48778: The docs and the guide are excellent, but i wish there were more examples
16:08 Guest48778: exercises even
16:08 nathanpc: me too
16:08 nathanpc: :)
16:08 nathanpc: I'm having a hard time with the docs
16:08 asumu: Guest48778: at the Hackathon we worked on adding examples to several sections. I agree we need even more though.
16:09 nathanpc: let me try to understand something here...
16:09 nathanpc: I'm opening a TCP listener using (tcp-listen)
16:10 nathanpc: when I do that, that call automatically makes a new custodian correct?
16:10 Cryovat: Guest48778: Have you seen a book called Real World Haskell?
16:10 Cryovat: I think it's probably the best functional programming book I've seen
16:10 Guest48778: no
16:11 Cryovat: I wish there was a Racket equivalent
16:11 Cryovat: What makes is great is that it assumes the reader is familiar with an imperative language already
16:11 asumu: nathanpc: I don't think it makes a new custodian. I think it associates the socket with the current custodian.
16:11 Cryovat: And leads you to reimplementing a bunch of the base library to get a feel for things
16:12 nathanpc: hmm
16:12 Cryovat: Before jumping on to "real" programming tasks
16:12 asumu: nathanpc: i.e, if you tell the current custodian to shut everything down, it will shut down that port too
16:12 nathanpc: asumu: So if I call (make-custodian (tcp-listen...)) it will make a new one?
16:13 asumu: nathanpc: I think you want something like (parameterize ([current-custodian (make-custodian)]) (tcp-listen ...) ...)
16:13 Guest48778: Cryovat: now i feel compelled to start working through the Haskell book
16:13 nathanpc: hmm
16:14 asumu: nathanpc: within the parameterize, it will set the current custodian to a new one. The tcp-listen associates to that. Then any shutdowns within the parameterize will shut down the TCP (and any other ports) but not stuff outside.
16:14 Cryovat: Guest48778: Don't stray away from Racket D:
16:14 nathanpc: ahh nice
16:14 nathanpc: this custodian stuff is finally making sense to me
16:14 nathanpc: and as far as I can see they are incredibly powerful
16:15 asumu: nathanpc: of course, you don't have to use parameterize. You can just set the current-custodian: (define new-custodian (make-custodian)) (current-custodian new-custodian)
16:15 nathanpc: ahh, that syntax is better
16:15 asumu: They are quite powerful. DrRacket and other Racket systems apps use them extensively to manage resources.
16:16 Cryovat: I was wondering
16:17 Cryovat: Would it be hard to make a "try-Racket-in-the-browser" page?
16:17 stamourv: nathanpc: Careful, though. `parameterize' reset the current custodian to its old value when you get out of parameterize, whereas setting the parameter directly doesn't.
16:18 nathanpc: hmm
16:18 nathanpc: ok
16:18 stamourv: Cryovat: There's wescheme.org
16:18 neilv: Cryovat: is pretty easy if the racket evaluation is done server side, but you really should just install drracket and use that
16:18 stamourv: (doesn't support full Racket, yet!)
16:18 (quit) hash_table: Ping timeout: 240 seconds
16:19 Cryovat: I was  thinking of something like http://tryruby.org
16:19 stamourv: Hacking rudybot to get commands from a web page shouldn't be too hard, I guess.
16:20 neilv: you won't get the full racket experience if you are just messing around in a browser
16:20 asumu: Cryovat: I bet it could be easily done with the sandbox, Racket webserver, and https://github.com/chrisdone/jquery-console
16:21 asumu: It would be a nice thing to have, for people who like trying things on the web.
16:21 Cryovat: Yeah
16:21 Cryovat: It's a good way to sell things
16:21 neilv: if people can't be bothered to download an app, maybe they should come back when they're a bit more ready :)
16:22 neilv: if racket is sold too hard, then it becomes like java, where you can't google a question without finding page after page of people who have no idea what they're talking about
16:22 carleastlund: whereas, neilv, if racket isn't sold hard enough, everybody's busy learning about java instead
16:23 Cryovat: Yeah
16:23 Cryovat: And a browser based page like that is a nice way to give people a feel of the language
16:23 Cryovat: You could play a bit in the lunch break at work, for instance
16:24 Cryovat: See the link on Hacker News, spend twenty minutes tinkering with it and making a mental note of giving it a more thorough look when you get home
16:33 neilv: people browsing reddit, hacker news, etc., during work hours is a big thing, you have a point.  you think a significant number of people would use and get a good experience from a web-based repl, when they otherwise wouldn't install drracket?
16:34 (quit) Blkt: Ping timeout: 255 seconds
16:38 samth: neilv: a web-based repl is *much* lower-investment
16:40 Cryovat: A web-based repl with directions is a much lower barrier of entry than downloading 50 megs, running an installer and opening a tutorial article
16:40 Cryovat: As for how many converts it would bring? No idea
16:41 neilv: samth: i think it depends on the person.   10+ years ago, lots of us got very used to typing "apt-get install APPNAME" whenever we heard of an app.  nowadays, people are very accustomed to installing apps for phones, and internet pipes at the office should make for fast downloads
16:42 (join) mizu_no_oto
16:42 samth: neilv: (a) windows/mac don't have apt
16:42 samth: (b) i find rudybot useful in this channel, and i certainly already have racket installed
16:43 neilv: no doubt, there are many people who will spend some time with a web-based thing before they'll install an app. how much yield you get for the amount of effort to give a good experience on the web is the question
16:43 neilv: if someone wants to do it, i should've shut up a few messages ago :)
16:43 samth: this is the model we should go for, I think: http://cljbin.com/
16:44 (quit) Araq: Ping timeout: 246 seconds
16:45 (join) Araq_bnc
16:45 (nick) Araq_bnc -> Araq
16:45 samth: in fact, that site is by #racket's own gf3 :)
16:45 gf3: Oh hai!
16:45 gf3: BTW http://twatlr.com is Racket
16:47 gf3: Has anyone built a nice URL resolver?
16:47 Cryovat: cljbin looks really nice
16:48 gf3: Cryovat: Thanks
16:48 samth: and it's only about 1k of clojure code
16:49 samth: much of which should be stealable from rudybot
16:49 (join) bitonic
16:50 gf3: Here's an example with some output and a fork → http://cljbin.com/paste/4f4de5ebe4b06eb04cc275f7
16:50 gf3: samth: I'd be happy to help
16:50 Cryovat: The background is really efficient
16:50 neilv: gf3: i have an old library that resolves urls. there is a heavily-changed unreleased version that also does reverse-resolution.  "http://www.neilvandyke.org/racket-uri/#%28part._.Resolving_.Relative_.U.R.I%29"
16:50 rudybot: http://tinyurl.com/957cb8y
16:50 Cryovat: It makes it feel really calm and pleasant
16:51 samth: gf3: thanks
16:51 samth: i know that cky expressed interest in building something like that as well
16:54 gf3: neilv: Interesting, thank you
16:55 neilv: gf3: i wrote it in 2004 in portable scheme. i will probably rewrite it from scratch, using modern racket-isms. but in the interim, it might get you by
17:04 samth: well, i created https://github.com/samth/rktbin as a fork of gf3's code
17:04 samth: we'll see if i do something with it
17:06 (join) bitonic`
17:06 Cryovat: I volunteer to test/play with the result
17:07 gf3: samth: Are you planning to rewrite it all in Racket?
17:07 (quit) mizu_no_oto: Ping timeout: 246 seconds
17:07 samth: yeah, probably
17:08 chandler: I volunteer to kibitz and to complain when it's not working. :-)
17:08 (join) hash_table
17:08 samth: that way i won't have to learn how to run clojure, just read it :P
17:08 (quit) bitonic: Ping timeout: 260 seconds
17:08 gf3: samth: Haha
17:08 gf3: samth: Probably not worth a fork then; I doubt there's anything in there that's reusable
17:08 samth: gf3: all the client-side code, though
17:09 gf3: samth: It's basically only this → https://github.com/samth/rktbin/blob/master/resources/public/js/main.js
17:10 gf3: samth: All the CSS uses a clojure pre-processor → https://github.com/samth/rktbin/blob/master/src/cljbin/css/cljbin.clj
17:10 (nick) bitonic` -> bitonic
17:11 (quit) jrslepak: Ping timeout: 246 seconds
17:13 (quit) gciolli: Quit: Leaving.
17:13 (quit) MayDaniel: Read error: Connection reset by peer
17:14 (join) jrslepak
17:28 neilv: if anyone is on a computer that has multiple displays and is running x, could they email the output of the "xrandr" command to neil@neilvandyke.org ?
17:30 neilv: for the bootable rackout image, i'm working on heuristics for how to use multiple displays it detects. something better than http://i.imgur.com/RpHaL.jpg
17:31 chandler: neilv: sent
17:31 neilv: thanks
17:34 (quit) vkz: Quit: vkz
17:38 chandler: Let me know if you need any other output or info... that's a ThinkPad T420 with a monitor connected via DisplayPort to HDMI dongle
17:39 (join) none___
17:40 neilv: chandler: thanks. this is exactly the the worst-case scenario i had in mind :)
17:44 nathanpc: nice neilv
17:44 (join) didi
17:45 nathanpc: btw, how to know if a custodian was shutdown?
17:45 nathanpc: and how to remake the custodian if it was shutdown?
17:45 nathanpc: (sorry, but I'm trying to learn how to use custodians)
17:46 (join) RacketCommitBot
17:46 RacketCommitBot: [racket] plt pushed 8 new commits to release: http://git.io/twQa0w
17:46 RacketCommitBot: [racket/release] Only recommend dropping exactness in truly inexact contexts. - Vincent St-Amour
17:46 RacketCommitBot: [racket/release] raco setup: improve description of `--doc-pdf' - Matthew Flatt
17:46 RacketCommitBot: [racket/release] slideshow/pict repair - Matthew Flatt
17:46 (part) RacketCommitBot
17:47 (quit) jesyspa: Quit: leaving
17:48 didi: When calling a procedure that returns multiple values, for example (values 4 2), I have to obligatory make room for those two values or is there a way to only use the first one (4)?
17:52 Cryovat: It sounds a bit strange to do so
17:53 Cryovat: And personally I think it'd be better to put the second value somewhere
17:53 Cryovat: Just so that it's explicit to the reader that you're discarding it
17:54 (quit) carleastlund: Quit: carleastlund
17:55 samth: neilv: https://gist.github.com/06700e7998fa3ac0e81a
17:59 neilv: samth: thanks. this breaks my parser
18:01 neilv: didi: the syntax forms that receive multiple values strongly encourage you to acknowledge each value that is returned. this arguably reduces bugs. similar to how you can't normally call a procedure with fewer than the required number of arguments
18:02 didi: neilv: I see.
18:03 didi: Cryovat, neilv: Thanks.
18:05 Cryovat: This made me think a bit
18:06 Cryovat: It doesn't seem like let-values lets you use wildcards for unwanted values
18:06 Cryovat: But that kind of feature would probably be a bad idea
18:06 (quit) jeapostrophe: Ping timeout: 255 seconds
18:08 (quit) none___: Ping timeout: 245 seconds
18:08 neilv: using define-syntax, you could implement your own version of let-values that supported such a wildcard (lisp people might call it a "rest argument"), if you wanted to
18:09 neilv: one of the things i found i was doing when i was first learning scheme was to imagine all these general purpose syntactic shortcuts i would like, and implement them
18:09 neilv: later, i found i wasn't really using those shortcuts, since my style was different than i expected it to be, and the situations were then different
18:10 Cryovat: Sounds like good macro practice though
18:10 neilv: it's great practice, yes
18:12 Cryovat: I have a theory that the code style changes a bit once you get used to the functional mindset
18:12 Cryovat: As in, less recursion, more combinators and point-free code
18:12 asumu: stchang: http://planet.racket-lang.org/display.ss?package=snappy.plt&owner=plt
18:13 didi: That's a good point. Although I was originally thinking about something like (+ ((lambda () (values 4 2))) 3) => 7
18:14 stchang: cool thanks
18:16 neilv: didi: i once implemented an alternate application syntax that would've yielded 9 rather than 7 for that
18:17 neilv: http://www.neilvandyke.org/weblog/2009/04/#2009-04-06
18:21 jrslepak: anyone know why I get this error from online expansion while trying to edit the Racket docs? -- make-directory: cannot make directory path: /Users/jrslepak/github/racket/collects/scribblings/compiled/drracket system error: Permission denied; errno=13
18:23 didi: neilv: Cool.
18:23 gf3: Is there a Racket equivalent of Clojure's if-let ?
18:23 gf3: http://clojuredocs.org/clojure_core/clojure.core/if-let
18:27 Cryovat: That's... weird
18:28 (join) bitonic
18:28 didi: Hehe, I find the first example a little weird.
18:29 didi: Why returning a string when `0' seems appropriate when there is no even number?
18:29 Cryovat: It's a pretty bad example
18:30 gf3: didi, Cryovat: The example is a bit contrived
18:30 (quit) masm: Ping timeout: 272 seconds
18:30 (quit) gridaphobe: Read error: Operation timed out
18:30 (join) jesyspa
18:30 Cryovat: I guess by test, they mean that the expression produces a non-null/empty/false result?
18:31 (quit) jrslepak: Quit: What happened to Systems A through E?
18:32 didi: Cryovat: I'm guessing Clojere have generalized boolean?
18:32 didi: s/Clojere/Clojure
18:37 Cryovat: It has both a boolean type and null
18:37 Cryovat: Due to running on top of the JVM
18:46 gf3: How does one know what type of exception is thrown?
18:50 (join) masm
18:53 (quit) bitonic: Ping timeout: 252 seconds
18:56 (join) gridaphobe
19:01 didi: gf3: What do you mean by "type"? You use predicates. <http://docs.racket-lang.org/guide/exns.html>
19:02 gf3: didi: For example: exn:fail:contract:divide-by-zero?
19:02 neilv: gf3: if you are just seeing the error message, you probably have to add a "with-handlers" to catch "exn:fail?" and then write the exception object as debugging output, then go change the "exn:fail?" to the appropriate predicate
19:04 gf3: neilv: Ahh, thanks
19:11 gf3: neilv: Boom! exn:fail:network
19:11 (join) otterdam
19:13 (join) jeapostrophe
19:13 (quit) jeapostrophe: Changing host
19:13 (join) jeapostrophe
19:17 (quit) gridaphobe: Ping timeout: 256 seconds
19:19 (quit) jeapostrophe: Ping timeout: 265 seconds
19:21 (join) gridaphobe
19:25 (join) bitonic
19:25 (quit) nathanpc: Disconnected by services
19:25 (join) Nathan
19:26 (nick) Nathan -> Guest74962
19:30 (quit) neilv: Quit: Leaving
19:31 (quit) bitonic: Ping timeout: 246 seconds
19:43 (join) mizu_no_oto
19:58 (quit) ltbarcly: Ping timeout: 246 seconds
19:59 (join) Fare
20:03 (quit) gridaphobe: Ping timeout: 246 seconds
20:11 (join) dnolen
20:26 nathanpc: Why I can't use (define ..) inside an if/else?
20:27 offby1: nathanpc: I think you can, if it's the first thing in the "arm".
20:27 offby1: but the real answer to your question is: I dunno :)
20:27 nathanpc: no problem
20:27 offby1: rudybot: (if #t "it's true!" (define x 3) (format "x is ~a" x))
20:27 rudybot: *offby1: Should I supply font to set-default-font in -*-blabalabla-*-*-rabel-rabel-* format?
20:27 nathanpc: I tried that, but it didn't worked
20:27 offby1: *sigh*
20:27 offby1: rudybot: eval (if #t "it's true!" (define x 3) (format "x is ~a" x))
20:27 rudybot: *offby1: your sandbox is ready
20:27 rudybot: *offby1: error: #<undefined>:1:0: if: bad syntax (has 4 parts after keyword) in: (if #t "it's true!" (define x 3) (format "x is ~a" x))
20:27 offby1: that's why :)
20:28 offby1: 'cuz they don't let you :)
20:30 nathanpc: hmm
20:30 nathanpc: rudybot: eval (if (< 10 20) ((define t "test") (display t)) (display "false"))
20:30 rudybot: nathanpc: your sandbox is ready
20:30 rudybot: nathanpc: error: #<undefined>:1:15: define: not allowed in an expression context in: (define t "test")
20:30 nathanpc: see
20:30 nathanpc: that's the error I'm getting
20:30 offby1: sure
20:30 offby1: try this
20:30 offby1: rudybot: eval (if #t "it's true!" (begin (define x 3) (format "x is ~a" x)))
20:30 rudybot: *offby1: error: #<undefined>:1:27: define: not allowed in an expression context in: (define x 3)
20:31 offby1: oops
20:31 offby1 rubs chin
20:31 nathanpc: haha no problem :)
20:31 offby1: rudybot: eval (if #t "it's true!" (let ([x 3]) (format "x is ~a" x)))
20:31 rudybot: *offby1: ; Value: "it's true!"
20:31 nathanpc: oh
20:31 offby1: rudybot: eval (if #t (let ([z 9])(format "it's true! z really is ~a" z))  (let ([x 3]) (format "x is ~a" x)))
20:31 rudybot: *offby1: ; Value: "it's true! z really is 9"
20:32 offby1: so there's a workaround.  But it doesn't answer your question.
20:32 nathanpc: hmm
20:33 offby1: I would hope that the racket docs explain this somewhere; hopefully in the section that introduces "define".
20:33 offby1: Haven't read 'em in a while though.
20:33 offby1: rudybot: doc define
20:33 rudybot: *offby1: http://docs.racket-lang.org/reference/define.html#(form._((lib._racket%2Fprivate%2Fbase..rkt)._define))
20:33 nathanpc: hmm
20:34 nathanpc: I'm going to ask this on the mailing list to see if people know why it throws this error
20:34 offby1: yep
20:37 (join) gridaphobe
20:46 asumu: nathanpc: cond is better than if because it lets you use defins
20:46 asumu: *defines
20:46 nathanpc: ahh
20:46 offby1: ooh
20:46 nathanpc: thanks asumu :)
20:46 asumu: rudybot: (cond [(< 10 20) (define t "test") t] [else 3])
20:46 rudybot: asumu: your sandbox is ready
20:46 rudybot: asumu: ; Value: "test"
20:46 nathanpc: going to take a look at cond
20:46 asumu: for example
20:46 (join) RacketCommitBot
20:46 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/E7NsnA
20:46 RacketCommitBot: [racket/master] minor GCs: reduce constant factor on total memory use - Matthew Flatt
20:46 (part) RacketCommitBot
20:47 asumu: Also see `match`
20:47 offby1: rudybot: eval (cond (#t (define z 9) (format "it's true! z really is ~a" z)) (#t   (define x 3)))
20:47 rudybot: *offby1: error: #<undefined>:1:0: begin (possibly implicit): no expression after a sequence of internal definitions in: ((define x 3))
20:47 offby1: rudybot: eval (cond (#t (define z 9) (format "it's true! z really is ~a" z)) (#t   (define x 3)  (format "x is ~a" x)))
20:47 rudybot: *offby1: ; Value: "it's true! z really is 9"
20:47 offby1: asumu: w00t
20:56 nathanpc: How to change the indentation in DrRacket? 1 space isn't good, I would like to have it on 4 spaces.
21:06 (join) jeapostrophe
21:06 (quit) jeapostrophe: Changing host
21:06 (join) jeapostrophe
21:06 (quit) acarrico: Ping timeout: 260 seconds
21:19 (join) ltbarcly
21:20 (join) acarrico
21:22 ltbarcly: hi, I'm new to racket and scheme, and am thinking about porting over some CL code I have been working on for awhile
21:22 (join) jrslepak
21:23 ltbarcly: one question I'm having trouble figuring out from the docs, although there are hints:  does racket let threads run in parallel?
21:24 ltbarcly: I see that you can use future objects, but it's not clear that threads can run concurrently, other than the docs saying you should run SCHEME_USE_FUEL from time to time to allow other threads to be run, which makes me think there must be an interpreter lock in there somewhere
21:28 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
21:28 (quit) Aune: Quit: L?mnar
21:29 ltbarcly: nm, I found it in the docs
21:30 didi: ltbarcly: This was discussed a little while around here. <http://paste.kde.org/570758/>
21:30 (quit) gridaphobe: Ping timeout: 260 seconds
21:31 (quit) jesyspa: Quit: leaving
21:31 ltbarcly: I'm guessing that this is due to a limitation in the vm?  I was hoping to port a dsl to racket, but unfortunately the dsl would not map happily into futures
21:33 didi: ltbarcly: Unfortunately, I know nothing about Racket's underlying implementation. :^( But CL and Racket are indeed different languages.
21:33 didi: Although there is nothing on CL's language that talks about threads.
21:34 ltbarcly: I agree, CL only supports threads as an extension
21:34 didi: More like an implementation thing, I guess.
21:34 nathanpc: How to change the indentation in DrRacket? 1 space isn't good, I would like to have it on 4 spaces.
21:35 didi takes nathanpc on the side and talks about the wonders of Emacs
21:35 didi: ;^)
21:35 nathanpc: hshs
21:35 nathanpc: hah*
21:35 nathanpc: I'm a Chocolat, and Vim, user ;)
21:36 didi: nathanpc: That's not incurable.
21:36 nathanpc: I hate how you need to use all those keyboard shortcuts instead of commands
21:37 ltbarcly: emacs is a terrible editor, but unfortunately it is also the best editor
21:37 didi: Hehe. I feel you. Having >20 fingers helps.
21:37 didi: ltbarcly: Nah. Emacs is awesome all the way.
21:37 ltbarcly: have you done much elisp?
21:38 didi: ltbarcly: A certain amount, yes. But Emacs is not elisp.
21:38 ltbarcly: ahh, don't say that in #emacs :)
21:38 didi: ltbarcly: They would agree.
21:38 nathanpc: haha
21:38 nathanpc: you'll start an epic flame war there
21:38 ltbarcly: if you don't write elisp, emacs is just an editor with emacs key bindings, right?
21:39 didi: ltbarcly: Nope.
21:39 didi: Emacs is org-mode, gnus, geiser...
21:39 nathanpc: didi: I can take another look at Emacs if you help me to setup it for Racket. :)
21:39 SamB_MacG5: you can also use others' elisp
21:39 didi: nathanpc: geiser dude.
21:39 ltbarcly: I see what you mean, but of course all of those are written in elisp
21:39 SamB_MacG5: well, sometimes ;-)
21:39 didi: nathanpc: And paredit.
21:39 nathanpc: I have Emacs installed here on my Mac and on my Linux box
21:40 ltbarcly: so, am i right to say that futures become serialized as soon as they have to deal with side effects?
21:40 ltbarcly: IO for example
21:41 ltbarcly: the docs are very shady and just say they become serialized if they try to do something 'too complicated' whatever that means
21:42 asumu: ltbarcly: you can run paralell computations in Racket via places or futures.
21:42 asumu: My guess is that you want places for what you're doing.
21:43 asumu: *parallel
21:43 asumu: see http://docs.racket-lang.org/guide/performance.html#(part._effective-futures) and the next section
21:44 ltbarcly: places seems isomorphic to forking + a nice IPC library
21:45 ltbarcly: which is handy I guess, but very different than proper system threads
21:45 ltbarcly: it's particularly hard to write a nice, extensible vm that also handles threads nicely
21:46 nathanpc: oh, racket.vim works great!
21:46 nathanpc: just need some tweaks to highlight stuff like (display)
21:46 nathanpc: but the rest does pretty well
21:47 ltbarcly: I think of all the interpreted languages, only ruby does an ok job at system threads, but even then it is iffy, python has a global lock and perl doesn't really have threads at all, and it seems racket is somewhere in between
21:47 didi: ltbarcly: May I point out that threads are not for parallel computation?
21:48 ltbarcly: well, maybe in racket
21:49 didi: ltbarcly: Maybe historically.
21:49 ltbarcly: but I strongly suspect that is more an apology for the vm not being able to do it than a real assertion
21:49 ltbarcly: at one point threads didn't mean parallel because few systems had more than one cpu
21:50 didi: ltbarcly: In Python, for example, if you want to do parallel computation you spawn new processes.
21:50 ltbarcly: that's purely because of a limitation in the vm
21:50 didi: ltbarcly: That's a little debatable.
21:50 didi: ltbarcly: Not the limitation, but the reason.
21:51 ltbarcly: the reason they have an interpreter lock is that it makes reference counting GC much faster (finer grained locking has more overhead and causes huge cache-line issues due to one cpu wanting to update memory that another has cached to do reference count updates), and because it makes writing C extensions much easier
21:52 didi: ltbarcly: <https://www.youtube.com/watch?v=EBRMq2Ioxsc&feature=player_embedded#t=1953s>
21:52 rudybot: http://tinyurl.com/96tfeby
21:52 ltbarcly: but other implementations of python, Jython for example, don't have this restriction and run threads normally
21:52 jrslepak: ltbarcly: I think didi is trying to tell you that the "threads are for concurrency, not parallelism" predates that VM by quite a bit
21:53 ltbarcly: guido doesn't believe in threads for whatever reason
21:53 didi: (add1 jrslepak)
21:53 ltbarcly: but again I think that is more being defensive about the VM limitations than a real argument
21:53 ltbarcly: anyway, I'm not trying to start a flame war
21:54 didi hugs ltbarcly 
21:54 ltbarcly: but it seems like "threads are for concurrency" is really not something that should be enforced
21:54 nathanpc: good night
21:55 ltbarcly: especially since the rest of the world has long since decided that threads are in fact for parallelism
21:55 didi: nathanpc: Sleep tight.
21:55 nathanpc: I'll dream with Vim :P
21:55 didi: Hehe. Whatever floats your boat.
21:55 jrslepak wonders how dreaming with vim works
22:00 (part) ltbarcly: "PING 1350352804"
22:07 (quit) didi: Remote host closed the connection
22:13 (join) didi
22:19 (join) jackhammer2022
22:52 (join) mizu_no_oto
22:57 (quit) mizu_no_oto: Client Quit
22:57 (quit) masm: Quit: Leaving.
23:07 (quit) Fare: Ping timeout: 246 seconds
23:08 (join) mithos28
23:12 (part) mithos28
23:12 (join) mithos28
23:12 (part) mithos28
23:12 (join) mithos28
23:20 SamB_MacG5: Hmm, I want to do a sort of survey about &lang; and &rang; ...
23:25 SamB_MacG5: If anyone can find a browser that fails to render a glyph in http://jsbin.com/okizeb/2/edit please tell me -- though most likely they'll fail in pairs ...
23:26 (quit) Shviller: Ping timeout: 245 seconds
23:27 (join) Shviller
23:30 (quit) jao: Ping timeout: 245 seconds
23:33 SamB_MacG5: (Well, really, I'm more interested in failure to render http://jsbin.com/okizeb/2 -- but I thought it might be helpful to see the other link too...)
23:34 SamB_MacG5 notes that CPython still has concurrency-only threads, for the most part
23:35 SamB_MacG5: (Of course, this has long been seen as a bad thing ;-)
23:35 mithos28: SamB_MacG5: Thats not true, its that there is a global lock on the python interpreter
23:35 SamB_MacG5: mithos28: well, I mean, if *Python* code is running, none of the other threads can run
23:36 mithos28: Yes they can, they just can't run python code
23:39 SamB_MacG5: yeah, realized I misspoke and was about to clarify when my client more-or-less froze up
23:40 SamB_MacG5: anyway, my first thought on reading about Racket threads was "oh, racket has a GIL?"
23:41 mithos28: Its different because racket threads are not native threads
23:41 SamB_MacG5: yes, that is a bit different
23:42 SamB_MacG5 thinks of fibres, but realizes that that's probably not a good analogy because fibres sound quite dangerous ...
23:43 SamB_MacG5: you know, they wouldn't have to be native threads just to support parallelism
23:44 SamB_MacG5: I mean, sure you'd need *some* native threads
23:44 mithos28: right that would be n*m multiplexing
23:44 asumu: They're so-called green threads.
23:44 mithos28: much more complicated
23:44 mithos28: haskell does that
23:44 SamB_MacG5: ITYM GHC
23:44 SamB_MacG5: and yeah, that's what I was thinking of
23:45 mithos28: Sorry the haskell platform does that
23:45 asumu: Our field is pretty bad about not using confusing names that mean 10 different things.
23:45 SamB_MacG5: it's not flawless -- tight allocation-free loops can still avoid being preempted ...
23:46 SamB_MacG5: I haven't really used Haskell much since the platform thing happened. (Not that I have anything against it, particularly...)
23:47 SamB_MacG5: (Though I do tend to be low on diskspace, which would probably lead me not to use the platform...)
23:48 didi: SamB_MacG5: What is the size of the platform?
23:49 SamB_MacG5: not quite sure, but I expect it has things in that I wouldn't actually use
23:50 didi: I see. It is a preference then.
23:51 SamB_MacG5: well, I'd prefer to have a larger disk in the Debian box I tend to use for that sort of thing
23:52 (quit) jeapostrophe: Ping timeout: 265 seconds