00:02 (quit) yoklov: Quit: bye!
00:02 (join) jonrafkind
00:11 (quit) rudybot: Remote host closed the connection
00:13 (quit) francisl: Quit: francisl
00:13 (join) rudybot
00:38 (join) jeapostrophe
00:45 (quit) jeapostrophe: Ping timeout: 248 seconds
00:54 (quit) jonrafkind: Ping timeout: 272 seconds
00:56 (quit) Shvillr: Ping timeout: 252 seconds
00:56 (join) Shvillr
00:59 (join) vkz
01:06 (join) jonrafkind
01:35 (quit) vkz: Quit: vkz
01:45 (join) vkz
01:47 (join) realitygrill_
01:48 (quit) realitygrill_: Client Quit
01:49 (quit) realitygrill: Ping timeout: 265 seconds
01:55 (quit) djcb`: Remote host closed the connection
02:47 (quit) jonrafkind: Ping timeout: 272 seconds
02:52 (quit) mithos28: Quit: mithos28
04:21 (quit) Shviller: Quit: bye
04:21 (join) Shviller
04:22 (join) bas_
04:22 (nick) bas_ -> Skola
04:23 (quit) vkz: Quit: vkz
04:48 (nick) noelw_away -> noelw
05:21 (join) tim-brown
05:21 tim-brown: mornin all
05:21 noelw: morning
05:29 (join) ahinki
05:49 (join) realitygrill
06:08 (join) bitonic
06:22 (join) kr
06:23 kr: hello. I try to run drracket on awesome wm, and it not work. it write this on terminal: fl-: expects type <flonum> as 2nd argument, given: 0; other arguments were: 14.0 about to suspend in atomic mode
06:24 (quit) realitygrill: Quit: realitygrill
06:25 kr: i can see logo and loading bar. but when it loaded it crush
06:26 noelw: Is it a recent DrRacket?
06:26 noelw: If not, you might try upgrading to the latest release
06:26 (join) vkz
06:29 kr: it is rocket 5.2.1-1 latest version in my repository
06:39 (join) veer
06:45 (join) masm
07:12 mrcarrot: how can i get racket to compile against the installed libffi instead of the bundled?
07:13 bremner: hmm. somehow debian does it.
07:18 bremner: mrcarrot: afaict, if the include files are found by the rack configuration process, it uses the system libffi
07:34 (join) mceier
07:39 (join) francisl
07:40 (quit) bitonic: Quit: WeeChat 0.3.5
07:45 mrcarrot: really strange, as i can not get it to pick up libffi
07:45 bremner: what OS?
07:45 mrcarrot: openbsd
07:45 bremner: hmm. Maybe it looks in the wrong place
07:48 bremner: mrcarrot: are your libs maybe in /usr/local or something?
07:48 mrcarrot: yes
07:48 bremner: it's just a wild guess, but maybe configure doesn't look there.
07:48 (join) samth_
07:51 bremner: try setting some the environment variables listed by ./configure --help
07:57 (join) jeapostrophe
08:01 (quit) francisl: Quit: francisl
08:01 mrcarrot: env LDFLAGS="-L/usr/local/lib" CPPFLAGS="-I/usr/local/bin/include" LIBS="-lffi" ./configure
08:01 mrcarrot: that is failing
08:02 mrcarrot: i get "configure: error: C compiler cannot create executables"
08:02 mrcarrot: if i remove LIBS="-lffi" it is working, but with included libffi instead of the system one
08:02 bremner: I guess you should write to the dev list
08:03 mrcarrot: okay, i will try :)
08:09 (quit) kanak: Quit: Leaving.
08:10 (join) kanak
08:20 (join) bitonic
08:36 (quit) samth_: Ping timeout: 246 seconds
08:44 (quit) kr: Quit: Page closed
08:49 (join) jrslepak
08:59 (join) dnolen
09:10 (nick) samth_away -> samth
09:14 (quit) vkz: Ping timeout: 250 seconds
09:39 (join) vkz
09:52 (join) bmp
10:11 (quit) jeapostrophe: Ping timeout: 260 seconds
10:33 (quit) mceier: Quit: leaving
10:38 (join) mceier
10:39 (join) realitygrill
10:44 (quit) karswell: Remote host closed the connection
10:53 (quit) dnolen: Ping timeout: 260 seconds
10:55 (join) karswell
11:13 (quit) jrslepak: Quit: What happened to Systems A through E?
11:19 (join) gciolli
11:20 (quit) Skola: Quit: Lost terminal
11:29 (quit) gciolli: Ping timeout: 245 seconds
11:36 (quit) realitygrill: Quit: realitygrill
11:46 (join) GeneralMaximus
11:46 (quit) GeneralMaximus: Client Quit
11:48 (join) DyingWeb
11:49 DyingWeb: anyone else felt the slowness of page http://racket-lang.org today???
11:51 (quit) ahinki: Quit: ChatZilla 0.9.88.1 [Firefox 12.0/20120314195616]
11:52 DyingWeb: maybe slownes is caused by "super scalable" racket web-stack written by superb jay?
11:57 asumu: DyingWeb: Probably not. There were some hardware or config issues with the server hosted here, I think.
11:58 DyingWeb: asumu: how do you know for sure?
11:58 asumu: DyingWeb: because eli said so.
11:58 asumu: Also the computer is, uhh, here.
11:58 DyingWeb: oh, when eli said so...
12:00 DyingWeb: 7GB of continuations not reclaimed
12:00 DyingWeb: just kidding
12:00 DyingWeb: can you imagine ebay running on racket?
12:01 DyingWeb: or amazon?
12:08 bremner: DyingWeb: are you racketnoob aka shen-troll?
12:08 DyingWeb: no
12:08 bitonic: bremner: yes
12:09 DyingWeb: i'm person who is honestly interested in answers
12:09 bremner: next time I'll just ask rudybot.
12:14 mrcarrot: DyingWeb: i noticed it
12:19 (quit) noelw: Ping timeout: 250 seconds
12:36 (quit) DyingWeb: Ping timeout: 245 seconds
12:39 (join) jonrafkind
12:41 (join) mithos28
12:49 (quit) bitonic: Ping timeout: 260 seconds
12:59 (quit) vkz: Ping timeout: 246 seconds
12:59 (quit) dous: Remote host closed the connection
13:00 (join) dous
13:01 (join) tommc
13:05 (join) dzhus
13:05 (quit) dous: Ping timeout: 244 seconds
13:09 (join) jao
13:09 (join) dous
13:10 (quit) veer: Quit: Leaving
13:15 (join) DyingWeb
13:16 (join) anRch
13:17 DyingWeb: bremner, others, don't avoid me, i'm not a troll!
13:17 bremner: so you are highlighting to demonstrate your non-trollishness? interesting strategy.
13:18 DyingWeb: i only love shen and hate the racket web documentation
13:18 DyingWeb: and i don't love some people behind racket, but that's all! Can't love everyone, no?
13:19 bremner: please don't keep changing nick. It makes it harder to ignore you.
13:19 (quit) stchang: Read error: Connection reset by peer
13:20 DyingWeb: why you'll want to ignore me at all? You can simply ignore text here, i don't come on private conversation to you
13:21 DyingWeb: here is "sound of silence" most of the time, anyway, so i don't think i'm doing any harm to anybody
13:24 (join) realitygrill
13:24 (join) stchang
13:31 (join) Cryovat
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:31 DyingWeb: HAHAHAHA, racket suckers!!!!!!!!!!!
13:32 Cryovat: ...I seem to have come in at a bad time
13:32 DyingWeb: one men, Mark Tarver beats you all!!!!!!!!!!!
13:32 DyingWeb: one genius!!!!!!!!!!!!!!
13:32 DyingWeb: beats all team of ridged PLT academic geeks!!!!!!!!!!!!!! HAHAHAHHA
13:33 tommc: DyingWeb ... I honestly cannot name a language with better documentation than racket. What is so confusing to you?
13:33 DyingWeb: look at reference for web server in racket
13:34 tommc: Yes, I've read every word of it.
13:34 DyingWeb: I throw up when I see it!
13:34 tommc: ... then produced a web app.
13:34 DyingWeb: you are lier and cretino
13:34 DyingWeb: HAHAHHAA
13:34 tommc: ?
13:34 tommc: Take a look at the Continue tutorial.  That's where I started.
13:35 asumu: tommc: No feeding the troll. :)
13:35 DyingWeb: i already am
13:35 DyingWeb: no templates, no nothing, just a bad looking vlog
13:35 DyingWeb: blog
13:36 DyingWeb: cretinetto
13:36 DyingWeb: cretinetto
13:36 DyingWeb: cretinetto
13:36 DyingWeb: cretinetto
13:36 DyingWeb: hahahaha
13:36 DyingWeb: http://www.shenlanguage.org/  -  all-time best language and best documentation ever!
13:36 tommc: I'm told you are trolling, and it appears to be so.  Otherwise I would have helped.
13:37 jonrafkind: stamourv, ban his ip
13:37 stamourv: Doing it now.
13:37 stamourv: Figuring out how to do it.
13:37 jonrafkind: yea haha im not quite sure how to do it myself
13:37 jonrafkind: the #freenode people know though
13:37 tommc: I don't understand... he was bashing a library, and yet recommending a language !?
13:37 stamourv: Ok, I'll go there.
13:38 Cryovat: You do /mode #racket +b *!*@hisip
13:38 bremner: meh, he's coming from a web gateway
13:38 stamourv: Actually, I'll check their FAQ first.
13:38 stamourv: Yeah, but the IP is recognizable, right?
13:38 stamourv: IIRC, he only ever used 2.
13:38 stamourv: Or we can ban all of Croatia, I guess...
13:38 jonrafkind: the web gateway uses his real ip
13:38 bremner: oh, ok.
13:38 jonrafkind: it does not use the ip of the web gateway
13:38 bremner: nifty
13:39 Cryovat: I'm kind of impressed
13:39 Cryovat: I would have expected language trolling in #ruby, but I thought the FP crowd would be a bit more civil ;)
13:39 bremner: you have never been on #lisp?
13:40 Cryovat: Years ago, when I had a Lisp class at uni
13:40 tommc: Being trolled has to be a sign of success? No?
13:40 Cryovat: Def
13:41 stamourv: Ok, his latest IP should be banned.
13:41 stamourv: Going after the others.
13:44 stamourv: I just hope he won't go back to the mailing list instead...
13:47 Cryovat: Anyone know what his problem is?
13:48 jonrafkind: i wouldn't expend effort thinking about it
13:48 stamourv: Meh, same as any troll, I guess.
13:50 (quit) cdidd: Remote host closed the connection
13:53 (quit) rudybot: Quit:  returned for regrooving
13:56 (join) Kaylin
13:59 Kaylin: ban is your only way to talk, PLT loosers! Mark Tarver is a hero! Shen is best lisp ever with best documentation!
13:59 bremner: itym losers. funny for a doc troll not to be able to spell.
13:59 (quit) dous: Remote host closed the connection
13:59 (part) tim-brown: "Leaving"
14:00 (join) rudybot
14:00 Kaylin: i can read enough to know web doc sucks!
14:01 Kaylin: but you are too prout to admit that truth!
14:01 Kaylin: proud
14:04 tommc: *boom*
14:04 bremner: stamourv: can you ban by ircname?
14:05 danl_ndi: whack-a-troll
14:05 stamourv: True Kaylin: Sorry about that.
14:05 stamourv: bremner: What do you mean?
14:05 stamourv: I'm just going by IP.
14:05 jonrafkind: if he has access to amazon can't he get inifinte ip's?
14:05 stamourv: Next step: ban all of Croatia.
14:05 stamourv: jonrafkind: They've all been from the same block.
14:06 jonrafkind: the same .8 you mean?
14:06 stamourv: Yeah.
14:06 stamourv: I think.
14:06 jrslepak_neu: two from the same .16, one from a separate block
14:08 jonrafkind: hopefully that guy is not actually mark tarvers. but that would be funny
14:08 bremner: stamourv: I was thinking of extbans under http://freenode.net/using_the_network.shtml, but those seem to be about authenticated users.
14:19 (quit) anRch: Quit: anRch
14:32 chandler: I think it might be better to mute him/her/it (+q) than to kickban.
14:33 stamourv: chandler: He/she/it keeps changing nicks, would that still work?
14:33 stamourv: That, and taking legitimate users' usual nicks, which could lead to collateral damage.
14:34 chandler: +q takes the same syntax as +b. You'll still need to play whack-a-mole, but I've found that a lot of users don't even notice they're muted, since they're still joined to the channel and jabbering away.
14:35 rapacity: make a bot that recognizes his patterns
14:35 stamourv: chandler: Good point.
14:35 rapacity: shen, web doc sucks etc, and ban on that
14:35 stamourv: If he/she/it comes back, I'll do that.
14:35 bremner: ban rapacity!
14:35 rapacity: T_T
14:36 stamourv: rapacity: Sounds NLP-complete.
14:36 chandler: Ah, but it isn't really. I've done it before.
14:36 rapacity: also for repeating the same message several times
14:36 rapacity: doesn't have to be perfect
14:36 stamourv: Right, I'm just afraid of false positives.
14:36 asumu: There's also a mode that only lets authenticated users join and/or talk, but that would have collateral.
14:36 stamourv: We're a friendly community, you know. ;)
14:38 (join) djcb
14:40 chandler: In this case, the bot could also probably use the hostmask as part of the decision too. It's better than banning or muting the whole ISP.
14:42 bremner: sounds like work though
14:42 (join) RacketCommitBot
14:42 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/KmTt_w
14:42 RacketCommitBot: [racket/master] revise phases section in small ways to match the guide's style - Matthew Flatt
14:42 (part) RacketCommitBot
14:42 (join) vkz
14:43 chandler: Yes, it is work. And work requires round tuits.
14:45 (quit) tommc: Remote host closed the connection
14:45 (join) dous
14:45 dous: chandler: I think it might be better to mute him/her/it (+q) than to kickban.
14:45 dous: stamourv: We're a friendly community, you know. ;)
14:45 dous: chandler: In this case, the bot could also probably use the hostmask as part of the decision too. It's better than banning or muting the whole ISP.
14:46 dous: such energy on wrong thing!
14:46 dous: you shoud rather ask: why i'm doing this?
14:46 chandler: dous: ??
14:46 dous: i'll stop
14:46 chandler: Oh, it's you.
14:46 dous: if you become smarter, guys
14:47 (join) dous_
14:47 dous: shame on you, folks!
14:47 chandler: dous: You do realize that dous_ can ghost you, right?
14:47 dous: stamourv: Next step: ban all of Croatia.
14:48 dous: such a shame!
14:48 (join) tommc
14:48 (part) dous
14:51 (quit) dous_: Ping timeout: 246 seconds
14:51 jonrafkind: so setting +q doesn't seem to help because he can see its being set
14:51 jonrafkind: so unless he has no idea what +q means it won't have the intended effect
14:52 chandler: I don't think it will work, given that he was reading the logs. But I've had some success with it in the past.
14:52 jonrafkind: which seems to be that he will chatter away not knowing that he is secretly ignored
14:58 (quit) _AlbireoX: Read error: Connection reset by peer
15:02 (join) fftb
15:12 (join) antithesis
15:14 (quit) snorble_: Read error: Connection reset by peer
15:14 (join) jeapostrophe
15:15 (join) snorble_
15:15 (join) dous_
15:19 (quit) dous_: Ping timeout: 260 seconds
15:24 tommc: jeapostrophe: I repeated my webserver test yesterday, and the current version still spews "thread: the custodian has been shut down: #<custodian>" under heavy load.
15:24 jeapostrophe: tommc: what version of racket did you use?
15:24 tommc: Latest head.
15:25 jeapostrophe: Are all the responses answered and does it stall?
15:25 samth: tommc: how are you running the test? can you post something so i can try to replicate it?
15:25 (join) MayDaniel
15:26 (join) untrusted
15:26 tommc: Requests all answered, and it eventually does stall if the test is made long enough.   I figured the stalling might be because the server/tester are on the same machine, and I'm starving file desc's
15:28 (quit) jao: Ping timeout: 246 seconds
15:28 tommc: http://pastebin.com/bHauPXcQ
15:30 tommc: Pretty no-thrills example, just testing for basic performance.
15:30 jeapostrophe: What number stalls it?
15:30 tommc: 11.3.0 Darwin Kernel Version 11.3.0: Thu Jan 12 18:47:41 PST 2012; root:xnu-1699.24.23~1/RELEASE_X86_64 x86_64
15:33 samth: tommc, jeapostrophe: i can get 20k on that experiment just fine
15:33 tommc: I see the error messages immediately, regardless of the number.  Throughput seems to drop (stall) after 3,000 or so connections.
15:33 jeapostrophe: I just did 300k requests with no drops or stalls
15:34 tommc: samth, jeapostrophe: Using ab, or another bench tool?
15:34 jeapostrophe: yes, ab
15:34 samth: tommc: ab
15:34 samth: your code and command line
15:34 tommc: Which OS?
15:34 samth: linux x86-64
15:34 tommc: It could be an osx issue...
15:34 jeapostrophe: with 1000 connections, it died
15:35 jeapostrophe: same with 500
15:35 samth: i think this output is a feature request for incremental gc:
15:35 samth: Percentage of the requests served within a certain time (ms)
15:35 samth: 50%     12
15:35 samth: 66%     12
15:35 samth: 75%     13
15:35 samth: 80%     13
15:35 samth: 90%     15
15:35 samth: 95%     16
15:35 samth: 98%     18
15:35 samth: 99%     20
15:35 samth: 100%    117 (longest request)
15:35 jeapostrophe: I agree
15:36 jeapostrophe: anyways, tommc, the error message is not a problem. i couldn't get it to not happen, but it used to cause the whole process to die, and now just that thread dies
15:37 (quit) vkz: Quit: vkz
15:37 jeapostrophe: i think that the concurrent connection limit is inherited from the racket i/o subsystem and not something the web server can do much about
15:37 tommc: Ok.  It does seem orthogonal to stalling... which is probably GC or file descriptor related.
15:37 samth: jeapostrophe: are you trying this on a mac?
15:37 jeapostrophe: no, ubuntu
15:39 samth: jeapostrophe: do you have a mac around you can test?
15:39 tommc: jeapostrophe: Didn't I read that racket had moved to a new I/O system? Using kqueue/epoll?
15:40 samth: tommc: yes, it did
15:40 samth: but that's about how it waits for things
15:40 samth: not how many concurrent things it waits for
15:40 samth: but i don't think either i or jeapostrophe could give a definitive answer on that
15:46 jeapostrophe: yes, and just because it using a scalable system call doesn't mean its own thread algorithms are scalabl
15:46 jeapostrophe: samth: i do not. i don't have a mac anymore
15:47 jeapostrophe: I am on a macbook, but it runs ubuntu
15:47 samth: jeapostrophe: how well does that work for you?
15:47 Cryovat: Do you bring it to Mac conventions to scare people?
15:47 Cryovat: ;)
15:47 samth is jealous of apple hardware design
15:47 offby1 is tempted to follow jeapostrophe's lead
15:47 jonrafkind: jeapostrophe, can you explain why you don't just run ubuntu in a vm inside osx?
15:47 offby1: love the hardware; software is mmm ...
15:47 Cryovat: My problem is that I like their hardware, but generally find myself horrified by the company behind it
15:48 offby1: on the other hand, sound and video always works, which is more than I can say for Linux :-|
15:48 Cryovat: I've been considering getting back on the Linux bandwagon
15:49 Cryovat: Been curious about Mint
15:49 Cryovat: I used to be an Ubuntu fan, but I'm among the Unity naysayers
15:49 tommc: samth: ok.
15:50 samth: tommc: since we don't have any mac users around atm, you might post on the mailing list about this
15:51 offby1 is among the Unity "Please, God, NOO!!!"-sayers
15:52 offby1 raises hand
15:52 offby1: I'm a mac user; is there something quick I can do to help?
15:52 Cryovat: Haha
15:52 chandler: Heh. It's not *that* bad.
15:52 offby1: yes it is
15:52 chandler: ... Once you turn off the stupid global menu bar thing.
15:52 offby1: I switched to xmonad because of it
15:52 offby1: maybe I should try that
15:52 (join) vkz
15:53 tommc: Interesting, Googled my original issue and found a thread from last year describing my initial problem. (Answered by Jay)
15:53 tommc: http://groups.google.com/group/racket-users/browse_thread/thread/62e5bc0f39aaae78
15:53 offby1: I hadda spend an hour figuring out how to programmatically disable all the new keyboard "shortcuts"
15:53 rudybot: http://tinyurl.com/6qh436c
15:56 samth: offby1: can you run apachebench following the instructions here: http://pastebin.com/bHauPXcQ
15:56 samth: and see if you get weird custodian errors?
15:57 offby1: well, sure; may take me a while since I assume apachebench will take a while to download & install
15:57 tommc: the apache bench which ships with OSX is broken, IIRC I had to patch it to make it work.
15:57 tommc: 10.7 that is.
15:58 offby1: is it worth me even trying the built-in one?
15:58 tommc: Try it in the shell
15:58 tommc: it either fails to run, or fails on the first connect.
15:58 chandler: samth: I'm getting a bunch of these: thread: the custodian has been shut down: #<custodian>
15:58 offby1: yep, failed
15:58 chandler: (5.2.1 on OS X 10.6.8)
15:58 samth: chandler: that's what tommc is seeing too
15:59 samth: but jeapostrophe and i couldn't repro on linux
15:59 chandler: 100%  18992 (longest request)
15:59 samth: yikes
15:59 samth: that's 100x worse than on my machine
15:59 tommc: offby1: httperf is a good alternative, one sec.
16:00 offby1: tommc: did someone say "siege" would work?  I see that's available from macports.
16:00 offby1: ah, httperf; I'll wait
16:01 tommc: brew install httperf (or however you do it), then: httperf --port=8888 --num-conns=10000  (Produces the same custodian errors)
16:01 chandler: samth: I'm seeing those custodian errors on Ubuntu Precise too.
16:01 offby1: tommc: ok, gimme a few minutes.
16:01 chandler: It's a virtual machine on the same system.
16:01 tommc: offby1: Yeah, I think siege is a wrapper around ab though...
16:01 chandler: 100%    617 (longest request)
16:02 samth: chandler: oh surprising
16:02 chandler: I'm going to try it on real hardware now.
16:03 tommc: Sorry, siege is distinct.  I was thinking about autobench.
16:03 chandler: Yes, I'm seeing it on real hardware too.
16:04 (quit) turon: Ping timeout: 248 seconds
16:04 offby1: installing httperf now ...
16:04 offby1 lights a pipe
16:04 tommc: ab -c 10 -n 10000 http://127.0.0.1:8888/
16:04 (quit) kanak: Ping timeout: 252 seconds
16:04 tommc: 100%    130 (longest request)
16:05 offby1: tommc: once I start httperf, what am I looking for?
16:05 tommc: Requests per second:    1173.45 [#/sec] (mean)
16:05 tommc: Time per request:       8.522 [ms] (mean)
16:05 tommc: Time per request:       0.852 [ms] (mean, across all concurrent requests)
16:05 tommc: Transfer rate:          229.19 [Kbytes/sec] received
16:05 (join) turon
16:06 offby1: still building OpenSSL ... /me drums fingers
16:06 tommc: In the terminal showing the webserver output, you might see custodian errors.
16:07 tommc: On the plus side... the server seems pretty quick ;-)  1173.45 req/s isn't too shabby.  I should probably benchmark nginx for a good upperbound.
16:07 tommc: Apple's and oranges I know.
16:07 samth: ah, i'm now getting it on ubuntu as well
16:08 offby1: aha
16:08 offby1: and I'm still building OpenSSL :)
16:08 samth: ok, tommc, you should either report this as a bug or post to dev@racket-lang.org
16:09 tommc: Will do. (In an hour or so)
16:09 samth: tommc: great
16:09 (nick) tommc -> tommc_away
16:11 offby1: ah done finally
16:11 offby1: might's well give it a whirl
16:12 offby1: yep, lotta ``thread: the custodian has been shut down: #<custodian>''
16:18 aidy: What would be the easiest way to change a canvas' coordinates to right-handed orientation?
16:18 aidy: s/canvas/dc
16:19 (join) yoklov
16:21 offby1: send it to one of those camps
16:22 (part) untrusted: "ERC Version 5.3 (IRC client for Emacs)"
16:28 (quit) djcb: Remote host closed the connection
16:34 aidy: that's mean :(
16:40 (quit) MayDaniel: Read error: Connection reset by peer
16:40 (join) djcb
16:48 jeapostrophe: jonrafkind: i don't run it in a VM because I want a full X11 experience
16:48 jeapostrophe: jonrafkind: with Xmonad and all that goodness
16:49 jeapostrophe: samth: using the mac hardware works out well, getting the right configs was a pain. there are still things i don't have i want (like right click with a key and the one click) but overall it is good
16:49 jeapostrophe: samth: i think my battery and video card are spottily supported, so i intend to buy a system76 laptop someday
16:50 jeapostrophe: tommc_away: offby1: the custodian errors are safe; i'd like to get rid of them, but i think it is a problem in the 'run-server' function from mzlib/thread
16:50 jeapostrophe: sorry for the hour delay... students kidnapped me
16:51 (nick) tommc_away -> tommc
16:51 tommc: jeapostrophe: np at all... lunch kidnapped me.
16:52 tommc: jeapostrophe: Do you want me to file it as a bug?
16:55 tommc: BTW, one suggestion for the web server is to have a scaffolding generator.  Nothing too complex, but enough to get someone started with some sensible defaults. Static file dir, html5 template etc.
16:55 (join) RacketCommitBot
16:55 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/JpkoUg
16:55 RacketCommitBot: [racket/master] Hold reference to original thread during BEGIN_ESCAPABLE in place-channel sync - Kevin Tew
16:55 (part) RacketCommitBot
16:56 tommc: (It might even help that troll guy, who complains about the docs)
17:00 offby1 idly wonders who ransomed jeapostrophe
17:04 jeapostrophe: tommc: no, i don't want you to file it, i've already marked it as analyze somewhere else
17:05 jeapostrophe: tommc: i'm not sure what scaffolding you could need, the web-server is pretty scaffold-less... you just need one file, you don't even need to save it. the default arguments to serve/servlet kind of are the sensible defaults. but if you have specific suggestions, i'm open
17:05 (quit) realitygrill: Quit: realitygrill
17:07 tommc: Basically just a project creator. To create a directory structure for a 'real world' app.  I'm not really talking about a convenience function.
17:08 (quit) fftb: Remote host closed the connection
17:09 samth: jeapostrophe: it might be worth a discussion on the list
17:09 (join) bitonic
17:09 samth: since it's not obvious how to fix run-server
17:09 tommc: Frameworks like Rails, Django, Play etc, offer something equivalent.
17:10 jeapostrophe: i emailed mflatt+bugs about it when i first discovered it and got no answer
17:10 samth: jeapostrophe: just re-ping on dev@
17:10 samth: mflatt often requires multiple ping attempts
17:11 samth: i think the problem is "kill-safety", and can be fixed by rereading the PLDI paper 10 or 20 times :)
17:12 jeapostrophe: tommc: i think i know what you mean, but a web-server in racket does not need any directory or any other files
17:12 jeapostrophe: and if you want one, you just do: #:server-root-path (current-directory)
17:13 jeapostrophe: and all the static files can go right next to the source
17:13 jeapostrophe: a script that does "mkdir www; touch www/go.rkt" and echos that seems a bit...
17:13 jeapostrophe: trivial?
17:14 tommc: I know, but in this day and age, people sadly don't read manuals.  They want to type one command, have a mini site, html5 templates, directories created for them etc.  Personally, I'm not an ideal use case for this, just trying to advocate for greater adoption.
17:15 tommc: *If* the simplicity captures them, then they'll invest time into reading deeper.
17:15 samth: jeapostrophe: it should do more than touch go.rkt
17:15 samth: it should create a "hello world" html file
17:15 samth: and a css
17:15 samth: file
17:15 samth: and a racket file that serves it
17:15 Cryovat: You know tommc
17:15 samth: preferably demonstrating templating in a trivial way
17:15 Cryovat: I used to love stuff like that
17:15 Cryovat: But after working with web frameworks for a while, I get more and more wary of them
17:15 tommc: samth: Yes, perhaps with something like http://html5boilerplate.com/
17:15 Cryovat: There's just too many surprises
17:16 jeapostrophe: i don't understand what the difference between that and copy&pasting out of the docs is
17:17 tommc: Cryovat: Agreed, I avoid frameworks for that reason myself. I always end up having to customize them.
17:19 tommc: jeapostrophe: I suppose it's not worth creating until someone demands it, and don't count my suggestion as a demand :)
17:20 (quit) vkz: Quit: vkz
17:21 jeapostrophe: tommc: i can count it if you want :)
17:23 tommc: One thing I find quite helpful though. Many frameworks/servers provide a way to start up in 'devel' mode, that auto reloads when changes occur.  I made a silly script to do just that for working on a racket web app. https://raw.github.com/zussitarze/rackmon/master/rackmon
17:23 jeapostrophe: that is not generally possible with continuations
17:23 tommc: My solution is rather brutal, it kills/restarts a racket instance on any change.
17:24 jeapostrophe: you can't port a continuation to another new source file
17:24 jeapostrophe: the default server gives you /conf/refresh-servlets if you use module servlets
17:24 (quit) yoklov: Quit: computer sleeping
17:25 jeapostrophe: but killing and restarting is basically the only option since the servlet is in the same module namespace / address space as the server
17:26 Cryovat: Might be my Java upbringing, but every time I see the term servlet I think of jsp and shudder :/
17:27 tommc: Right. And this is simply for developer ease, rather than a production hot-swap.  So it doesn't matter if the continuations are lost.
17:27 tommc: Your stateless servlets should survive restarts though, right? (Assuming the servlet code is unchanged)
17:28 jeapostrophe: yes, but then why would you restart if you hadn't changed anything (in this "dev" mode)
17:31 tommc: jeapostrophe: What if the change is outside of the lexical scope of the captured continuation?
17:31 jeapostrophe: it computes a checksum of your code to ensure that no data-structures, etc are changed
17:31 jeapostrophe: only if it outside of the continuation's module will it survive
17:31 jeapostrophe: but that's a soundness flaw
17:32 tommc: Ok.
17:34 tommc: Most of the time, a restart is useful for viewing a design change in a template.
17:46 tommc: jeapostrophe: Looking down the line, are there any big changes coming to the webserver?
17:46 (quit) jonrafkind: Ping timeout: 260 seconds
17:47 jeapostrophe: tommc: nope, it is pretty stable
17:48 (join) jrslepak
17:49 tommc: For what it's worth, I find the documentation to be quite readable.  I'm not sure if it's a common stumbling block, or just for one person on the mailing list (whom spends his life complaining about it)
17:50 jeapostrophe: tommc: here is a dev mode: https://github.com/jeapostrophe/exp/blob/master/devmode/dev.rkt
17:50 jeapostrophe: you just give it a path to your servlet's module
17:50 jeapostrophe: then you write the call to serve/servlet
17:50 jeapostrophe: and it will notice changes and kill/recall your server
17:50 tommc: Cool. I'm going to try that now.
17:51 jeapostrophe: it monitors once per second... you could change that
17:52 tommc: Right. So this is similar to mine, but preserves the vm.
17:53 jeapostrophe: yup
17:53 jeapostrophe: and only looks at one file
17:55 jeapostrophe: gtg
17:55 jeapostrophe: <3
17:55 tommc: thanks for your help. by.
17:55 tommc: bye.
17:56 (quit) antithesis: Quit: yes leaving
18:00 (quit) jeapostrophe: Ping timeout: 272 seconds
18:15 (join) drmark
18:15 drmark: tommc: For what it's worth, I find the documentation to be quite readable.  I'm not sure if it's a common stumbling block, or just for one person on the mailing list (whom spends his life complaining about it)
18:15 drmark: what a sleezy toady!!!!!!!!!
18:15 (join) dous
18:17 drmark: that documentation is completly unreadable. If you want to know what the good doc looks like, then read MSDN, or Charles Petzold or MArk Tarver!
18:18 (part) drmark
18:20 (quit) dous: Ping timeout: 246 seconds
18:23 (join) yoklov
18:45 danl_ndi: doesn't irc have a "ban" mode where the person isn't heard but doesn't know it?
18:47 Cryovat: Yes, /ignore <nick>
18:47 Cryovat: Unfortunately, that depends on the person in question not changing his nickname
18:48 Cryovat: I guess you could ignore *!*@gateway/web/freenode/*, but that would filter out legitimate people too
18:49 danl_ndi: should take him longer to realize he's ignored, so he'll switch nicks less often (especially it he's been banned up til now)
18:50 tommc: I ignored his 'indiscretion', and he left naturally.
18:56 bremner: I find the irssi alias  "/alias mute ignore -time 1200 -replies" handy
19:26 (quit) dzhus: Remote host closed the connection
19:33 (nick) samth -> samth_away
19:37 (quit) tommc: Remote host closed the connection
20:01 Diarmid: A more interesting tactic might be to point out that Shen can't be so great, since it doesn't even come with support for the web.  At least he'd have to find a new thing to whinge about, since he's always on about Racket documentation and web support.
20:02 Diarmid: Of course I already pointed out to him that Shen has very limited I/O support.
20:03 Diarmid: Or there's the old "shut up and code".  "Why don't you show us how to do it better."  That'd require actual work, so should get rid of him.
20:08 offby1: or there's just ignore him.
20:09 Diarmid: offby1: Seems to be hard to do with him changing nicks and IP's.
20:09 (quit) yoklov: Quit: computer sleeping
20:10 offby1: *shrug*
20:10 offby1: in my experience he announces himself pretty clearly with "hello community" or similar.
20:11 chandler: Diarmid: Have you ever heard the expression "don't feed the troll"?
20:11 offby1 nods gravely
20:11 Diarmid: chandler: Feeding trolls usually has to do with getting angry at them, and becoming irrational.  I don't think I've ever seen a troll 'fed' by having their argument scuttled.
20:12 chandler: Heh. I definitely have seen that. They don't admit defeat.
20:12 offby1: depends what you mean by "fed", I guess.
20:12 chandler: They're not rational animals.
20:12 offby1: to me, any attention at all counts as feeding.
20:14 bremner: Diarmid: I certainly don't want to start discussing shen with some maniac. For all I know, shen is OK.
20:14 Diarmid: Iit would be nice to have a way to ban or ignore him, or get him to wander off by making him look like the jackass he is.  I enjoy exposing myself to the risk of learning something by reading channel history, and he adds a lot of noise to the signal there.
20:17 Diarmid: bremner: Shen basically lets you specify 'computation' just fine.  It's not a terribly practical language, due to lack of I/O facilities (GUI, graphics, sound, networking), but might become so someday.
20:18 Diarmid: Judging from the writing on lambdaassociates (e.g. http://www.lambdassociates.org/blog/nextlisp%281%29.htm and http://www.lambdassociates.org/blog/the_problems_of_open_source.htm) I don't find it that hard to believe that Racketnoob *is* the Qi/Shen author, since the essays are pretty trollish.
20:18 bremner: if so, that is really sad.
20:19 (join) yoklov
20:20 Diarmid: bremner: I suppose so.  The main thing that makes life on earth miserable is each other. :P
20:21 bremner: "hell is other people" is shorter ;)
20:24 Diarmid: Pretty much.
20:25 (join) tommc
20:27 asumu: I doubt Racketnoob is the Shen author, given that I think Mark Tarver is the author and is British.
20:27 asumu: While Racketnoob is from Croatia or Hungary.
20:28 chandler: For all we know, Racketnoob is trolling #shen and saying that Racket is better!
20:28 asumu: That would be pretty funny actually.
20:34 (join) francisl
20:35 Diarmid: If it's not Tarver, my next guess is Xah Lee.
20:35 Diarmid: Either way, it'd be nice to be rid of them.
20:47 (join) jeapostrophe
20:52 (quit) francisl: Remote host closed the connection
20:53 (join) francisl
20:57 offby1: Le enfer -- es les autres!
20:57 offby1 's attempt at reverse-engineering the original French
20:59 (quit) mceier: Quit: leaving
21:01 (part) jschuster
21:01 (join) realitygrill
21:02 (quit) jrslepak: Quit: This computer has gone to sleep
21:03 tauntaun: offby1: to which language?
21:03 (quit) francisl: Quit: francisl
21:03 (quit) tommc: Remote host closed the connection
21:05 offby1: _to_ french _from_ english
21:05 offby1: sans caps
21:05 offby1 's attempt at reconstructing the original French via the technique of reverse-engineering
21:06 offby1: is what I should have said.
21:06 offby1: tauntaun: if you're a Francophone, have at it
21:06 tauntaun: FYI: "L'enfer, ç'est les autres."
21:06 offby1: tx
21:06 offby1: _I_ would say I was pretty close given that I don't speak a word of French.
21:06 tauntaun: One of Sartre's sad truths.
21:07 offby1: ah, another IRC-on-linode fella
21:07 tauntaun: Yeah, I'm so lumpen.
21:08 offby1: I used to do that; I think EC2 is a bit cheaper.
21:11 (quit) turon: Remote host closed the connection
21:16 (join) dous
21:21 (quit) dous: Ping timeout: 246 seconds
21:23 (join) francisl
21:31 chandler: offby1: if it's cheap you want, there's always prgmr.com too
21:34 offby1: I think I tried that.
21:34 offby1: turned out to be unreliable.
21:35 chandler: Well, it *is* cheap.
21:35 offby1: ayup
21:36 chandler runs his own virtualized hosting
21:36 chandler: because I'm just that kind of crazy
21:37 offby1: why am I not surprised
21:41 (join) tommc
21:43 (join) samth_
21:43 (quit) bitonic: Quit: WeeChat 0.3.5
21:46 (join) dous
21:56 asumu: Hmm. If you macro step this code: (provide (struct-out point)) (struct point (x y))
21:56 asumu: you end up getting this primitive provide: '(#%provide point struct:point point point? point-y point-x)
21:57 asumu: Which doesn't make sense because that provide line lacks the constructor function (which is alpha-renamed to point1).
21:57 asumu: Anyone have any idea what's going on?
22:03 (join) jrslepak
22:03 offby1: nope!
22:03 offby1: isn't "point" -- the first name provided -- the constructor?
22:04 offby1: oh wait, it's provided "point" _twice_?
22:04 offby1: that seems weird
22:04 offby1 doesn't know what "alpha-renaming" is anyway
22:04 (quit) karswell: Read error: Connection reset by peer
22:05 (quit) francisl: Quit: francisl
22:08 (quit) tommc: Remote host closed the connection
22:09 (join) AlbireoX
22:09 (join) karswell
22:12 (quit) dous: Remote host closed the connection
22:12 (join) dous
22:16 (quit) dous: Ping timeout: 244 seconds
22:24 (quit) mario-goulart: Read error: Connection reset by peer
22:24 asumu: offby1: The macro expander renames the `point` constructor to not conflict with the `point` struct type definition. I think.
22:24 (quit) masm: Quit: Leaving.
22:25 (join) mario-goulart
22:25 asumu: But for some reason `struct-out` doesn't reflect that(?).
22:25 offby1: I'd have assumed the struct type definition was struct:point
22:25 asumu: Which doesn't seem possible since struct-out actually works.
22:25 offby1: anyway as you can see I don't have a clue
22:26 asumu: offby1: there are two bindings: `struct` at the transformer level and `struct:point` at runtime.
22:26 asumu is confused too
22:28 (part) mimcpher
22:49 (join) realitygrill_
22:49 (quit) realitygrill: Ping timeout: 260 seconds
22:49 (nick) realitygrill_ -> realitygrill
23:03 tauntaun: offby1: of course you know what alpha-renaming is.
23:04 offby1: of course.
23:04 offby1: gna
23:04 offby1 glances around nervously.
23:04 tauntaun: AFAIR, it's renaming a bound variable so that it doesn't shadow or get shadowed.
23:16 asumu: Alpha renaming is just what IDEs do when they use the "rename" refactoring. :p
23:26 Demosthenes: ugh, racket's so much more awesome than python. why does it get all the libraries?! :P
23:36 (join) cdidd
23:50 (quit) samth_: Ping timeout: 246 seconds
23:51 (quit) jrslepak: Quit: What happened to Systems A through E?
23:55 (join) jonrafkind