00:50 (quit) jonrafkind: Ping timeout: 265 seconds
03:19 (join) masm
04:08 (join) jt`
04:09 (quit) jtolds: *.net *.split
05:07 (quit) jao: Ping timeout: 265 seconds
05:18 (quit) carleastlund: Quit: carleastlund
06:01 (join) jao
09:05 askhader: I'm trying to do some string comparison, basically I have a large data structure of words and I want to check to see how many times a word occurs in in that datastructure. The problem is that I am currently using string-ci=? to compare and so say I am looking for the word 'apple' it won't match 'apple?'- I could write a function to do some smart comparison but I feel like this may be slow and inefficient. Do you fellas have any suggestions?
09:09 chandler: askhader: I'd suggest using a regexp to match whatever you consider to be non-word characters.
09:10 askhader: Hm good idea.
09:10 askhader: Probably for non-word characters at the beginng and the end of the block
09:11 askhader: as well as 's' for the plural
09:18 askhader: is there going to be a dr racket?
09:24 chandler: There already is.
09:24 (quit) masm: Quit: Leaving.
09:25 askhader: badass
09:30 askhader: Is there any way for me to stipulate the ignorance of case when matching a regexp?
09:33 chandler: It's like I'm getting your questions in stereo! :-)
09:33 chandler: Try using an internal modifier like (?i) in the regexp.
09:33 askhader: =P these /are/ two different channels and the questions just so happen to be pertinent to both :o)
09:34 (quit) sstrickl: Quit: sstrickl
09:50 (join) sstrickl
10:03 (join) superjudge
10:06 (quit) superjudge: Quit: superjudge
10:09 (quit) atak: Read error: Connection reset by peer
10:11 (join) djahandarie
10:25 (join) atak
10:57 (nick) jt` -> jtolds
11:30 (join) jonrafkind
11:35 (join) stamourv
12:09 (join) carleastlund
12:14 (quit) jao: Ping timeout: 240 seconds
14:11 (join) Hezy
14:12 (part) Hezy
14:22 (join) superjudge
14:26 (quit) carleastlund: Quit: carleastlund
14:28 (join) schang
14:43 (join) jao
14:47 (join) elibarzilay
14:47 elibarzilay: jonrafkind: ping
14:48 (quit) elibarzilay: Client Quit
14:48 jonrafkind: eli, ping
14:48 eli: jonrafkind: You're supposed to be ponging.
14:48 jonrafkind: pongification
14:48 eli: Anyway, I added the irc logs and the browser widget thing.
14:49 jonrafkind: whats the url for the logs
14:49 eli: See both on the community page.
14:49 jonrafkind: i pressed shift-f5 3 times and i odnt see anything new
14:50 eli: Oh right, I forgot to update the public ones.
14:50 eli: Try now.
14:51 jonrafkind: ok. my recommendation would to make "#racket on freenode.net" a link to the webchat thing. but if you dont want to change it I won't press for it
14:51 eli: I don't mind either way -- I tried that first, but then I thought that it's less obvious to do it that way.
14:51 jonrafkind: i find it more obvious, but whatever
14:52 eli: Maybe both?
14:52 jonrafkind: ok
14:52 eli: It also looks ugly when the color is different.
14:52 eli: (Refresh to see it.)
14:52 jonrafkind: you can use css for that
14:53 eli: Since the mail links above that are not links.
14:53 eli: Changing the css will be a bad idea.  The ugliness follows from the email things not being links and this is a link.
14:53 jonrafkind: ok maybe a "click here to join" wording instead of "a browser client"
14:54 eli: It should be common knowledge that "click here" should not appear in link texts...
14:54 jonrafkind: its not clear what link the "browser client" goes to
14:54 jonrafkind: it might goto a wiki page that describes what a browser client is
14:55 (join) carleastlund
14:56 eli: jonrafkind: So?
14:57 jonrafkind: yay
14:57 eli: I wasn't asking for a "yay" reply, but for a suggestion to avoid the "click here" thing...
14:58 jonrafkind: ok in my own page where I use the freenode webchat thing, I just have a link that says "Chat on IRC"
14:58 (quit) carleastlund: Client Quit
14:58 (join) carleastlund
14:58 jonrafkind: for racket I would have "Chat on IRC" as a link followed by " -- using #racket on freenode.net. Logs are at blah"
14:59 eli: jonrafkind: You know what, here's the whole line:
14:59 eli: @text{@TT{@big{@strong{#racket}}} on @a[href: "http://freenode.net"]{@tt{freenode.net}} @mdash an informal discussion channel for all things related to Racket. @irc-chat{Click here to join}, or @irc-logs{browse the logs}.}
15:00 eli: jonrafkind: Just do whatever you want with it and mail me a new copy.
15:00 eli: (Or if you do that after I push out, then just edit the code directly.)
15:03 samth: eli, http://mg.pov.lt/irclog2html/
15:04 eli: samth: I am definitely not going to use that.
15:04 samth: eli, why not?
15:04 eli: jonrafkind: That looks even more odd but whatever.
15:05 eli: samth: Because text logs are much more useful, for example, Robby uses that to send himself a note when someone mentions him, I think, and I definitely do that for #scheme.
15:05 (topic) samth: Racket: http://racket-lang.org (logged @ http://racket-lang.org/racket/)
15:06 samth: text logs are good, but it's nice to have readable logs as well
15:06 eli: If you want colorization etc, then that should be done in addition, and goes well beyond my quota of stuff.
15:06 eli: So feel free to set whatever up...
15:06 samth: you could just have the text file and the output of the colorizer
15:06 samth: where would i set it up?
15:06 samth: surely you have some script somewhere doing this
15:07 eli: FWIW, there's more than just the colored output -- there's pastes, quotes, searches, interactive searches, statistics, etc etc etc.
15:07 eli: I've seen some pretty huge irc things do such things.
15:07 eli: And writing one of these things from scratch is wasted time.
15:07 samth: well, my itch is readable logs
15:07 samth: i agree
15:07 eli: You know what you shouldn't do with bad itches, right?
15:08 eli: In any case, just do whatever (I recommend looking for something over doing your own thing), and I'll just link to wherever the results are.
15:09 samth: that's what that link was - a script
15:11 eli: Pointing somewhere is three orders of magnitudes easier than installing it, running it, making sure it stays alive, making it accessible, etc.
15:13 jtolds: so i talked to matthew flatt about this a few weeks ago, but he mentioned that the underlying racket event driven architecture just uses select on file descriptors. is that still true?
15:13 eli: Yes.
15:14 jtolds: any plans to switch to something like libevent?
15:14 eli: What's the difference?
15:14 jtolds: huge. select is incredibly unperformant when you have tons of file descriptors
15:14 jtolds: the sort of canonical explanation of the problem is http://www.kegel.com/c10k.html
15:15 eli: But we're not really using select, we're using a layer on top of that.
15:15 jtolds: anyway, event driven architectures that use select are sort of non-starters in hpc environments, but it would be cool to be able to use racket in such an environment
15:15 djahandarie: Haskell recently moved to a epoll-based event library
15:15 jtolds: no i know, but the way select is implemented in the kernel is, while cross-platform, not good enough
15:16 jtolds: oh beautiful djahandarie. i didn't know that
15:16 djahandarie: Yeah, it was near the start of this year
15:16 eli: The issue of cross platformed-ness is important.
15:16 eli: Specifically, Windows and OSX...
15:16 jtolds: right, so that's why i suggested libevent
15:17 jtolds: there are a few other options
15:17 eli: But since this is an implementation detail (that is, below what's visible from inside racket), then it's best to send it on the list.
15:17 jtolds: oh okay, i was more just curious about if there had been any discussion
15:17 eli: And a patch always works better -- though including a new library is not that easy.
15:17 jtolds: yes i understand :)
15:17 eli: I don't think there was any...
15:17 jtolds: okay
15:24 (join) ice_man`
15:26 mario-goulart: Hi.  Do the web libraries for racket support multipart/form-data requests?  If so, can you give me a link to docs/source?
15:28 eli: mario-goulart: I think so, but Jay would have more to say.
15:29 mario-goulart: Ok, eli, thanks.  I'm gonna ask Jay when he shows up here or on #scheme.
15:51 (join) asdf_
15:52 (quit) asdf_: Client Quit
16:46 (quit) jonrafkind: Read error: Connection reset by peer
16:47 (join) jonrafkind
16:51 askhader: are there libraries for constructing graphs?
16:53 samth: askhader, what kind of graphs?
16:53 askhader: for ploting nodes and edges .etc
16:54 samth: i think people normally just call out to 'dot'
16:54 samth: that's what redex does, for example
16:54 askhader: dot?
16:54 askhader: ah
16:56 (quit) superjudge: Quit: superjudge
16:56 askhader: Actually, I'm not certain to what you're referring.
16:56 askhader: I got confused by something that looked graphical and was called 'dot in the docs.
16:56 mario-goulart: askhader: maybe http://chicken.wiki.br/eggref/4/format-graph
16:56 sstrickl: askhader: dot is part of graphviz: http://www.graphviz.org/
16:57 askhader: thanks
16:58 mario-goulart thought here was #scheme
16:58 mario-goulart: Sorry for the chicken link.
16:59 samth: we forgive you, mario-goulart :)
16:59 askhader: Have any of you tested http://docs.racket-lang.org/profile/index.html?q=graphviz#(mod-path._profile/render-graphviz)  ?
17:00 (join) masm
17:01 eli: askhader: That's intended to be used with graphviz.  I have some code that tries to do the layout in scheme, but never finished it.
17:01 askhader: Ah
17:01 askhader: So how do you go about making calls to 'dot'?
17:01 askhader: I have a bunch of data that I would like to plot.
17:03 eli: Then this is not what you want -- all it does is take profiling results and dumps them out in a graphviz syntax.
17:04 eli: You probably want to create such output yourself.
17:04 eli: The graphviz syntax is very simple, btw.
17:04 askhader: Oh so I'd be constructing graphiz output
17:05 mario-goulart: Thanks, samth. :-)
18:29 (join) Accidus
18:31 Accidus: The installer fails on move/copy-distribution of a README file (if I parse the error message correctly). I saw a bug report and a reply by Eli that it's fixed, but apparently it isn't. I'm using the Ubuntu jaunty installer.
18:38 jonrafkind: Accidus, when did you download the installer
18:38 Accidus: just now
18:51 (quit) jao: Read error: Connection reset by peer
18:59 Accidus: Okay, managed to by-pass the bug
18:59 Accidus: I halted the installer in the middle, and manually moved the README file.
19:02 stamourv: doing a self-contained install shpuld work
19:02 stamourv: you tried a Unix-style install, right?
19:12 Accidus: stamourv, Aye, I downloaded the Ubuntu jaunty installer, did a unix-style install. The rest are the defaults
19:12 Accidus: stamourv, I think that the problem with the script is that it forgets to mv the README file from the tmp directory
19:13 (join) jao
19:13 Accidus: stamourv, and then, in the move/copy-distribution function, when it checks (line 406) that it has moved everything, the check fails
19:13 Accidus: stamourv, so the installation is unrolled
19:14 Accidus: That's why, if I put the process to sleep in the middle, and manually move the README file, the installation succeeds
19:14 stamourv: it's fixed in git
19:14 stamourv: I guess it should have been included in the release but wasn't
19:14 Accidus: I guess
19:25 (quit) jtolds: Ping timeout: 252 seconds
19:26 (join) jt`
19:26 (nick) jt` -> jtolds
19:35 (nick) samth -> samth_away
20:05 (quit) masm: Quit: Leaving.
20:13 (quit) schang: Quit: leaving
20:21 (quit) atak: Ping timeout: 240 seconds
20:40 (quit) jonrafkind: Ping timeout: 258 seconds
20:52 (part) stamourv: "ERC Version 5.2 (IRC client for Emacs)"
21:25 (quit) carleastlund: Quit: carleastlund
21:31 (quit) mario-goulart: Ping timeout: 260 seconds
21:51 (join) Shimei2
21:58 (quit) Shimei2: Ping timeout: 240 seconds
22:06 (join) carleastlund
22:09 (join) Quetzalcoatl_
22:12 eli: Accidus: No, I didn't get to rebuild the installers, yet.
22:53 (quit) Quetzalcoatl_: Remote host closed the connection