00:13 (quit) jeapostrophe: Quit: jeapostrophe
00:35 (quit) alexsuraci: Ping timeout: 276 seconds
00:36 (join) alexsuraci
00:48 (join) jonrafkind
01:33 (join) vu3rdd
01:36 (quit) jonrafkind: Ping timeout: 276 seconds
02:24 (join) rbarraud
02:58 (quit) rbarraud: Remote host closed the connection
02:58 (join) rbarraud
03:28 (quit) jao: Ping timeout: 255 seconds
03:32 (quit) rbarraud: Excess Flood
03:32 (join) rbarraud
04:06 (quit) rbarraud: Remote host closed the connection
04:08 (join) rbarraud
04:12 (quit) rbarraud: Remote host closed the connection
04:12 (join) rbarraud
05:22 (join) hanDerPeder
05:22 (quit) hanDerPeder: Client Quit
05:23 (quit) shofetim: Ping timeout: 240 seconds
05:46 (join) hanDerPeder
06:06 (quit) rbarraud: Ping timeout: 240 seconds
07:47 (join) jeapostrophe
08:58 (quit) vu3rdd: Remote host closed the connection
09:01 (quit) fjl: Remote host closed the connection
09:03 (quit) jeapostrophe: Quit: jeapostrophe
09:28 (join) shofetim
09:34 (quit) shofetim: Ping timeout: 252 seconds
09:41 (join) shofetim
09:50 (join) tonyg
10:02 (quit) tonyg: Ping timeout: 272 seconds
10:45 (join) b-man_
10:50 (join) jonrafkind
10:51 (quit) stamourv: Remote host closed the connection
10:55 (join) stamourv
11:15 (quit) hanDerPeder: Quit: hanDerPeder
11:28 (quit) jmbr: Remote host closed the connection
11:41 (join) tonyg
11:42 (join) jao
11:42 jonrafkind: why isn't there an r6rs language option in the language dialog box?
11:43 samth: jonrafkind, b/c r6rs can be specified in the source
11:43 samth: the r6rs provides a standard way of doing that
11:43 samth: #!r6rs
11:43 jonrafkind: so what
11:43 jonrafkind: why cant it also just be inserted by drracket
11:44 samth: it can - just change the default #lang line to #!r6rs
11:44 jonrafkind: r6rs mode could just insert that text and drop you into "get language from the # line"
11:44 jonrafkind: people dont know to use #!r6rs
11:44 jonrafkind: nor do I think they should have to
11:44 samth: we don't have a #lang typed/racket mode, or a #lang racket/base mode, or anything like that
11:44 jonrafkind: i dont see why not actually
11:45 jonrafkind: we might as well have all these options in the langauge dialog box
11:45 samth: b/c the language dialog should eventually go away
11:45 jonrafkind: and replaced with what?
11:45 samth: rather than being a gui way of inserting a single line of text
11:45 samth: replaced with #lang
11:45 jonrafkind: it gives you a list of options to choose from, if you just use #lang then you have to already know what the language is (and how to write it syntactically correct)
11:46 jonrafkind: seems to me the point o fusing a gui is to be user friendly
11:46 jonrafkind: you seem to be suggesting that because they provide the same functionality we can do without one or the other
11:46 samth: no, the #lang version provides much more functionality
11:47 samth: b/c it's extensible
11:47 samth: the general idea is that things should be specified in the program, not in the tool
11:48 jonrafkind: i think that places a burden on newbie programs
11:48 jonrafkind: and there is such a trivial way to help out those people its mind boggling why we dont help them
11:59 (join) sstrickl
12:08 samth: tewk, did you make compiling already-compiled collections faster when you added parallel build?
12:10 (join) hanDerPeder
12:30 tonyg: Why is "any" not permitted in the domain of a function contract? (Putting (lambda (x) #t) works ok)
12:30 sstrickl: any is a special marker
12:31 sstrickl: basically, it ensures that the contract does not break tail calls by not checking _anything_ about the return type
12:31 sstrickl: not even the number of values
12:31 tonyg: interesting, ok. Still: why not permit it in domain position?
12:33 sstrickl: there's unconstrained-domain-> if you're interested in a function contract that doesn't do any domain checking, but as for why (-> any some-ctc) doesn't turn into the former, I'm not sure.  Probably just the usual imbalance between multiple arguments and return values.
12:33 sstrickl: I'd have to ask Robby if there was a good technical reason or not, but I haven't thought of one in a few moments' thought.
12:34 tonyg: ok. Thanks
12:35 clklein: tonyg: Note that there is an `any/c' which I assume is roughly \x.#t
12:35 tonyg: I'm still reading the contract section of the guide
12:35 tonyg: clklein: ah, ok. I'll keep reading to see if the asymmetry is explained further down :) I haven't even reached the bit describing multiple-values yet
12:37 tonyg: yep, any/c works fine in domain position
12:38 sstrickl: tonyg: Right, any/c is an actual contract value, so it'll work anywhere you want a contract.  "any" is just a special syntactic marker.
12:43 (quit) tonyg: Quit: leaving
12:43 (join) tonyg
12:53 (join) vu3rdd
13:34 (quit) b-man_: Ping timeout: 240 seconds
13:35 (join) fjl
13:35 (part) fjl
13:38 (join) pavelludiq
13:41 (quit) jao: Read error: Operation timed out
13:48 (join) dc_
13:49 (join) jao
14:04 (quit) jonrafkind: Read error: Connection reset by peer
14:05 (join) jonrafkind
14:11 (quit) dc_: Quit: Page closed
14:19 (quit) vu3rdd: Remote host closed the connection
15:00 (join) anRch
15:04 clklein: Is there special syntax for multi-line strings containing characters that need to be escaped?
15:05 chandler: Something like a heredoc?
15:06 samth: clklein, there are here strings
15:06 samth: but i usually just use at-exp
15:07 clklein: Ah, yes, #<< is what I was remembering
15:13 clklein: thanks!
15:38 (quit) anRch: Quit: anRch
15:55 (join) _ryanc_
16:01 jonrafkind: http://www.larcenists.org/benchmarksGenuineR6Linux.html -- i didnt know larceny was so fast
16:03 (quit) jonrafkind: Read error: Connection reset by peer
16:04 (join) jonrafkind
16:12 jonrafkind: clklein, wow a use of dynamic-wind.. amazing
16:13 clklein: jonrafkind: I suppose I should have used `with-handlers' but I didn't think too hard about it
16:14 clklein: err, that's not enough actually
16:14 jonrafkind: so automatically creating typed programs is hard for redex?
16:15 (join) InzerMex
16:15 clklein: Yea, because it just works from the grammar, which isn't nearly enough information
16:15 jonrafkind: oh i see
16:19 jonrafkind: have you tried these ssorts of things for the simply typed lambda calclus?
16:19 clklein: which sorts?
16:20 jonrafkind: i mean can you generate simply typed lambda calculus terms
16:20 jonrafkind: with redex-test orw hatever
16:23 clklein: Redex's testing forms never take typing constraints into account during term generation.
16:24 clklein: Just generate terms freely from the grammar and throw out the ones that don't end up typechecking
16:24 jonrafkind: well, thats probably a lot of terms
16:24 jonrafkind: also wont i have to generate some type constructs as well?
16:25 clklein: Yes, it's stupid but it's less stupid then it sounds. We've found bugs in ICFP papers this way.
16:26 jonrafkind: ok
16:26 clklein: You can also write code that takes the freely generated term and tries to fix some of the more common mistakes
16:27 clklein: For example, find the free variables replace them (or add bindings if your term has a store)
16:27 clklein: I don't understand the question about generating type constructs.
16:27 jonrafkind: (lambda x -> x) -- I might have to generate (lambda (x : A) -> x)
16:29 clklein: A random term drawn freely from the grammar is much less likely to type check when the syntax uses type annotations. Is that what you mean?
16:34 clklein: I would not spend much time trying to get useful results out of redex-check in your particular case.
16:34 clklein: As I said in email, the problem seems hard enough that you'll need to write your own term generator (which is a pain)
16:47 jonrafkind: i wonder if i can generate terms constructively, that is, take some term that type checks and add more terms to it (at random) until one of those type checks
16:47 (quit) InzerMex: Ping timeout: 252 seconds
16:53 clklein: jonrafkind: Yes, that might work well. Randoop, a test generation tool for Java, uses that kind of bottom-up construction.
16:53 (quit) sstrickl: Quit: sstrickl
16:53 (join) sstrickl
16:55 clklein: jonrafkind: I need to run, but I'll be back later.
16:55 (quit) alexsuraci: Ping timeout: 272 seconds
16:55 jonrafkind: ok
17:01 (join) alexsuraci
17:27 (quit) sstrickl: Quit: sstrickl
17:29 (quit) martinhex: Read error: Connection reset by peer
17:30 (join) martinhex
17:45 (quit) pavelludiq: Remote host closed the connection
18:22 (join) rbarraud
18:37 (join) evhan
18:41 (join) rbarraud_
18:46 (quit) rbarraud: Ping timeout: 272 seconds
18:48 (part) shofetim: "ERC Version 5.3 (IRC client for Emacs)"
18:50 (nick) samth -> samth_away
18:56 (quit) rbarraud_: Remote host closed the connection
18:57 (join) rbarraud_
19:04 (quit) _ryanc_: Quit: Leaving
19:16 (join) jeapostrophe
20:16 (quit) jeapostrophe: Quit: jeapostrophe
20:21 (quit) hanDerPeder: Quit: hanDerPeder
20:22 (join) jeapostrophe
20:29 (quit) evhan: Read error: Connection reset by peer
20:41 (quit) jeapostrophe: Quit: jeapostrophe
20:44 (quit) rbarraud_: Remote host closed the connection
20:44 (join) rbarraud
20:48 (join) jeapostrophe
21:40 (quit) jonrafkind: Ping timeout: 245 seconds
21:44 (quit) jeapostrophe: Quit: jeapostrophe
21:51 (quit) jao: Ping timeout: 252 seconds
22:09 (join) rbarraud_
22:10 (quit) rbarraud: Ping timeout: 265 seconds
22:14 (join) jao
22:16 (quit) rbarraud_: Remote host closed the connection
22:16 (join) rbarraud_
22:23 (quit) rbarraud_: Remote host closed the connection
22:23 (join) rbarraud_
22:28 (join) jcowan
22:28 (part) jcowan: "Leaving"