00:00 jimmypk: Ya, like I said this is really bizarre to me. Maybe I'll email the developers list tomorrow.
00:00 Shambles_: jimmypk: I haven't tried to do that in ages.  On Windows the stack size for the main thread (other threads are determined at run time by the arguments sent to the functions that start them) is stored inside the executable's binary headers.  I don't know how it works on Linux.  I wouldn't be surprised if it was similar.  That'd mean you'd need to recompile it.
00:00 Shambles_: jimmypk: I'd also be surprised if setting a system variable would effect the executables, if that is, indeed, how it works.
00:01 jimmypk: stackoverflow is telling me it's an environment variable set by ulimit -s on linux.
00:02 Shambles_: http://stackoverflow.com/questions/2275550/change-stack-size-for-a-c-application-in-linux-during-compilation-with-gnu-com
00:02 rudybot: http://tinyurl.com/87eqaxd
00:07 jimmypk: Alright, well I'm going to give up for the night. I'm going to try emailing the developers tomorrow. I really appreciate all of your help :) Take care.
00:07 (join) jrslepak
00:09 (quit) jimmypk: Quit: leaving
00:15 (quit) jao: Ping timeout: 246 seconds
00:39 (quit) jeapostrophe: Ping timeout: 245 seconds
00:45 (quit) hash_table: Ping timeout: 260 seconds
00:45 (quit) getpwnam: Ping timeout: 272 seconds
00:48 (join) veer
00:53 (quit) veer: Client Quit
00:57 (quit) genbattle: Ping timeout: 245 seconds
01:06 (quit) ozzloy: Quit: leaving
01:07 (join) ozzloy
01:07 (quit) ozzloy: Changing host
01:07 (join) ozzloy
01:12 (join) otterdam
01:14 (quit) mceier: Quit: leaving
01:18 (join) kreol[Ukr]
01:25 (join) gciolli
01:32 (quit) gciolli: Quit: Leaving.
01:33 (join) gciolli
01:34 (quit) adu: Quit: adu
01:43 (quit) yoklov: Quit: bye!
01:52 (join) hkBst
01:56 (quit) gciolli: Ping timeout: 255 seconds
02:00 (join) Shviller
02:04 mithos28: rudybot: (impersonate-struct s impersonator-prop:application-mark 2)
02:04 rudybot: mithos28: error: reference to an identifier before its definition: s in module: 'program
02:04 mithos28: rudybot: (impersonate-struct (set) impersonator-prop:application-mark 2)
02:04 rudybot: mithos28: error: impersonate-struct: expects type <structure accessor or structure mutator> as 2nd argument, given: #<chaperone-property>; other arguments were: (set) 2
02:05 mithos28: I am using impersonate-struct wrong or is this a bug?
02:10 (quit) dnolen: Ping timeout: 276 seconds
02:10 mithos28: chaperone-struct seems to have a restriction that would imply this, but not impersonate-struct
02:32 (join) adu
02:39 (join) mceier
02:52 (join) RacketCommitBot
02:52 RacketCommitBot: [racket] plt pushed 5 new commits to master: http://git.io/X7FnIQ
02:52 RacketCommitBot: [racket/master] reformatting only - John Clements
02:52 RacketCommitBot: [racket/master] working on adding stopping dialog - John Clements
02:52 RacketCommitBot: [racket/master] adding halt dialog for runaway process - John Clements
02:52 (part) RacketCommitBot
02:54 (quit) jonrafkind: Ping timeout: 248 seconds
02:58 (quit) adu: Quit: adu
03:02 (quit) mithos28: Quit: mithos28
03:21 (join) djcoin
04:07 (quit) SHODAN: Remote host closed the connection
04:10 (join) SHODAN
04:15 (join) dalaing
04:27 (join) wahjava
04:38 (join) bitonic
05:11 (join) tfb
05:29 (quit) ivan\: Ping timeout: 245 seconds
05:30 (join) ivan\
05:34 (quit) dalaing: Quit: Leaving.
05:35 (join) dalaing
05:46 (join) antithesis
05:47 (quit) tfb: Quit: gone
06:04 (quit) antithesis: Quit: antithesis
06:04 (join) antithesis
06:09 (join) gciolli
06:09 (quit) dalaing: Ping timeout: 248 seconds
06:22 (quit) antithesis: Quit: antithesis
06:22 (join) antithesis
06:31 (quit) gciolli: Ping timeout: 244 seconds
06:36 (join) lamefun3
06:36 lamefun3: hello
06:36 lamefun3: can I run several independent Racket contexts in one program?
06:41 (part) lamefun3
06:41 (quit) antithesis: Quit: antithesis
06:42 (join) antithesis
06:50 (join) jesyspa
06:54 (join) lamefun3
06:54 lamefun3: hi
06:54 lamefun3: racket download page is confusing
06:55 lamefun3: is Windows, UNIX and Mac OS sources same?
06:55 bremner: iirc, no
06:56 (quit) noam: Ping timeout: 246 seconds
06:56 (join) noam
06:57 (join) gciolli
06:57 lamefun3: and how to verify download?
07:03 (quit) gciolli: Ping timeout: 252 seconds
07:04 (join) nilyaK
07:04 (quit) vu3rdd: Ping timeout: 245 seconds
07:09 (quit) rbarraud: Ping timeout: 276 seconds
07:16 bremner: are there no md5sums or whatever listed on the download page?
07:16 lamefun3: no
07:16 lamefun3: http://download.racket-lang.org/racket-5-2-1-src-unix-tgz.html
07:16 (quit) kreol[Ukr]: Ping timeout: 244 seconds
07:20 (part) lamefun3
07:21 bremner: ok then, I guess I will stop investigating
07:33 (join) snearch
07:52 (quit) nilyaK: Quit: Leaving.
07:55 (join) jeapostrophe
07:55 (quit) jeapostrophe: Changing host
07:55 (join) jeapostrophe
08:03 (quit) antithesis: Quit: antithesis
08:05 (join) untrusted
08:09 (join) kanak
08:09 (join) masm
08:23 (quit) kanak: Quit: Leaving.
08:34 (join) kanak
08:51 (join) kreol[Ukr]
08:58 Shambles_: Does anybody know when the GUI guy will be back, and if so, if he tends to read backlogged stuff on the mailing list?
09:00 (join) getpwnam
09:01 (join) hash_table
09:02 samth: Shambles_: matthew will probably see your message at some point, but i recommend patience
09:02 samth: i don't know how long he's in china
09:16 (quit) jeapostrophe: Ping timeout: 244 seconds
09:21 (quit) snearch: Quit: Verlassend
09:26 (join) RacketCommitBot
09:26 RacketCommitBot: [racket] plt pushed 3 new commits to master: http://git.io/KT5OUw
09:26 RacketCommitBot: [racket/master] make scribble's part/document/element/etc structs all be #:transparent - Robby Findler
09:26 RacketCommitBot: [racket/master] remove uppercasing of the tag names - Robby Findler
09:26 RacketCommitBot: [racket/master] a bunch of fixes to the diagram: - Robby Findler
09:26 (part) RacketCommitBot
09:27 (quit) bitonic: Remote host closed the connection
09:29 (join) bitonic
09:33 (quit) hash_table: Ping timeout: 246 seconds
09:33 (quit) getpwnam: Ping timeout: 246 seconds
09:34 (join) adu
09:47 (quit) Shviller: Disconnected by services
09:47 (nick) Shvillr -> Shviller
09:47 (join) Shvillr
09:48 asumu: Shambles_: if you want to make a new widget and make it look native, I think the only "trick" is that you need three separate FFI bindings to build the widget on each of the three platforms.
09:48 Shambles_: asumu: Why would FFI use be necessary?
09:49 asumu: Shambles_: If you want to use a native widget that's not currently available, yes. An alternative is to build a widget using a canvas and racket/draw.
09:50 Shambles_: asumu: I more had the building it in Racket thing in mind.  There seem to be a lot of "you can't do that's" involved though.  Even basic things like figuring out what the color scheme of the desktop is don't seem to be supported.
09:50 asumu: Well, doing that in a cross-platform way is hard.
09:51 Shambles_: asumu: It appears that Racket is using GTK to provide the vast majority of the GUI.  I'm sure it has some cross-platform way to do that, since people use custom widgets in it, and it's been around a very long time.
09:51 asumu: Racket only uses GTK on Linux.
09:52 asumu: (or rather, non-Win/Mac systems)
09:52 Shambles_: asumu: That's odd.  The 'requirements' section had GTK as the only potentially-GUI-providing library.  I guess in reality it's binding to win32 and whatever Mac's use.
09:53 asumu: And one of the ways that people write custom widgets in GTK is to use Cairo to draw a widget, which is pretty much like using racket/draw.
09:53 Shambles_: I wonder how much it would /really/ hurt to just wrap something like Qt and have it over with.  Apparently Chicken did that.
09:54 asumu: That was probably omitted since Mac/Windows users don't actually need to install anything to use their native GUIs, whereas Unix people may not have GTK installed.
09:54 asumu: Qt is pretty hard since it uses its own preprocessing step to build and it's C++.
09:55 Shambles_: Given that GTK can look like the native widgets I'm not really sure why it wasn't just wrapped in the first place.
09:55 (quit) jrslepak: Quit: This computer has gone to sleep
09:55 Shambles_: It has pretty much everything, even 'weird' stuff I'd like to have like a toolbox widget.
09:58 Shambles_: asumu: Any idea why?
10:03 (quit) karswell: Read error: Connection reset by peer
10:08 (join) karswell
10:11 (join) veer
10:20 (join) jeapostrophe
10:20 (quit) jeapostrophe: Changing host
10:20 (join) jeapostrophe
10:26 (join) anRch
10:34 (quit) mceier: Quit: leaving
10:38 (join) jrslepak
10:49 (join) getpwnam
10:50 (join) hash_table
10:57 samth: Shambles_: (a) it would require either shipping gtk or expecting windows/mac users to download it
10:58 samth: (b) gtk doesn't really seem like a native app on windows
10:58 samth: (c) the gtk mac port is not up to the completeness standards of the other two
11:00 (quit) s_p_a_c_e_d_o_u_: Remote host closed the connection
11:00 Shambles_: samth: Would (a) be a big deal given all the other libraries?  (b) seems possible but not too convincing.  GIMP doesn't look native but isn't confusing on Windows.  As for (c) I can't say since I don't own a Mac, but I would expect it to be more complete than the GUI library we have at the moment.  For example I'm certain it has spinners (spinboxes) and tooltips, along with more obscure things (e.g. toolbox, grid layout...).
11:00 (join) s_p_a_c_e_d_o_u_
11:01 samth: (a) would indeed be unfortunate -- increasing the download size is never a good thing
11:02 samth: (b) you're welcome not to place a premium on interface consistency, but it is a drawback
11:02 samth: (c) the problem is not completeness, but functionality
11:02 samth: plus racket/gui does support tooltips -- you can see them in drracket
11:02 Shambles_: I can't think of a single 'living' project whose download size doesn't grow over time.
11:03 Shambles_: samth: I've actually investigated the tooltips thing rather thoroughly.  DrRacket rolls its own version, as does MrEd Designer, since the toolkit itself doesn't support them.
11:03 samth: Shambles_: sure, but keeping the dl size down is a good thing
11:03 Shambles_: samth: There's actually quite a bit of basic functionality missing.  The hardest thing I've encountered to work around, though, is that you can't drag across pasteboards.
11:05 Shambles_: Which, as far as I can tell, is just a hard limit.  There's not even a way to 'break off' scrollbars from editors, since they're not provided separately.  Otherwise you could have one great big pasteboard and then plaster scrollbars and other stuff over top of it to make it look like separate widgets.
11:07 Shambles_: You could make your own widget library, which, in the long run, might be the best thing to do, but then you run up on the problem of what happens if the user isn't using a generic Windows Classic desktop theme.  Then part of your window will look purple (or whatever) with your custom widgets looking gray and white.
11:07 Shambles_: That problem could be fixed if there was some way to get basic information from the OS, like the colors of various things.
11:08 (join) gciolli
11:08 (quit) gciolli: Client Quit
11:11 (quit) hkBst: Quit: Konversation terminated!
11:19 samth: eli: first, a bunch of the typed benchmarks, such as: http://drdr.racket-lang.org/24867/collects/tests/racket/benchmarks/shootout/spellcheck.rkt , which was previously run with 'mzc'
11:20 rudybot: http://tinyurl.com/6mhv9n2
11:20 samth: or http://drdr.racket-lang.org/24865/collects/tests/racket/benchmarks/shootout/binarytrees-places.rkt which was run with a command line argument
11:20 rudybot: http://tinyurl.com/7jvv3jw
11:21 samth: also, these: http://drdr.racket-lang.org/24866/collects/tests/racket/benchmarks/shootout/typed/ used to have a 'responsible' prop set
11:21 rudybot: http://tinyurl.com/7st54ra
11:22 samth: eli: more generally, there are now 500+ failures in drdr
11:22 samth: which date to your props changes
11:22 samth: so i think that change should be either fixed or temporarily reverted
11:27 eli rolls eyes at sam's direction
11:27 eli: Did you look in the props file?
11:27 (quit) Fare: Ping timeout: 246 seconds
11:27 (join) mceier
11:29 (join) RacketCommitBot
11:29 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://git.io/JQn27A
11:29 RacketCommitBot: [racket/master] when moving the focus to the replace window, also select the text - Robby Findler
11:29 RacketCommitBot: [racket/master] Added @literal-algol{} and started an algol60 test suite (apologies - Robby Findler
11:29 (part) RacketCommitBot
11:31 eli: samth: Did you look in the props file?
11:32 samth: eli: no, not yet
11:32 samth: what would i see there?
11:32 eli: That the properties *are* there.
11:32 eli: And therefore the problem is in some other place, which I didn't find yet.
11:33 eli: (Something in the code maybe.)
11:33 eli: jeapostrophe: ping
11:36 eli: samth: And now I have a good guess about the total breakdown.
11:36 samth: eli: ok, great
11:37 eli: it would be nice if jeapostrophe would be interactive and confirm it...
11:38 jeapostrophe: hmm?
11:38 eli: jeapostrophe: Do you have some log of the drdr run?
11:39 eli: jeapostrophe: More specifically, is the build derived by a racket script of version 5.2.1 (or earlier)
11:39 (join) Fare
11:39 jeapostrophe: drdr runs on 5.1.1
11:40 jeapostrophe: and when the props file can't be loaded, everything uses the defaults (nobody and racket)
11:40 eli: jeapostrophe: OK, I'll put a conditional around the module+ which is probably what broke.
11:41 eli: jeapostrophe: Also, it would be good if you upgrade to 5.2.1, since I don't have an easy way to check 5.1.1...
11:41 jeapostrophe: i could update drdr's install too and re-run the pushes
11:42 eli: You mean re-run the prop change pushes?
11:42 eli: That would be nice, but I think that they were all fine up to the commit that added the `module+'.
11:42 jeapostrophe: ya
11:43 jeapostrophe: 5.2.1 will work? even with the module+?
11:44 eli: jeapostrophe: No, it would break also -- that's how I caught it.
11:44 jeapostrophe: ah, so it wouldn't matter if i re-ran the pushes
11:44 eli: I also caught a use of `string-join' with one argument, which is a recent change.
11:45 eli: Anyway, I'm suggestion 5.2.1 since that I can check, and will know to check from now when/if I change that code.
11:45 eli: Oh, you meant re-run them as is -- no that won't work.
11:45 eli: I'm comitting a change that puts the `module+' in only in versions bigger than 5.2.1
11:47 jeapostrophe: i just updated the install
11:49 (quit) anRch: Quit: anRch
11:51 (join) RacketCommitBot
11:51 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://git.io/CtDHGg
11:51 RacketCommitBot: [racket/master] No need to run this wrapper in drdr. - Eli Barzilay
11:51 RacketCommitBot: [racket/master] Conditional use of `module+'. - Eli Barzilay
11:51 (part) RacketCommitBot
11:51 asumu: eli: what do you think of linking this (http://www.ccs.neu.edu/home/matthias/Thoughts/Racket_is____.html) from the Racket home page. As in make "Racket is a programming language" a link that points there.
11:51 asumu: s/\./\?/
11:52 eli: jeapostrophe: I pushed a version that should work -- if you can use that version with older pushes it might be good to run it over the last few broken pushes.
11:53 eli: asumu: good point, just a minute.
11:54 (join) nilyaK
11:55 asumu: (or perhaps "for more, see [link]" or whatever you see fit)
11:55 eli: asumu: On a brief read, it looks fine to do that.
11:55 eli: I was worried that it might be draftish or quasi-private-ish or something.
11:56 asumu: Should I just make the change and push it then? Or would you rather do it?
11:58 (join) jonrafkind
11:58 (quit) jonrafkind: Changing host
11:58 (join) jonrafkind
12:00 jeapostrophe: eli: i can't use anything but the version of the props file for the push
12:00 eli: jeapostrophe: ok
12:01 eli: jeapostrophe: In that case, if you have a way to skip builds, then you should do that since they'll all fail similarly.
12:01 (quit) djcoin: Quit: WeeChat 0.3.2
12:01 eli: If you have no way, you can just do what I do with nightly builds: kill the process.
12:01 eli: asumu: I'll do it, I have some related changes.
12:02 eli: asumu: That's assuming that I'll remember it by the time I finish going over the daily pile of emails.
12:03 (quit) hash_table: Ping timeout: 246 seconds
12:03 (quit) getpwnam: Ping timeout: 246 seconds
12:03 jonrafkind: was the error about stepper/private fixed?
12:06 jonrafkind: oh drdr is going to fail
12:06 (join) mithos28
12:09 (join) RacketCommitBot
12:09 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/G-B1Fg
12:09 RacketCommitBot: [racket/master] oops add one more reference to new file - John Clements
12:09 (part) RacketCommitBot
12:09 (join) MayDaniel
12:11 jonrafkind: eli, I found out from the nightly ppa build :p
12:11 eli: jonrafkind: So this was the reason for the flood of messages?
12:12 jonrafkind: what flood
12:12 eli: jonrafkind: I got about 10 emails on builds
12:12 jonrafkind: from launchpad?
12:13 jonrafkind: i thought only I got those
12:13 eli: From: Launchpad Buildd System <noreply@launchpad.net>
12:13 jonrafkind: ok yea those
12:13 eli: [Build #3590647] lpia build of racket 5.3.0.11-20120620~karmic in	ubuntu karmic RELEASE (plt-racket-nightly PPA)
12:13 jonrafkind: yea every build for each version of ubuntu failed, on 2 different architectures
12:13 jonrafkind: or possibly 3 in some cases, i think lpia might be arm?
12:13 eli: You mean their buildd, right?
12:14 jonrafkind: uhm, lpia is an ubuntu thing yes
12:14 eli: semi joke about the apparent typo
12:14 jonrafkind: oh rofl
12:14 (part) untrusted: "ERC Version 5.3 (IRC client for Emacs)"
12:15 eli: (Which my guess is a common name for a "build daemon")
12:15 jonrafkind: ya
12:19 jonrafkind: eli, so how come you didn't send an email about the broken push?
12:20 (quit) BeLucid: Ping timeout: 252 seconds
12:23 (join) RacketCommitBot
12:23 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/eXHP0Q
12:23 RacketCommitBot: [racket/master] fix struct/c for the case that the selectors are not available at the use site. - Robby Findler
12:23 (part) RacketCommitBot
12:24 (quit) jrslepak: Quit: This computer has gone to sleep
12:24 eli: jonrafkind: I woke up not too long ago, and didn't finish my email scan yet.
12:24 jonrafkind: oh ok
12:33 ozzloy: i'd like to store stuff in a configuration file and read that in from code.  what's a good way to do that?
12:33 ozzloy: stuff like username, pw, server for connecting to a postgres db
12:39 jeapostrophe: ozzloy: i tend to use read and write, or just write the config file as a module that exports the config data as bindings
12:44 ozzloy: jeapostrophe, and then you require it in the code?
12:53 jeapostrophe: ya
12:55 FareWell: does racket have some notion of (unit?) signatures satisfied by several (unit?), and is that feature actually used to develop libraries?
12:55 jeapostrophe: yes
12:55 jeapostrophe: search for signature in the docs
12:56 ozzloy: jeapostrophe, thanks
12:56 mithos28: Are units still used that much? the only recent code that I have seen it used in is TR and that is just to allow mutual recursion across source files.
12:57 (join) BeLucid
12:58 FareWell: mithos28, what is TR? Typed Racket?
12:59 mithos28: FareWell: Yes
12:59 jeapostrophe: mithos28: they pre-date modules, so they were used mean modules were all that was needed
12:59 jeapostrophe: mithos28: now they are only used when necessary, which is rare
12:59 jeapostrophe: but when used, they are great and important
12:59 FareWell: are do swindle generic functions allow for cross-module specialization, and are they actually used as interface / signatures to develop libraries?
13:00 FareWell: I'm trying to develop / standardize some generic datastructure libraries for CL, and I'm trying to find out what other languages have done.
13:00 (join) getpwnam
13:01 mithos28: jeapostrophe: it looked like you were answering yes to the actually used to develop libraries, where most uses of them are now in maintenence mode.
13:01 (join) hash_table
13:01 mithos28: Farewell: Look at generics in the prerelease version of racket
13:01 eli: FareWell: I need to run soon, but the sort answer is that you should avoid them for good code, which is different if you're coming from a CL world where that's considered good enough.
13:01 mithos28: you can find the docs at pre.racket-lang.org/docs/html
13:02 eli: mithos28: These two kinds of generics are *very* different.
13:02 FareWell: I intend to use my infamous "interface-passing style" for that - i.e. a lower-level approach to what units do, but one that allows for generic functions in the style of swindle.
13:02 mithos28: eli: Ah, they sound similar
13:04 eli: FareWell: I don't know what that is, but you *should* look at the other generics library that mithos28 mentioned,
13:05 eli: it's somewhat closer to haskell type classes, so it might be relevant, but it's not doing the super flexible thing that CLOS does (which is a feature, and ultimately the reason that I recommend not using them).
13:08 (join) tewk
13:12 (quit) veer: Remote host closed the connection
13:15 (quit) dented42: Quit: Computer has gone to sleep.
13:21 (join) dspt
13:37 (join) anRch
13:48 (join) dented42
13:52 (join) gciolli
13:59 (quit) gciolli: Ping timeout: 260 seconds
14:01 FareWell: it looks like a fork of racket. Is there documentation?
14:02 FareWell: https://github.com/takikawa/racket/tree/generics
14:04 asumu: FareWell: you should ignore that repo. It was just me brainstorming.
14:05 asumu: The latest version of Racket has a generics library that eli originally wrote that is useful.
14:05 asumu: (by latest I mean the pre-release version, not the latest release)
14:06 FareWell: URL?
14:07 FareWell: or should I just checkout the git tree?
14:12 (quit) nilyaK: Quit: Leaving.
14:24 asumu: FareWell: you can find docs for it here: http://pre.racket-lang.org/docs/html/reference/struct-generics.html
14:25 asumu: To use it, you'll need to check out the git tree or use a nightly build.
14:28 (join) yoklov
14:34 (quit) anRch: Quit: anRch
14:44 (quit) hash_table: Ping timeout: 248 seconds
14:44 (quit) getpwnam: Ping timeout: 248 seconds
14:45 FareWell: asumu: thanks!
14:45 asumu: NP, also I actually looked at that link and I thought it was something else. That was our development branch, but it's already been merged into core Racket.
14:48 (quit) ivan\: Quit: ERC Version 5.3 (IRC client for Emacs)
14:52 (join) antithesis
15:05 (join) ivan\
15:06 (quit) dented42: Quit: Computer has gone to sleep.
15:22 (join) soegaard
15:24 soegaard: asumu: Thanks for trying out "Populate "compiled" directories".
15:24 soegaard: asumu: Does testing.rkt work for you if you choose "syntactic test suite coverage" ?
15:27 soegaard: Another file that doesn't work for me with "population" on is bracket-lang.rkt
15:31 (join) gciolli
15:31 FareWell: asumu: do racket generics allow for any kind of method combinations?
15:31 FareWell: and/or inheritance?
15:43 asumu: soegaard: with syntactic test suite coverage, it breaks.
15:43 asumu: FareWell: what do you mean by method combinations?
15:43 stamourv: FareWell: They can't do method combinations yet.
15:44 stamourv: But, it would be great if it did.
15:44 asumu: It doesn't really have any inheritance. You could do it manually by having one set of generic functions just dispatch to another set.
15:44 stamourv: As for inheritance, Racket structs do provide inheritance.
15:44 stamourv: And I think that they would inherit generic functions. I'
15:44 stamourv: d need to double-check, though.
15:45 FareWell: method combinations: :around :after :before methods, or progn append cons + method combinations, or programmable method combinations.
15:45 FareWell: see CLOS.
15:45 soegaard: asumu: Hmm.
15:45 asumu: Ah, nevermind what I said. Structs will inherit in the sense that stamourv means.
15:45 FareWell: is it single-inheritance?
15:45 asumu: Yes.
15:46 FareWell: also, your method definitions are all in the defstruct form - can there be mutual recursion at the module scope?
15:46 stamourv: FareWell: Racket's generic functions are not as flexible as CLOS's, but they should cover a lot of their use cases, while being less scary and more robust.
15:47 stamourv: FareWell: Yes.
15:49 (quit) gciolli: Quit: Leaving.
15:49 soegaard: asumu: Is bracket/bracket-lang.rkt  breaking too?
15:50 soegaard: If so, then the problem is not the related to the custom language, but something else.
15:50 stamourv: soegaard: I really like the idea of a CAS specifically for high-school math.
15:50 (quit) kanak: Ping timeout: 240 seconds
15:51 stamourv: I remember using CASes in HS, and the advanced math / advanced features would always pop up and confuse me.
15:51 asumu: soegaard: with "syntactic test coverage" on, I get an error but a different one. I get a phase error from "../graphics/graphics.rkt"
15:51 stamourv: So I really like the idea of "language levels" for a CAS.
15:51 soegaard: stamourv: Thanks. I have been looking at some other CASes, and they more or less have a complete lisp at the core...
15:52 soegaard: asumu: Okay, I'll go look for references to graphics.
15:52 soegaard: Language levels... Nice idea. One with reals and one without!
15:53 stamourv: soegaard: Mark Weaver, a Maxima developer, has been working on numeric in Guile. He sometimes hangs out in #scheme.
15:53 soegaard: Errr. I mean one with reals and one with complex numbers...
15:53 stamourv: s/numeric/numerics/
15:53 soegaard: Numeric, is that the GNU spreadsheet?
15:56 (join) SrPx
15:56 stamourv: That's Gnumeric. I was talking about numerics, in the sense of the implementation of numbers in Guile.
15:57 stamourv: Hence my "s/numeric/numerics/" above.
15:57 stamourv: But I think he mentioned something about writing a CAS in Scheme/Guile. I don't remember.
15:58 stamourv: Actually, the guy is on #scheme right now: mark_weaver.
16:01 soegaard: Thanks for the tip. On GitHub I found a project by dharmatech implementing a small CAS in R6RS.
16:07 soegaard: asumu: The tip about the graphics module turned out to be quite useful. Removing the graphics module worked testing.rkt now works for me with populated directories. Now I only need to figure out why the breakage occurred in the first place.
16:13 (join) genbattle
16:16 asumu: Great. :)
16:18 (quit) FareWell: Quit: Leaving
16:23 (join) rbarraud
16:26 (join) ynniv
16:27 soegaard: The plot thickens. If I comment out everything inside the graphics module, I get the error (which refers to the expression module) and if I comment out the graphcis module, the error disappears. And the graphics module isn't used any where...
16:36 (join) lamefun3
16:37 lamefun3: is it possible to make Racket sandboxed just like JavaScript?
16:38 asumu: rudybot: doc racket/sandbox
16:38 rudybot: asumu: your sandbox is ready
16:38 rudybot: asumu: not found in any library's documentation: racket/sandbox
16:38 asumu: lamefun3: see http://docs.racket-lang.org/reference/Sandboxed_Evaluation.html
16:38 stamourv: lamefun3: rudybot here is a great example of a Racket sandbox.
16:39 stamourv: Each user has his own sandbox, and you're restricted in what you can do.
16:39 asumu: rudybot: ((lambda (x) (x)) (lambda (x) (x)))
16:39 rudybot: asumu: error: #<procedure>: expects 1 argument, given 0
16:39 stamourv: rudybot: source
16:39 rudybot: stamourv: git clone git://github.com/offby1/rudybot.git
16:39 stamourv: rudybot: (open-input-file "/etc/passwd")
16:39 rudybot: stamourv: your sandbox is ready
16:39 rudybot: stamourv: error: open-input-file: `read' access denied for /etc/passwd
16:40 stamourv: rudybot: (for ([i (in-naturals)]) i)
16:40 rudybot: stamourv: error: with-limit: out of time
16:40 asumu: rudybot: ((lambda (x) (x x)) (lambda (x) (x x)))
16:40 stamourv: lamefun3: ^
16:40 rudybot: asumu: error: with-limit: out of time
16:40 asumu: There we go.
16:49 m4burns: rudybot: (require ffi/unsafe)
16:49 rudybot: m4burns: your sandbox is ready
16:49 rudybot: m4burns: Done.
16:50 m4burns: rudybot: (ffi-lib 'libc)
16:50 rudybot: m4burns: error: #<undefined>:1:1: compile: access disallowed by code inspector to protected variable from module: "/mnt/racket-5.2.900.1/collects/ffi/unsafe.rkt" at: ffi-lib in: get-ffi-lib
16:50 m4burns: cool
16:51 soegaard: Interesting. If I move the graphics module upwards in the same file, the error moves from the expression module to the bracket module.
16:54 (part) lamefun3
17:11 (join) getpwnam
17:12 (join) hash_table
17:12 (nick) Shvillr -> Shviller
17:12 (join) Shvillr
17:17 (quit) mceier: Quit: leaving
17:19 bartbes: soegaard: again sounds like memory issues
17:19 (quit) ssbr_: Ping timeout: 260 seconds
17:20 soegaard: bartbes: Maybe. The behaviour is consistent though.
17:25 bartbes: out-of-memory issues, specifically
17:41 eli: Fare: The generics library is something that I did a long time ago, and recently asumu and stamourv made a proper library out of it,
17:42 eli: then, they went over places that used struct properties and made them use these generics too.
17:42 eli: So to find uses of that, it's probably easiest to just grab the current repository and grep for uses of `racket/generic'.
17:50 Fare: ok
17:51 sethalves: the google tells me there's no built-in way to make a file executable (other than ffi).  still true?
17:51 Fare: eli: these days, do you use emacs or DrRacket for editing?
17:52 eli: Fare: Emacs.  Well too invested in it.
17:52 (join) tonyg
17:52 tonyg: has anyone had issues before with bytes going mysteriously missing when reading from tcp sockets using read-bytes-evt?
17:53 tonyg: i'm starting to get worried
17:53 asumu: sethalves: what do you mean? As in build a Racket executable?
17:56 (quit) adu: Quit: adu
17:58 asumu: Or you mean like `chmod`?
17:58 bartbes: is the ffi easy to use, btw? since the "more examples" page in the docs seems to be missing
17:59 bartbes: (it refers to the non-existent ffi/examples)
17:59 asumu: It should exist. https://github.com/plt/racket/tree/master/collects/ffi/examples
17:59 asumu: I think the FFI is pretty easy to use.
17:59 asumu: Easier than other FFIs I've used anyway.
17:59 bartbes: ah, I expected it to be in the docs
18:00 asumu: Maybe that could be clearer.
18:03 (join) dented42
18:05 bartbes: hmm, I'd say the code looks overwhelmingly low-level
18:05 bartbes: but it's pretty easy, if you actually get down to reading it
18:06 bartbes: I guess type-massaging just does that
18:07 (quit) MayDaniel: Read error: Connection reset by peer
18:08 (quit) antithesis: Quit: antithesis
18:09 asumu: Depends on the library you're binding to, I think.
18:10 bartbes: of course, but transitioning between different types is always 'painful'
18:10 bartbes: ever seen luajit's ffi? it's pretty damned great
18:11 bartbes: that said, it won't make the functions make sense from a lua standpoint, so you'll still need to wrap it, but it's relatively painless
18:11 (quit) jeapostrophe: Ping timeout: 264 seconds
18:11 bartbes: basically, you drop a header into a function (cdef), and it generates a binding for you
18:12 asumu: Does it marshal/unmarshal values to/from C?
18:12 asumu: *marshall
18:12 bartbes: I never really tried structs, but those should be fine
18:12 bartbes: otherwise it automagically picks the right types and transparently converts
18:15 asumu: That's neat. I bet you could write a macro that did that as a high-level API over the Racket FFI.
18:16 asumu: (i.e., the macro would parse some C headers you write)
18:22 bartbes: of course you'll still have to do pointers, but you apparently can even assign metamethods to c types
18:22 bartbes: where metamethods allow you to overload operators, one of them indexing
18:23 bartbes: anyway, I might take a poke at racket's ffi soon-ish
18:28 (join) jimmypk
18:55 (join) ssbr_
19:11 (quit) masm: Quit: Leaving.
19:15 (join) RacketCommitBot
19:15 RacketCommitBot: [racket] plt pushed 7 new commits to master: http://git.io/hq79Dw
19:15 RacketCommitBot: [racket/master] another ASL error-message repair - Matthew Flatt
19:15 RacketCommitBot: [racket/master] error message repair - Matthew Flatt
19:15 RacketCommitBot: [racket/master] futures: make tail-call allocation local or synchronizable - Matthew Flatt
19:15 (part) RacketCommitBot
19:41 (join) gciolli
19:54 (quit) ssbr_: Ping timeout: 252 seconds
19:55 tonyg: yay
19:56 tonyg: i think i've found a bug in racket's I/O handling
19:56 (quit) jonrafkind: Ping timeout: 255 seconds
19:56 tonyg: i'm just packaging up my little test case for sending to the mailing list
20:05 (join) jeapostrophe
20:05 (quit) jeapostrophe: Changing host
20:05 (join) jeapostrophe
20:17 (quit) gciolli: Ping timeout: 250 seconds
20:25 (quit) rbarraud: Ping timeout: 245 seconds
20:34 (join) ssbr_
20:41 (quit) ynniv: Quit: ynniv
20:42 offby1: \o/
20:42 offby1: ~
20:42 offby1: / \
20:56 (quit) jesyspa: Quit: leaving
21:02 (join) rbarraud
21:02 (quit) ssbr_: Read error: Connection reset by peer
21:07 (join) ssbr_
21:11 (quit) kreol[Ukr]: Ping timeout: 246 seconds
21:31 (quit) yoklov: Quit: computer sleeping
21:39 (quit) dented42: Ping timeout: 244 seconds
21:42 sethalves: asumu -- i could shell out to /bin/chmod
21:42 sethalves: but i was looking for a way to affect the file from withing the interpreter
21:42 sethalves: (within)
21:42 Fare: grep the sources for chmod
21:43 sethalves: because... i will find something useful but undocumented?
21:45 Fare: maybe
21:46 asumu: sethalves: you can do that, one sec...
21:46 sethalves: file.util has file-perm, but that looks read-only
21:46 asumu: rudybot: doc file-or-directory-permissions
21:46 rudybot: asumu: http://docs.racket-lang.org/reference/Filesystem.html#(def._((quote._~23~25kernel)._file-or-directory-permissions))
21:46 asumu: sethalves: ^
21:47 asumu: (I would demonstrate with rudybot, but the sandbox probably won't let me touch any files)
21:47 sethalves: ah, thanks
21:48 (quit) jeapostrophe: Ping timeout: 248 seconds
21:49 sethalves: okay wait... this is another way to look at the permission bits?
21:53 sethalves: fare -- i see some chmod'ing happening in copy file
21:53 sethalves: but nothing else looks promising
21:55 sethalves: If an integer is supplied as the second argument, its is used as an encoding of properties (mostly permissions) to install for the file.
21:56 (join) adu
22:03 sethalves: file-or-directory-permissions is what I was looking for.  thanks asumu
22:03 asumu: NP
22:07 (join) yoklov
22:12 jimmypk: Shambles_: It was a namespace issue. DrRacket was linking to another library with a function with the same name as the one that was causing my program to crash.
22:29 (join) jonrafkind
22:35 (quit) ChanServ: *.net *.split
22:50 (join) ChanServ
22:52 (join) vu3rdd
22:52 (quit) vu3rdd: Changing host
22:52 (join) vu3rdd
22:53 (join) RacketCommitBot
22:53 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://git.io/SN7Lvg
22:53 RacketCommitBot: [racket/master] Racketize the `version' collection. - Eli Barzilay
22:53 RacketCommitBot: [racket/master] Dump attempt to be cute with stderr. - Eli Barzilay
22:53 (part) RacketCommitBot
23:16 (join) jrslepak
23:26 Shambles_: jimmypk: Now that, is amazing.  I would have never thought you'd get that behavior.  Thanks for letting me know.
23:30 jimmypk: Ya, me either. Thanks again for helping!
23:30 (quit) jimmypk: Quit: leaving
23:32 (quit) SrPx: Ping timeout: 245 seconds
23:46 mithos28: does anyone know why make-contract is needed, and make-chaperone-contract is not enough?
23:49 (quit) ivan\: Ping timeout: 245 seconds
23:50 (join) ivan\
23:57 (join) dented42