00:16 (quit) ivan\: Ping timeout: 265 seconds
00:17 asumu: mithos28: `make-contract` produces an impersonator contract, which is not the same as a chaperone contract.
00:17 asumu: rudybot: (impersonator-contract? (make-contract))
00:17 rudybot: asumu: ; Value: #t
00:17 mithos28: asumu: What is the use case?
00:20 mithos28: I know that contracts existed before impersonators and thus it might be a relic to support old bad contracts, but I cannot think of a contract that cannot be a chaperone
00:21 mithos28: except the parametric contracts, but I don't believe those are contracts as they break equality
00:21 (join) ivan\
00:29 asumu: Combinators that take impersonator contracts themselves have to be impersonator contracts too.
00:30 asumu: So in general, most combinators are built so they can be either.
00:30 asumu: (since you might get something like a parametric contract as an argument)
00:31 mithos28: But other than parametric contracts, what other base non-chaperone contracts are there?
00:32 asumu: I don't know of any off the top of my head.
00:32 asumu: Oh, object/c.
00:33 mithos28: why is that not a chaperone contract?
00:36 asumu: Since it uses class/c, which doesn't produce a chaperone, but a new class that has contracts built into the method tables.
00:38 mithos28: why cannot class/c produce a chaperone that adds it to the method tables, I am assuming it is just a field within the struct
00:40 mithos28: I can understand engineering effort as a reason, but that still means to me that impersonator contracts are not needed
00:40 (quit) soegaard: Ping timeout: 245 seconds
00:54 (quit) ssbr_: Ping timeout: 245 seconds
00:55 asumu: Yeah, it might be engineering effort and also efficiency, but not sure.
00:56 (quit) genbattle: Ping timeout: 255 seconds
01:03 (join) jeapostrophe
01:03 eli: samth: ping
01:07 asumu: mithos28: Contracts don't actually guarantee eq?-ness. See http://pre.racket-lang.org/docs/html/guide/contracts-gotchas.html
01:08 mithos28: right they don't guarantee eq?-ness but therefore the do guarantee equal-ness
01:12 asumu: Well, a promise is just checked with eq?, so equal? and eq? are the same in this case.
01:15 mithos28: what do you mean a promise is just checked with eq?
01:16 mithos28: If you use equal? you pass through impersonators and chaperones which is different than eq?
01:21 mithos28: The only way to extract a value from a promise is through the force procedure, which can be redirected through the use of a chaperone on the structure property prop:force
01:22 mithos28: thus promise/c could be implemented as a chaperone maintaining equality
01:27 (quit) jeapostrophe: Ping timeout: 245 seconds
01:30 (join) Kaylin
01:32 (join) kreol[Ukr]
01:45 (quit) hash_table: Ping timeout: 250 seconds
01:46 (quit) getpwnam: Ping timeout: 250 seconds
01:48 (join) hkBst
01:48 (quit) hkBst: Changing host
01:48 (join) hkBst
01:51 (join) veer
01:56 (quit) Kaylin: Quit: Leaving.
01:56 (quit) yoklov: Quit: computer sleeping
02:04 (join) Shviller
02:13 (quit) Fare: Quit: Leaving
02:21 (quit) dented42: Quit: Computer has gone to sleep.
02:29 (quit) veer: Quit: Leaving
02:45 (quit) jonrafkind: Ping timeout: 246 seconds
02:48 (quit) adu: Quit: adu
03:01 (join) djcoin
03:16 (quit) mithos28: Quit: mithos28
03:16 (join) untrusted
03:25 (join) antithesis
03:28 (quit) peeeep: Remote host closed the connection
03:28 (quit) sizz: Quit: No Ping reply in 180 seconds.
03:29 (join) sizz
03:29 (join) peeeep
03:33 (join) RacketCommitBot
03:33 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/BIQp8g
03:33 RacketCommitBot: [racket/master] rackety - John Clements
03:33 (part) RacketCommitBot
03:42 (quit) untrusted: Remote host closed the connection
03:43 (join) mceier
03:50 (join) ssbr_
04:07 (join) dalaing
04:24 (join) jesyspa
04:28 (join) untrusted
04:29 (join) gciolli
05:11 (quit) dalaing: Quit: Leaving.
05:11 (quit) gciolli: Quit: Leaving.
05:26 (join) tim-brown
05:45 (quit) mceier: Quit: leaving
05:47 (join) snearch
05:48 (join) dalaing
05:57 (join) gciolli
05:58 (quit) gciolli: Client Quit
06:01 (quit) noam: Read error: Connection reset by peer
06:02 (join) noam
06:05 (quit) ssbr_: Ping timeout: 246 seconds
06:17 (part) dalaing
06:28 (join) ssbr_
06:50 (quit) noam: Read error: Connection reset by peer
06:51 (join) noam
06:55 (quit) noam: Read error: Connection reset by peer
06:56 (join) noam
06:58 (quit) noam: Read error: Connection reset by peer
06:59 (join) noam
06:59 (quit) noam: Read error: Connection reset by peer
06:59 (join) noam
07:00 (quit) noam: Read error: Connection reset by peer
07:01 (join) noam
07:07 tim-brown: just seen a message on racket-users from neil van dyke which mentions the named-let vs. internal-define controversy. what is this controversy? how passionate do feelings run? would i be considered a troll just for mentioning it?
07:10 Cryovat: I've been curious about that myself
07:11 Cryovat: I feel using let is a bit more proper from a FP point of view
07:17 (quit) noam: Read error: Connection reset by peer
07:18 (join) noam
07:18 tim-brown: hence the controversy, i guess -- there must be people who would disagree (?)
07:20 tim-brown: JFGI seems not to unearth much
07:21 (quit) cdidd: Remote host closed the connection
07:24 (join) veer
07:28 Cryovat: The only issue I see with lets is that when you have many items, it looks a bit ugly
07:28 Cryovat: But even with the lexical scope, defines feel like what my Lisp teacher would have described as Java-isms
07:32 tim-brown: you want your program to be one single let* (or even let)
07:32 (quit) noam: Read error: Connection reset by peer
07:32 (join) noam
07:33 (quit) rbarraud: Ping timeout: 240 seconds
07:40 (quit) noam: Read error: Connection reset by peer
07:40 (join) noam
07:45 bremner: Cryovat:  I think some people prefer to omit let early in the learning process, since in principle you can get by with only define.
07:47 (quit) Shviller: Disconnected by services
07:47 (nick) Shvillr -> Shviller
07:47 (join) Shvillr
07:52 (quit) antithesis: Remote host closed the connection
07:55 (quit) noam: Read error: Connection reset by peer
07:55 (join) noam
07:56 tim-brown: bremner: is this a difference in teaching approach, rather than coding approach?
07:56 bremner: I guess so.
07:57 bremner: fwiw, I'm not one of these people.
07:57 tim-brown: bremner: you reserve define for toplevel, then?
07:57 bremner: but others have _much_ more experience teaching racket than I do.
07:57 bremner: tim-brown: yes
07:58 tim-brown: let has a nice habit of making lexical scope obvious
07:58 tim-brown: (if you're going to have a language with too many parenthesis, you might as well get value from them)
07:58 tim-brown: :-)
08:08 (join) kanak
08:10 (quit) vu3rdd: Ping timeout: 272 seconds
08:17 (join) jeapostrophe
08:17 (quit) jeapostrophe: Changing host
08:17 (join) jeapostrophe
08:38 samth: tim-brown: bremner is incorrect here
08:38 samth: the preference for internal define is for "professional" programs
08:38 samth: and it's to avoid nesting and rightward drift
08:38 samth: plus to avoid having a myriad of different local binding forms
08:39 samth: you just use the different variants of `define`
08:39 tim-brown: of which there are myriad?
08:39 samth: but some people like `let`, and/or dislike the longer name of `define`
08:39 samth: tim-brown: you already need all the define forms, roughly
08:40 samth: such a `define-values`, and `match-define`, and `define-struct`
08:40 tim-brown: is rightward drift not a trigger for refactoring your code?
08:40 samth: tons of rightward drift is
08:40 samth: but if you use let and let* and let-values you can get a lot of drift on just a few lines
08:41 tim-brown: i don't get much rightward drift with let*
08:41 tim-brown: maybe i'm not indenting "correctly"
08:41 samth: no, it's not that much
08:41 samth: but it adds up
08:42 samth: but regardless, that's the recommended style, although not everyone (such a neil) likes it
08:42 samth: *asd
08:42 samth: *as
08:42 tim-brown: what "that's"... nested-let or inner-define?
08:43 samth: "that" = internal define
08:44 Cryovat: samth: I started wondering
08:44 Cryovat: In terms of performance, is there any difference to let vs define?
08:44 samth: Cryovat: no
08:45 tim-brown: doesn't the compiler/vm boil it all down to lets anyway?
08:45 samth: of course, if you use the recursive binding features of define, then potentially good
08:45 samth: s/good//
08:46 samth: tim-brown: it boils down to let or letrec
08:46 samth: and letrec can be more expensive
08:46 tim-brown: than recursive binding features of define?
08:48 samth: no, than let
08:50 tim-brown: but surely that same expense exists when you compare define with to define without recursive bindings
08:52 samth: yes
08:52 samth: although i believe there may be some cases where you pay the expense of letrec for define even when it isn't needed
08:52 samth: b/c the compiler can't prove that
08:52 samth: those cases are quite rare, though
08:53 tim-brown: i guess the programmer wouldn't type the extra ...rec suffix unless they hadn't done that proof manually
08:54 samth: tim-brown: you can just write the let
08:55 samth: and you get the 'proof' automatically
08:55 samth: because of the different scoping semantics
08:55 tim-brown: are you saying I don't need letrec, cos let will do recursive bindings?
08:57 bremner: samth: I won't argue with you about your own coding standards ;). Iiirc, PLAI uses internal defines as well.
08:58 samth: tim-brown: no, i'm saying that letrec and let have different semantics, and so transforming one into the other in the compiler requires a (conservative) analysis
08:58 samth: and that `define` has (mostly) the semantics of letrec, and so getting it to have the performance of let again requires analysis
08:58 samth: bremner: in the teaching languages, we use `local`, which is a combo of let and internal-define
08:59 bremner: ah, right. That is what confused me.
08:59 tim-brown: the compiler only transforms letrec to let, though? (if it allows itself)
08:59 bremner: I never understood the point of local ;)
09:05 (join) getpwnam
09:06 (join) hash_table
09:13 (join) MayDaniel
09:19 (join) anRch
09:28 (quit) jeapostrophe: Ping timeout: 240 seconds
09:33 (quit) hash_table: Ping timeout: 248 seconds
09:33 (quit) getpwnam: Ping timeout: 248 seconds
09:38 samth: tim-brown: i don't understand your question
09:38 samth: bremner: local re-uses the syntax of define, but it's less complicated semantically than internal define (although i think that may no longer be true)
09:39 tim-brown: samth: i was asking for confirmation that the compiler will not try to transform let into letrec, only letrec into lec
09:39 tim-brown: s/lec/let
09:40 samth: tim-brown: right
09:40 samth: the other way would be silly
09:41 (join) yoklov
09:42 tim-brown: it's worth asking the question - i sometimes find that what i assume to be "silly" behaviour can actually be quite cunning
10:07 (join) gciolli
10:14 (join) mceier
10:18 (quit) noam: Read error: Connection reset by peer
10:18 (join) noam
10:35 (quit) hkBst: Quit: Konversation terminated!
10:42 (join) antithesis
10:50 (join) noam_
10:50 (quit) noam: Read error: Connection reset by peer
10:52 (quit) anRch: Quit: anRch
10:53 (join) getpwnam
10:54 (join) hash_table
10:57 (join) nilyaK
10:59 (quit) maahes: Ping timeout: 246 seconds
11:00 (join) jeapostrophe
11:00 (quit) jeapostrophe: Changing host
11:00 (join) jeapostrophe
11:02 (join) maahes
11:14 (join) diginet
11:14 diginet: hello
11:15 diginet: does racket have any bindings to SDL? and/or any libraries for game development?
11:16 (join) graspee
11:19 danking: diginet: Probably not what you had in mind, but I've made some pretty sweet games with World and Universe: http://world.cs.brown.edu/1/htdw-v1.pdf
11:19 (quit) MayDaniel: Read error: Connection reset by peer
11:20 (quit) yoklov: Quit: computer sleeping
11:20 asumu: diginet: I think jonrafkind is working on allegro 5 bindings https://github.com/kazzmir/racket-allegro-5
11:21 asumu: (and more libraries on PLaneT: http://planet.racket-lang.org/)
11:23 (quit) gciolli: Ping timeout: 276 seconds
11:23 danking: http://www.ccs.neu.edu/home/matthias/HtDP2e/prologue.html One of the examples under the third bullet under terminology sees od. Why does (cos pi) evaluate to #i#i#i...#i-1.0 ?
11:24 (quit) noam_: Read error: Connection reset by peer
11:24 (join) noam_
11:27 (join) mithos28
11:28 Shambles_: diginet: Racket comes with double-buffered canvases stock.  Unfortunately it comes with absolutely no sound support at all.  You'll need the PortAudio and RSound packages off PLaneT, most likely, to make a game.  You might also want to get the RacketGL package if you want to do 3-D graphics or hardware accelerated 2-D graphics, since the sgl bindings Racket comes with are for a very old version of OpenGL.
11:44 (join) vu3rdd
11:44 (quit) vu3rdd: Changing host
11:44 (join) vu3rdd
11:48 (quit) noam_: Read error: Connection reset by peer
11:48 (join) noam_
11:49 (quit) hash_table: Ping timeout: 265 seconds
11:49 (quit) getpwnam: Ping timeout: 265 seconds
11:49 (join) yoklov
11:58 (part) tim-brown: "WeeChat 0.3.8"
12:01 (join) jonrafkind
12:01 (quit) jonrafkind: Changing host
12:01 (join) jonrafkind
12:02 (join) dented42
12:09 (quit) djcoin: Quit: WeeChat 0.3.2
12:25 (quit) BeLucid: Ping timeout: 265 seconds
12:26 (quit) jeapostrophe: Ping timeout: 245 seconds
12:29 (join) MayDaniel
12:33 (join) jeapostrophe
12:33 (quit) jeapostrophe: Changing host
12:33 (join) jeapostrophe
12:41 (join) aidy
12:49 aidy: Is there a way to disable the R5RS' language option 'disallow redefinition of initial bindings' with a #lang directive instead of using drracket?
12:54 (join) wedesoft
12:56 mithos28: aidy: you want to be able to redefine bindings using #lang r5rs? the negations are confusing.
12:56 (join) dyoo
12:57 aidy: mithos28: correct
12:57 mithos28: Because that is not allowed, http://docs.racket-lang.org/r5rs/r5rs-mod.html?q=r5rs
12:57 (quit) bitonic: Ping timeout: 252 seconds
12:57 mithos28: racket prohibts that in the module context, only at the toplevel does it allow it
12:58 (quit) vu3rdd: Remote host closed the connection
12:58 aidy: I know, but when you use the language dialoct in drracket you can set it to allow redefinitions
12:58 (part) wedesoft
12:59 mithos28: I'm not sure what you are trying to do, but I think to get strict conformance to r5rs you need to use load/eval with a namespace made by (namespace-require/copy 'r5rs)
12:59 mithos28: that is because the language dialog is broken, and is being replaced by modules
12:59 mithos28: It is likely doing the toplevel eval of each form
13:00 dyoo: aidy: offhand, I'm not sure how to make redefinition work in a module.  Is there anything in particular you're trying to do with it?
13:00 (part) untrusted: "ERC Version 5.3 (IRC client for Emacs)"
13:03 (join) getpwnam
13:04 (join) hash_table
13:07 aidy: I'm trying to run SICP code
13:08 jonrafkind: there should probably  be #lang sicp
13:10 (quit) dyoo: Ping timeout: 245 seconds
13:14 (join) gridaphobe
13:15 jonrafkind: how do I use an alternate cls file with scribble?
13:18 (join) BeLucid
13:19 (quit) veer: Quit: Leaving
13:39 samth: jonrafkind: roughly, you re-implement scribble/sigplan for your new cls file
13:43 jonrafkind: has anyone done this for sig-alternate.cls?
13:43 samth: jonrafkind: i don't think so
13:43 samth: what conference is that?
13:44 jonrafkind: gpce, with sig-alternate.cls
13:54 stamourv: aidy, jonrafkind: neilv has a PLaneT package for sicp.
13:54 stamourv: planet.racket-lang.org/display.ss?package=sicp.plt&owner=neil
13:54 samth: jonrafkind: if it's pretty similar to sigplan, then it won't be too hard
13:55 (quit) antithesis: Quit: antithesis
14:00 aidy: stamourv: (define apply-in-underlying-scheme apply) gives 'reference to an identifier before its definition identifier: apply
14:00 aidy: in neil/sicp
14:01 (quit) dented42: Quit: Computer has gone to sleep.
14:02 (join) bitonic
14:05 (quit) kreol[Ukr]: Quit: Leaving
14:06 (join) weirdo
14:06 weirdo: hey
14:06 weirdo: having a problem evaluating a r5rs file in nightly
14:07 weirdo: http://paste.lisp.org/display/130203
14:08 weirdo: am i doing something wrong?
14:14 weirdo: geiser is nigtly, emacs 24.1
14:15 weirdo: will switch to non-nightly. sad panda.
14:17 (quit) jeapostrophe: Ping timeout: 260 seconds
14:19 (join) jeapostrophe
14:19 (quit) jeapostrophe: Changing host
14:19 (join) jeapostrophe
14:25 weirdo: non-nightly racket works
14:37 (quit) getpwnam: Ping timeout: 248 seconds
14:37 (quit) hash_table: Ping timeout: 248 seconds
14:38 (join) getpwnam
14:38 (join) hash_table
14:43 (join) bro_grammer
14:44 (join) random_malice
14:44 (quit) hash_table: Ping timeout: 246 seconds
14:44 (quit) getpwnam: Ping timeout: 246 seconds
14:52 (quit) random_malice: Ping timeout: 246 seconds
14:52 (quit) snearch: Quit: Verlassend
14:52 (quit) bro_grammer: Ping timeout: 246 seconds
15:00 (quit) MayDaniel: Read error: Connection reset by peer
15:06 asumu: weirdo: can you reduce that to a small test case that fails/
15:06 asumu: *?
15:06 asumu: If it's a bug, a bug report would be appreciated.
15:06 weirdo: asumu, i can show a backtrace, if that helps
15:06 weirdo: oh, i already did
15:06 asumu: Well, what is seaking2/util? Is that code available somewhere?
15:07 weirdo: asumu, #lang r5rs some-random-defun
15:08 asumu: weirdo: some-random-defun being something like (define ...)?
15:08 weirdo: asumu, i'm sorry, but i'm largely unfamiliar with Racket or Geiser innards. i've just chosen this configuration for my r7rs implementation, that's al
15:08 (quit) jonrafkind: Ping timeout: 252 seconds
15:09 asumu: Well, if you want to use Racket the R5RS language is probably unideal.
15:09 weirdo: i know :-(
15:09 asumu: danking: that HtDP thing is weird. It prints correctly in my racket.
15:09 weirdo: but i need to bootstrap somehow. and other impls are far farther from ideal
15:09 asumu: Why not just bootstrap off of Racket?
15:10 asumu: Are you actually trying to extend #lang r5rs (as implemented in Racket) to R7?
15:10 asumu: You might want to look at how the #lang r6rs is implemented if so (for how to do the module system, for example).
15:10 weirdo: asumu, i'd have to spend time learning racket and r5rs differences. also there's a factor of using only standard stuff to ease the switch
15:11 weirdo: the target 'VM' is javascript
15:13 weirdo: another thing is that the only impls geiser supports are racket and guile
15:18 samth: weirdo: you might want to take a look at danny yoo's whalesong system
15:19 samth: which is an implementation of racket in JS
15:19 weirdo: samth, i'm going for r7rs though so it's an effort in itself. also, it's a learning project :-)
15:19 weirdo: not invented here :-)
15:24 stamourv: aidy: I haven't used that package, so I don't know.
15:25 stamourv: Does `(define (apply-in-underlying-scheme . x) (apply apply x))' work?
15:27 aidy: stamourv: it does actually, thanks :)
15:30 (join) anRch
15:34 (join) MayDaniel
15:38 stamourv: No problem.
15:40 (join) jonrafkind
15:48 (part) freakazoid: "Textual IRC Client: http://www.textualapp.com/"
15:53 (join) RacketCommitBot
15:53 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/bCd1Ng
15:53 RacketCommitBot: [racket/master] Added docs for pi.f, corrected pi's contract, added examples - Neil Toronto
15:53 (part) RacketCommitBot
15:56 (quit) ssbr_: Ping timeout: 245 seconds
15:58 (quit) gridaphobe: Remote host closed the connection
16:08 (join) RacketCommitBot
16:08 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://git.io/kCLOsQ
16:08 RacketCommitBot: [racket/master] mzlib/class100: deprecation notice - Asumu Takikawa
16:08 RacketCommitBot: [racket/master] racket/control: Racketize & remove trailing whitespace - Asumu Takikawa
16:08 (part) RacketCommitBot
16:13 (join) DanBurton
16:13 (join) untrusted
16:24 (join) genbattle
16:38 (quit) untrusted: Remote host closed the connection
16:39 (quit) anRch: Quit: anRch
16:40 (quit) MayDaniel: Read error: Connection reset by peer
16:44 (join) ynniv
17:02 (join) gciolli
17:08 (quit) nilyaK: Quit: Leaving.
17:17 (quit) jeapostrophe: Ping timeout: 260 seconds
17:17 (join) cdidd
17:17 (join) jeapostrophe
17:17 (quit) jeapostrophe: Changing host
17:17 (join) jeapostrophe
17:29 (join) Nisstyre
17:29 (quit) Myk267: Quit: Myk267
17:30 (quit) Nisstyre: Read error: Connection reset by peer
17:31 (join) getpwnam
17:32 (join) hash_table
17:37 (join) _
17:37 (nick) _ -> Shviller
17:39 (quit) Shvillr: Ping timeout: 255 seconds
17:39 (join) Shvillr
17:40 (join) Myk267
17:41 (quit) acarrico: Ping timeout: 265 seconds
17:45 (join) acarrico
17:46 (quit) graspee: Quit: leaving
17:48 (quit) noam_: Read error: Connection reset by peer
17:48 (quit) Myk267: Quit: Myk267
17:48 (join) noam_
17:53 (nick) samth -> samth_away
18:00 (quit) jeapostrophe: Ping timeout: 245 seconds
18:02 (quit) mceier: Quit: leaving
18:03 (join) Myk267
18:10 (quit) PfhorSlayer: 
18:16 (join) doomrobo
18:18 doomrobo: Why does this error in typed scheme? (define-type NN (Number -> Number)) (ann (lambda: ((x : Number)) (+ x 1)) NN)
18:18 doomrobo: gives me unbound type name
18:19 asumu: Works for me.
18:19 asumu: (in both 5.2.1 and in my git build)
18:20 asumu: doomrobo: you said "typed scheme". Are you using an old pre-Racket version?
18:21 doomrobo: no, I meant typed/racket
18:21 asumu: What version of Racket?
18:21 (join) gridaphobe
18:21 doomrobo: 5.2.1
18:21 doomrobo: ok, in the text box it works but not in the REPL
18:21 doomrobo: odd...
18:23 asumu: REPL behavior in TR is sometimes different from when you're in a module.
18:23 doomrobo: what do you mean?
18:24 asumu: If you write your program in a file with #lang typed/racket it should always work fine.
18:24 asumu: Unfortunately, some things are hard to get working in the REPL.
18:27 doomrobo: ah, ok
18:27 doomrobo: is it technically compiled when not in the REPL?
18:27 asumu: Racket always compiles your code to bytecode even when in the REPL.
18:28 asumu: There are some optimizations that don't happen in the REPL that can be done in modules, but that's a detail.
18:29 (join) adu
18:31 doomrobo: asumu, so what would be the problem in getting the code working in the REPL?
18:32 doomrobo: I mean, in Python for instance, there are no inconsistencies in running code between the REPL and the actual interpreter
18:40 stamourv: doomrobo: The big difference between a module and the REPL is that in a module, it's possible to see all the definitions at once. In the REPL, it's not.
18:40 stamourv: That means that a lot of things are going to be possible in a module that are not possible at the REPL.
18:41 asumu: Yes, in this regard Python is just less flexible.
18:41 stamourv: The REPL is fundamentally limited in what it can do.
18:41 stamourv: Python just propagates these limitations to the whole language.
18:42 doomrobo: ah, ok
18:42 (quit) Myk267: Quit: Myk267
18:42 stamourv: So in general, the REPL is more for quick experiments and exploration. Modules should be used for the rest.
18:43 stamourv: But since modules are pretty lightweigth, and can you can visit them easily at the REPL, you get the best of both worlds.
18:43 (join) Myk267
18:43 doomrobo: thanks
18:44 (quit) doomrobo: Quit: Leaving
18:44 (join) antithesis
18:45 (quit) acarrico: Ping timeout: 276 seconds
18:48 (join) acarrico
18:50 (quit) antithesis: Quit: antithesis
19:17 (quit) ynniv: Quit: ynniv
19:29 (quit) noam_: Read error: Connection reset by peer
19:30 (join) noam_
19:35 (quit) acarrico: Ping timeout: 265 seconds
19:45 (join) acarrico
19:50 (quit) yoklov: Quit: bye!
20:03 (join) dented42
20:20 (quit) bitonic: Remote host closed the connection
20:20 (join) bitonic
20:21 (quit) bitonic: Remote host closed the connection
20:24 (join) bitonic
20:25 (quit) bitonic: Remote host closed the connection
20:26 (quit) stchang: Read error: Operation timed out
20:28 (join) stchang
20:29 (join) bitonic
20:30 (quit) bitonic: Remote host closed the connection
20:30 (join) bitonic
20:32 (quit) diginet: Read error: Connection reset by peer
20:37 (join) jimmypk
20:44 (quit) jesyspa: Quit: leaving
20:45 (quit) jonrafkind: Ping timeout: 276 seconds
20:53 (quit) acarrico: Ping timeout: 246 seconds
21:05 (quit) DanBurton: Quit: leaving
21:06 (join) dnolen
21:09 (join) jeapostrophe
21:09 (quit) jeapostrophe: Changing host
21:09 (join) jeapostrophe
21:14 (join) acarrico
21:15 (join) Kaylin
21:28 (quit) dented42: Ping timeout: 265 seconds
21:28 (join) duomo
21:32 (quit) jeapostrophe: Ping timeout: 250 seconds
21:37 (quit) bitonic: Remote host closed the connection
21:45 (join) ynniv
21:46 (quit) ynniv: Client Quit
21:47 (quit) dnolen: Ping timeout: 244 seconds
22:07 (join) RacketCommitBot
22:07 RacketCommitBot: [racket] plt pushed 3 new commits to master: http://git.io/TTHNkw
22:07 RacketCommitBot: [racket/master] further error message convention refinements - Matthew Flatt
22:07 RacketCommitBot: [racket/master] ffi/com: allow a `com-object' as an 'iunknown argument - Matthew Flatt
22:07 RacketCommitBot: [racket/master] racket/port: fix `read-bytes-evt' and `read-string-evt' - Matthew Flatt
22:07 (part) RacketCommitBot
22:10 (quit) adu: Quit: adu
22:17 (join) mmalorni
22:24 (join) dented42
22:33 (quit) mmalorni: Remote host closed the connection
22:35 (join) mmalorni
22:36 (quit) mmalorni: Remote host closed the connection
22:37 (join) mmalorni
22:40 (quit) gridaphobe: Remote host closed the connection
22:43 (quit) mmalorni: 
22:44 (join) mmalorni
22:45 mithos28: does anyone know why hash/c on immutable hashes makes a copy versus making a chaperone of the hash?
22:49 (quit) dsantiago: Ping timeout: 248 seconds
22:53 (join) dsantiago
23:00 (quit) dsantiago: Ping timeout: 272 seconds
23:03 (quit) hash_table: Ping timeout: 272 seconds
23:03 (quit) getpwnam: Ping timeout: 272 seconds
23:04 (join) dsantiago
23:06 (join) jonrafkind
23:08 asumu: mithos28: immutable hashes aren't kept wrapped since the providing party can't observe any modifications (via hash-set) that the client party makes.
23:10 mithos28: asumu: That is a reason that chaperones are not required, but doesn't answer why copying is chosen over chaperoning.
23:10 asumu: Oh, I see. Maybe it's cheaper.
23:12 mithos28: It requires a copy on every application of the contract versus every access of a field,
23:12 asumu: Well, also the straightforward implementation with a chaperone would do more checking.
23:12 asumu: Since the chaperone retains the redirect functions.
23:12 asumu: So on hash-set, you'd get a new hash that still checks the contract.
23:12 mithos28: Ah, ok
23:19 (join) jeapostrophe
23:19 (quit) jeapostrophe: Changing host
23:19 (join) jeapostrophe
23:20 asumu: Hmm, getting `in-set` to work in TR might be hard.
23:20 (quit) mmalorni: Ping timeout: 276 seconds
23:21 asumu: (its expansion is safe but TR can't figure that out...)
23:21 mithos28: what is the problem with the expansion?
23:21 (join) doomrobo
23:22 doomrobo: I don't understand why the example for implementing Maybe from Haskell is so complicated at http://docs.racket-lang.org/ts-guide/types.html
23:22 doomrobo: This seems to work for me: (struct: None ())
23:22 doomrobo: (define-type (Maybe a) (U None a))
23:23 mithos28: doomrobo: What are the elements of the type (Maybe None)
23:23 doomrobo: nothing?
23:24 asumu: in-set does a trick where it checks if it's set?. When it's not, it uses `unless` to run an error case. Unfortunately, occurence typing can't figure out that if the body is executed, the code after the unless is unreachable.
23:25 mithos28: asumu: that just requires fixing the typing rule of begin, to add the correct filters to the following statements
23:25 mithos28: just might be a slight understatement
23:26 doomrobo: mithos28, is what I wrote incorrect?
23:26 asumu: Heh. Right, it's probably do-able. I was just hoping it'd be an easy hack.
23:26 doomrobo: and why do you need an element to the struct if you are just specifying a type?
23:26 mithos28: doomrobo: In haskell it would Be Nothing and Just Nothing
23:26 (quit) jeapostrophe: Ping timeout: 240 seconds
23:26 (join) mmalorni
23:27 doomrobo: mithos28, I don't know any Haskell
23:27 doomrobo: I just realize what Maybe is for
23:27 (join) dnolen
23:27 mithos28: Ah that would explain it, in racket you would define it how you did
23:27 mithos28: in haskell it is slightly different, but the example is true to the haskell origins
23:28 doomrobo: ah
23:28 doomrobo: ok, I wasn't too clear on that. Thanks
23:28 mithos28: Its the difference between true Unions (in racket) versus discriminated unions in other languages (Haskell)
23:29 mithos28: asumu: you could rewrite provide a slightly different in-set from dr racket that would do the right annotations
23:30 (quit) doomrobo: Quit: gn
23:31 (join) veer
23:45 (quit) mmalorni: Remote host closed the connection
23:51 (quit) dented42: Quit: Computer has gone to sleep.
23:53 (join) jacius