00:04 (quit) _p4bl0: Read error: Connection reset by peer
00:04 (join) cipher-
00:04 (join) _p4bl0`
00:04 (quit) cipher: Ping timeout: 252 seconds
00:04 (quit) jamessan: Ping timeout: 252 seconds
00:04 (join) jamessan`
00:04 (quit) jamessan`: Changing host
00:04 (join) jamessan`
00:19 danking: Is there a #lang name for the student langauges? Opening DrRacket to try something is a pain
00:20 mithos28: #lang htdp/bsl
00:20 danking: mithos28: Excellent.
00:20 mithos28: I think that works, it has minor differences with the language level
00:29 (quit) haruki_zaemon: Ping timeout: 240 seconds
00:29 (join) haruki_zaemon
00:52 mithos28: If I have two pattern variables both bound with ellipses is there an easy way to generate the sequence of them alternating
00:53 mithos28: say (a …) = (a1 a2) and (b …) = (b1 b2). I want something that => (a1 b1 a2 b2)
00:54 mithos28: ((a b) …) gets me ((a1 b1) (a2 b2)) but that has an extra level of parentheses
00:57 jonrafkind: i guess you could just zip them together
00:57 jonrafkind: (with-syntax ([((both ...) ...) (zip as bs)]) #'(both ...))
00:58 asumu: Need a flatten too?
00:58 jonrafkind: the nested destructuring matcher should be an effective flatten
00:58 asumu: Ah, right.
00:58 jonrafkind: or you can just use flatten, either way
00:58 jonrafkind: #'(both ... ...)
00:59 mithos28: that is my current solution, it just seems like it should be easier
01:01 jonrafkind: the alternative, and my dream, is to have a better set of tools in templates instead of just ellipses, possibly to bind ellipses to a function that can do arbitrary computation, and then you can bind other ellipses types things
01:02 jonrafkind: (define-my-ellipses-thing zip... blah) #'((as bs) zip...)
01:02 jonrafkind: but of course you would have to reimplement 'syntax' to do so, of course I do that in honu
01:03 mithos28: that would be cool
01:06 mithos28: what is the racket style on make-noun vs noun as the constructor for a data type?
01:06 jonrafkind: it seems to have shifted to just noun
01:07 asumu: It seems inconsistent though. make-parameter vs. class.
01:07 asumu: (okay class is a bad example)
01:09 mithos28: ok, I just see so much make- names from the core libraries, it is disconcerting making my constructors not have the make-
02:08 (quit) jonrafkind: Ping timeout: 240 seconds
02:09 (quit) realitygrill: Quit: realitygrill
02:10 (join) realitygrill
02:11 (join) wishbone_
02:13 (quit) wishbone4: Ping timeout: 256 seconds
02:13 (quit) MayDaniel: Read error: Connection reset by peer
02:52 (quit) mithos28: Quit: mithos28
03:16 Gertm: morning
03:20 (join) hkBst
03:25 (join) Blkt
03:29 Blkt: good morning everyone
03:37 (join) ahinki
03:42 (join) noelw
03:47 noelw: epw: if you're still around, i might be able to help
04:08 (nick) _p4bl0` -> _p4bl0
05:05 (quit) ahinki: Read error: Connection reset by peer
05:05 (join) ahinki_
05:05 (nick) ahinki_ -> ahinki
05:08 (join) ahinki_
05:11 (quit) ahinki: Ping timeout: 260 seconds
05:11 (nick) ahinki_ -> ahinki
05:32 (nick) jamessan` -> jamessan
05:42 (quit) noelw: Quit: noelw
05:43 (quit) Blkt: Remote host closed the connection
05:43 (join) Blkt
05:54 (part) haruki_zaemon: "Laterz"
06:11 (join) oiiii
06:20 (join) tim-brown
06:30 _p4bl0: is there any reason Racket's webserver enforce to have a "Content-Type" header in the HTTP response? Even when there is no content at all (HEADER or OPTIONS methods for instance)
06:30 (quit) cipher-: Read error: Operation timed out
06:37 (join) cipher
06:37 (nick) cipher -> Guest935
07:06 (quit) tim-brown: Quit: Leaving
07:12 (join) francisl
07:14 RacketCommitBot: [racket] plt pushed 7 new commits to release: http://git.io/MGZLhw
07:14 RacketCommitBot: [racket/release] fix cocoa `get-clipboard-string' to never return #f; fix docs - Matthew Flatt
07:14 RacketCommitBot: [racket/release] update Racket history for v5.2 - Matthew Flatt
07:14 RacketCommitBot: [racket/release] history for teachpacks updated; please propagate - Matthias Felleisen
07:33 (join) tim-brown
08:17 (quit) francisl: Quit: francisl
08:32 (join) noelw
08:44 (join) jeapostrophe
08:47 (quit) noelw: Quit: noelw
09:22 (quit) si14: Remote host closed the connection
09:23 (join) si14
09:23 (quit) si14: Remote host closed the connection
09:47 (join) noelw
10:23 (quit) asumu: Ping timeout: 240 seconds
10:24 (join) asumu
10:28 (join) epw
10:29 (nick) samth_away -> samth
10:30 samth: _p4bl0: probably it would be easy to fix
10:30 samth: make a patch, submit a pull request
10:34 (part) Gertm: "WeeChat 0.3.5"
10:47 _p4bl0: samth: oh, actually it seems it's already been taken care of (commit e9a9d7949030faab94d0223a6ebebe3ad9e8b602)
10:49 _p4bl0: I guess it will be in 5.2
10:49 _p4bl0: good :-)
10:50 _p4bl0: I'll use #"text/plain; charset=utf-8" instead of #f as request-mime in the mean time
10:50 (quit) realitygrill: Quit: realitygrill
11:08 (quit) ahinki: Quit: ChatZilla 0.9.87 [Firefox 8.0/20111011182523]
11:17 (join) dnolen
11:25 (quit) jeapostrophe: Quit: jeapostrophe
11:36 danking: My favorite error message: "expected the identifier `TOKENS' at: TOKENS"
11:37 noelw: Error messages regarding generative structures are fun
11:37 noelw: Expected value of type foo, but got value of type foo
11:37 noelw: or similar
11:37 noelw: Hilarity ensues
11:41 (join) jonrafkind
11:44 (join) mithos28
11:51 (join) jeapostrophe
11:52 (join) anRch
11:59 (nick) wishbone_ -> wishbone4
12:03 (join) davegurnell
12:05 tim-brown worries increasingly about noelw's sense of humour
12:07 davegurnell: you're not the only one :)
12:11 noelw: Alright haters, I'm out of here
12:14 (join) dingfeng
12:19 (quit) oiiii: Remote host closed the connection
12:22 (join) mye
12:22 dingfeng: what's a good way to wrap a data structure to give it it's own type? using a struct with 1 field?
12:22 mithos28: Thats what I usually use
12:23 (quit) Blkt: Remote host closed the connection
12:24 dingfeng: and unwrap it every time it's used?
12:24 mithos28: yep
12:24 mithos28: what is the point of making it its own type?
12:25 dingfeng: hmm.... what if i just want to know that the data structure is true for some predicate
12:25 dingfeng: like, (< (length x) 10). i see things like pairof types, positive-bytes.
12:26 mithos28: In typed racket, or at the contract level?
12:26 dingfeng: typed racket
12:26 mithos28: You can use refinement types
12:27 mithos28: not sure how well they work, they are experimental
12:27 mithos28: but the basic idea is you get a subtype for values which pass the predicate
12:27 dingfeng: oh
12:27 dingfeng: a subtype?
12:27 mithos28: and then you pass the proof that it passes the predicate throughout the type system
12:28 dingfeng: not a value that satisfies two types?
12:28 mithos28: Well integer is a subtype of number
12:28 mithos28: and an integer value is a member of both types
12:28 dingfeng: like the number 12, could it be a type of integer, a type of number
12:29 mithos28: yep
12:29 dingfeng: and a type of even number, and a type of divisible by 3 number?
12:29 mithos28: yes
12:30 dingfeng: ok!
12:30 dingfeng: it's int he guide?
12:30 mithos28: multiple partially overlapping refinement types might not work all that well though
12:30 mithos28: rudy
12:30 mithos28: rudy-bot: help define-refinement-type
12:30 mithos28: rudybot: doc define-refinement-type
12:30 rudybot: mithos28: your sandbox is ready
12:30 rudybot: mithos28: not found in any library's documentation: define-refinement-type
12:31 mithos28: rudybot: (require typed/racket)
12:31 rudybot: mithos28: Done.
12:31 mithos28: rudybot: doc define-refinement-type
12:31 rudybot: mithos28: not found in any library's documentation: define-refinement-type
12:31 dingfeng: was just curious. i don't have an actual use-case for multiple partially overlapping refinement types. only for refinement types, yes.
12:31 mithos28: it should be, I cannot remember the name
12:31 (quit) hkBst: Remote host closed the connection
12:32 mithos28: rudybot: doc declare-refinement
12:32 rudybot: mithos28: http://docs.racket-lang.org/ts-reference/Experimental_Features.html#(form._((lib._typed-scheme%2Fbase-env%2Fprims..rkt)._declare-refinement))
12:32 dingfeng: subtyping?
12:32 (join) polann
12:32 dingfeng: mm
12:32 dingfeng: but if it's experimental, how were the positive bytes and integers done?
12:33 mithos28: union types
12:33 mithos28: the integer type is actually a union of other types
12:34 dingfeng: oh!
12:36 mithos28: and all of those don't partially overlap
12:36 dingfeng: this is cool. i've never used refinements in any other language at work.
12:37 mithos28: java sub typing with instance of casts is of similar behavior
12:37 mithos28: but only for the instance of predicate instead of arbitrary ones
12:40 dingfeng: which leads to a huge difference in practice
12:40 dingfeng: what about the type of list?
12:40 dingfeng: (Any -> Boolean : (Listof Any))
12:40 dingfeng: is that a refinement type?
12:41 mithos28: no, that is a predicate type
12:41 mithos28: it says that if the function returns true, then the argument has the type (Listof Any)
12:41 mithos28: i've got to go, I'll be on later if you have more questions
12:41 (quit) mithos28: Quit: mithos28
12:41 dingfeng: alright. thanks!
12:44 (join) haruki_zaemon
12:55 (quit) anRch: Quit: anRch
12:55 (quit) jeapostrophe: Quit: jeapostrophe
12:59 (part) tim-brown: "Leaving"
13:10 (join) mithos28
13:12 (join) MayDaniel
13:25 (quit) noelw: Quit: noelw
13:33 danking: How come define-binary-check expects to be passed two identifiers for actual and expected? http://docs.racket-lang.org/rackunit/api.html?q=rackunit#(form._((lib._rackunit/main..rkt)._define-binary-check))
13:36 (quit) davegurnell: Quit: davegurnell
13:39 (quit) MayDaniel: Read error: Connection reset by peer
13:41 asumu: danking: Looks like it binds them in the expression.
13:42 (quit) dnolen: Quit: dnolen
13:43 (quit) jonrafkind: Ping timeout: 244 seconds
13:58 (join) MayDaniel
14:02 (join) DanBurton
14:03 (join) jonrafkind
14:07 (join) dnolen
14:19 (quit) dnolen: Remote host closed the connection
14:48 dingfeng: how do i call untyped functions in typed racket?
14:49 mithos28: you cannot, you can import untyped functions as typed functions using require/typed
14:49 mithos28: rudybot: doc require/typed
14:49 rudybot: mithos28: http://docs.racket-lang.org/ts-reference/special-forms.html#(form._((lib._typed-scheme%2Fbase-env%2Fprims..rkt)._require%2Ftyped))
14:51 dingfeng: thanks (:
14:53 (quit) polann: Quit: Lost terminal
15:03 (join) shofetim
15:11 (join) mceier
15:14 DanBurton: Can Typed Racket infer the type of a top level function?
15:14 (quit) Nanakhiel: Read error: Connection reset by peer
15:14 (join) Nanakhiel
15:15 mithos28: Typed Racket does local inference, so in practice you need to annotate every function
15:17 DanBurton: *could* it infer the type of top level functions?
15:17 DanBurton: or is there some type weirdness that would prevent that?
15:19 dingfeng: *could* it infer the type of top level functions where it is possible?
15:19 mithos28: I think the reason is exponential time and not necessarily useful types instead of decidability
15:20 mithos28: So maybe, samth or stamourv would be able to better answer this if they are around
15:22 dingfeng: typed racket compiles in linear time?
15:22 dingfeng: it takes around 6s to compile my code
15:23 mithos28: I think some things are quadratic, but the ones I know about are bugs
15:23 (join) jeapostrophe
15:23 mithos28: do you use raco make?
15:23 mithos28: I have never had one file take 6s to compile
15:24 mithos28: actually maybe I have, 6s isn't that long
15:29 dingfeng: i just press ctrl+T in drracket
15:29 dingfeng: my typed racket code consistently compiles at 6s for a single file of around 2 pages
15:30 dingfeng: that's on an i5 mobile processor, windows
15:30 mithos28: Check if you have the option to save compiled files in the change language menu
15:30 mithos28: I mispoke about never seeing 6s, that is not that long
15:31 mithos28: also I assume you are on a released version of racket
15:31 mithos28: there were speedups that I think are still only in the development version, and will be in the soon to be released version
15:32 (join) dingfeng_quek
15:34 dingfeng_quek: gg sleep. i'll try it another time =)
15:34 (quit) dingfeng_quek: Client Quit
15:34 (quit) dingfeng: Ping timeout: 265 seconds
15:54 (join) EmmanuelOga
16:01 samth: mithos28: the subtyping algorithm is defintely super-linear
16:02 samth: also, the reasons not to do global inference are (a) that it's very hard in the presence of subtyping and (b) it produces terrible error messages
16:28 mithos28: samth: what part is super linear?
16:30 RacketCommitBot: [racket] plt pushed 3 new commits to master: http://git.io/xylO5A
16:30 RacketCommitBot: [racket/master] switch Mac OS X build default (32-bit vs 64-bit) to compiler default - Matthew Flatt
16:30 RacketCommitBot: [racket/master] CGC fix for OpenBSD x86_64 - Matthew Flatt
16:30 RacketCommitBot: [racket/master] fix JIT early reference to JIT generated address - Matthew Flatt
16:36 samth: mithos28: union subtyping, for example, is quadratic
16:37 mithos28: ok, is it linear in the number of expressions?
16:47 DanBurton: does #racket have a favorite pastebin?
16:49 mithos28: I use gist.github.com
16:52 samth: mithos28: the number of types is linear in the number of expressions in the worst case
16:52 DanBurton: Thoughts on this? https://gist.github.com/1302347
16:53 (quit) epw: Quit: Lost terminal
16:54 mithos28: What is the problem with it being too generic? I don't have racket installed on my current machine.
16:54 samth: DanBurton: try this:
16:54 samth: (: tree-sum-int ((Tree Integer) -> Integer))
16:54 samth: (define (tree-sum-int t)
16:54 samth: (tree-fold (inst id Integer) + t))
16:55 samth: mithos28, DanBurton: the problem is applying a polymorphic function to a polymorphic argument
16:55 samth: which Typed Racket doesn't allow
16:55 samth: which is why the explicit instantiation is required there
16:55 samth: DanBurton: note that the `plus' function isn't necessary in either version
16:56 samth: other than that, i like it :)
16:56 DanBurton: :)
17:00 (quit) shofetim: Read error: Connection reset by peer
17:00 DanBurton: Is there a reason Racket can't apply polymorphic functions to polymorphic arguments?
17:01 (quit) MayDaniel: Read error: Connection reset by peer
17:01 samth: DanBurton: it's hard to infer the instantiation
17:02 samth: DanBurton: hopefully, we'll eventually be able to do it for simple cases like the one you have
17:04 DanBurton: I guess Haskell has spoiled me when it comes to my expectations on a type system :P
17:04 samth: DanBurton: haskell makes different choices that make inference possible in some cases where it isn't in Typed Racket
17:06 DanBurton: I do like the ease of union types in Racket
17:06 DanBurton: TR, I mean
17:06 samth: yes, that's one of the big differences
17:10 DanBurton: variadic arguments look a lot easier, too
17:10 (part) haruki_zaemon: "Laterz"
17:11 (quit) karswell: Excess Flood
17:12 (join) karswell
17:23 (quit) DanBurton: Quit: ChatZilla 0.9.87 [Firefox 6.0/20110811165603]
17:35 (quit) jeapostrophe: Ping timeout: 255 seconds
17:39 danking: Does anyone have examples of using a syntax-class which accepts formal arguments? name:(stx-class arg1) doesn't seem to work as expected, nor does it look parsable in a sane way.
17:40 jonrafkind: (~var name (stx-class arg1))
17:41 stamourv: danking: There's an example in collects/typed-racket/optimizer/float.rkt 31:0
17:42 stamourv: Use site at 93:24.
17:42 stamourv: (Which is basically what jonrafkind said.)
17:43 danking: stamourv, jonrafkind: Thanks.
17:43 danking: stamourv: Syntax-classes definitely aren't higher order in the manner we were speaking, the arguments seem to be normal variables you can use in, for example, #:with clauses.
17:45 jonrafkind: you mean syntax-classes aren't first class
17:51 (quit) mceier: Quit: leaving
17:54 (join) jeapostrophe
18:07 (join) Fare
18:10 (join) francisl
18:45 (quit) jeapostrophe: Quit: jeapostrophe
19:12 (join) jeapostrophe
19:13 (quit) Fare: Quit: Leaving
19:15 (join) Fare
19:17 (quit) karswell: Excess Flood
19:17 (join) karswell
19:56 (quit) mithos28: Quit: mithos28
20:08 (nick) samth -> samth_away
20:09 (quit) Fare: Ping timeout: 240 seconds
20:10 (join) mithos28
20:18 (join) jrslepak_
20:21 (quit) jrslepak_: Client Quit
20:30 (join) spdegabrielle
20:30 (quit) jonrafkind: Read error: Operation timed out
20:36 (join) spdegabrielle_
20:36 (quit) spdegabrielle: Read error: Connection reset by peer
20:36 (nick) spdegabrielle_ -> spdegabrielle
20:40 (quit) spdegabrielle: Client Quit
20:45 (quit) mithos28: Quit: mithos28
20:57 (join) mithos28
21:03 (quit) acarrico: Remote host closed the connection
21:15 (join) jrslepak_
21:23 (quit) mithos28: Ping timeout: 240 seconds
21:41 (join) niarch
21:44 (quit) Nanakhiel: Ping timeout: 258 seconds
21:58 (join) anRch
22:02 (join) acarrico
22:05 (join) realitygrill
22:26 (join) anRch_
22:30 (quit) anRch: Ping timeout: 248 seconds
22:38 (quit) realitygrill: Ping timeout: 256 seconds
22:38 (join) realitygrill
22:43 (quit) anRch_: Read error: Connection timed out
22:44 (join) anRch
22:45 (quit) anRch: Client Quit
22:45 (join) jonrafkind
22:51 (quit) jonrafkind: Ping timeout: 253 seconds
22:52 (quit) francisl: Remote host closed the connection
22:52 (join) francisl
23:03 (join) jonrafkind
23:09 (join) dingfeng
23:24 dingfeng: hii
23:24 dingfeng: i get this issue with typed racket
23:24 dingfeng: after i do some predicate checks on a list, it becomes a
23:24 dingfeng: (mkdir   ((dir "${temp.config.dir}" . #<Typed Value: ()>) . #<Typed Value: ()>)   .   #<Typed Value: ()>)
23:24 dingfeng: it no longer cons to a (), and it's no longer considered a list!
23:42 (join) francisl_
23:46 (quit) francisl: Ping timeout: 248 seconds
23:46 (nick) francisl_ -> francisl