00:02 (join) dnolen
00:03 (quit) mceier: Quit: leaving
00:08 (quit) jeapostrophe: Ping timeout: 265 seconds
00:18 (quit) jao: Ping timeout: 246 seconds
00:25 (join) cdidd
00:27 (quit) hash_table: Ping timeout: 244 seconds
00:27 (quit) getpwnam: Ping timeout: 244 seconds
01:12 (quit) yoklov: Quit: computer sleeping
01:43 (quit) Kaylin: Quit: Leaving.
01:54 (quit) ski: Ping timeout: 246 seconds
01:56 (join) ski
02:10 (join) jonrafkind
02:18 (join) mceier
02:37 (join) djcoin
02:38 (quit) dnolen: Remote host closed the connection
02:51 (quit) jonrafkind: Ping timeout: 248 seconds
02:52 (join) hkBst
02:52 (quit) hkBst: Changing host
02:52 (join) hkBst
03:04 (quit) mceier: Quit: leaving
03:36 (quit) maahes: Ping timeout: 252 seconds
03:48 (quit) jacius: Remote host closed the connection
03:49 (join) maahes
03:57 (join) snearch
04:09 (quit) bjz: Quit: Leaving...
04:10 (join) bjz
04:29 (join) noelw
04:39 (quit) maahes: Read error: Operation timed out
04:40 (join) bitonic
04:52 (join) MightyFoo
04:53 (join) maahes
04:53 (nick) MightyFoo -> tim-brown
05:05 tim-brown: i discovered yesterday that duckduckgo has a "bang-racket" syntax e.g. "!racket let/cc"
05:07 (join) noelw_
05:08 tim-brown: i was interested by jonrafkind's futures browser, but i don't know how to get a hold of it. any pointers?
05:09 (quit) noelw: Ping timeout: 246 seconds
05:09 (nick) noelw_ -> noelw
05:46 (quit) rbarraud: Ping timeout: 248 seconds
06:10 (quit) mithos28: Read error: Connection reset by peer
06:11 (join) mithos28_
06:40 (quit) ski: Ping timeout: 240 seconds
06:40 (join) mceier
06:54 (join) ski
06:57 (quit) mithos28_: Quit: mithos28_
07:00 (quit) ivan\: Ping timeout: 250 seconds
07:03 (join) ivan\
07:05 tim-brown: with ffi, i have a function which returns a pointer which has been malloc'd using a GC_malloc (linked with my foreign library)
07:05 tim-brown: i need this pointer to remain visible within racket (i.e. on the stack or anywhere else that the CGC would find it)
07:06 tim-brown: this is fine for racketcgc (since GC_malloc is shared between my library and racket)
07:06 tim-brown: but how do i keep a pointer visible with racket3m?
07:07 tim-brown: [if possible, without "remembering" it explicitly in my library]
07:11 (quit) ski: Ping timeout: 240 seconds
07:25 (quit) antithesis: Remote host closed the connection
07:42 (quit) vu3rdd: Ping timeout: 246 seconds
07:43 (quit) mceier: *.net *.split
07:43 (quit) djcoin: *.net *.split
07:43 (quit) jrslepak: *.net *.split
07:43 (quit) sethalves: *.net *.split
07:43 (quit) mattmight: *.net *.split
07:43 (quit) ppilate: *.net *.split
07:43 (quit) hyko: *.net *.split
07:43 (quit) offby1: *.net *.split
07:43 (quit) dspt: *.net *.split
07:43 (quit) rotty: *.net *.split
07:43 (join) mceier
07:44 (join) djcoin
07:44 (join) jrslepak
07:44 (join) sethalves
07:44 (join) mattmight
07:44 (join) ppilate
07:44 (join) hyko
07:44 (join) offby1
07:44 (join) dspt
07:44 (join) rotty
07:50 (join) masm
07:54 (quit) snearch: Quit: Verlassend
08:12 (quit) Shviller: Ping timeout: 246 seconds
08:13 (join) Shviller
08:20 (quit) jyc: Read error: Connection reset by peer
08:26 (join) jeapostrophe
08:26 (quit) jeapostrophe: Changing host
08:26 (join) jeapostrophe
08:28 (join) Kaylin
08:34 (join) jeremyheiler
08:44 (nick) samth_away -> samth
08:48 (join) jtpercon
08:51 (join) yoklov
08:53 (join) getpwnam
08:54 (join) hash_table
09:04 (join) antithesis
09:06 (quit) jeapostrophe: Ping timeout: 248 seconds
09:13 (quit) yoklov: Quit: computer sleeping
09:19 (join) MayDaniel
09:22 (quit) hash_table: Ping timeout: 246 seconds
09:22 (quit) getpwnam: Ping timeout: 246 seconds
09:26 (quit) jeremyheiler: Quit: Computer has gone to sleep.
09:44 (join) jeremyheiler
09:45 (topic) -: Racket -- http://racket-lang.org -- logs at http://racket-lang.org/irc-logs/
09:45 (names) -: gabot jeremyheiler MayDaniel antithesis jtpercon Kaylin Shviller masm rotty dspt offby1 hyko ppilate mattmight sethalves jrslepak djcoin mceier ivan\ noelw maahes tim-brown bitonic bjz hkBst cdidd jrslepak_neu noam__ mungojelly acarrico Shambles_ sizz_ Nisstyre aidy Znudzon rudybot cataska stchang jimmypk Shvillr surrounder rapacity BeLucid SHODAN gf3 stamourv imami|afk basepi m4burns eli eMBee danl_ndi asumu mario-goulart DraX em ivan` elliottcable
09:45 (names) -: tauntaun SeanTAllen dsantiago samth Araq karswell errstr danking ozzloy igibson @ChanServ anonus s_p_a_c_e_d_o_u_ spanner tewk lusory otterdam bartbes antono cky bremner chandler friscosam Cryovat jamessan
09:45 (join) gciolli
09:52 (join) yoklov
09:54 (quit) yoklov: Client Quit
10:09 (join) soegaard
10:09 (join) soegaard_
10:09 (quit) soegaard: Read error: Connection reset by peer
10:09 (nick) soegaard_ -> soegaard
10:10 (join) jeapostrophe
10:12 (quit) mceier: Quit: leaving
10:17 (quit) jrslepak: Quit: This computer has gone to sleep
10:23 (quit) gciolli: Ping timeout: 250 seconds
10:31 (join) gciolli
10:32 (join) jrslepak
10:45 (join) RacketCommitBot
10:45 RacketCommitBot: [racket] plt pushed 7 new commits to master: http://git.io/_3XBxQ
10:45 RacketCommitBot: [racket/master] Make TR compile cleanly with contracts enabled. - Eric Dobson
10:45 RacketCommitBot: [racket/master] Fix bad contract and clashing provides in TR when contracts are enabled. - Eric Dobson
10:45 RacketCommitBot: [racket/master] Added cond-check-below - Eric Dobson
10:45 (part) RacketCommitBot
10:47 (join) kvda
10:54 (quit) bjz: Ping timeout: 250 seconds
10:56 (join) getpwnam
10:56 (join) anRch
10:56 (join) hash_table
10:57 (quit) hkBst: Quit: Konversation terminated!
11:02 (quit) jrslepak: Quit: This computer has gone to sleep
11:06 (quit) kvda: Quit:  x___x
11:06 (join) DanBurton
11:08 (join) mmajchrzak_
11:09 (quit) noelw: Quit: noelw
11:11 (quit) Znudzon: Ping timeout: 248 seconds
11:15 (quit) mmajchrzak_: Quit: Leaving
11:19 (join) jrslepak
11:19 (quit) jeapostrophe: Read error: Operation timed out
11:23 (quit) MayDaniel: Read error: Connection reset by peer
11:27 (join) MayDaniel
11:30 (quit) MayDaniel: Read error: Connection reset by peer
11:36 (quit) jeremyheiler: Quit: Computer has gone to sleep.
11:39 (join) jeremyheiler
11:39 (join) mceier
11:43 (join) mithos28
11:48 (join) kanak
11:53 (quit) acarrico: Ping timeout: 248 seconds
11:57 (join) jonrafkind
11:57 (quit) jonrafkind: Changing host
11:57 (join) jonrafkind
12:02 mithos28: samth: ping
12:02 samth: mithos28: pong
12:02 samth: thanks for the refactoring!
12:02 mithos28: No problem. I have some more coming
12:02 (join) RacketCommitBot
12:02 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/NrcHrw
12:02 RacketCommitBot: [racket/master] more submodule repairs for `raco exe' - Matthew Flatt
12:02 (part) RacketCommitBot
12:03 mithos28: I was wondering what check-below is supposed to return
12:03 samth: it should return a Result if one of the arguments is a result, otherwise it should return a Type
12:04 mithos28: Bad question. Let me actually phrase it better.
12:04 mithos28: Is it ok to return a subtype of the expected, instead of the expected type?
12:05 samth: mithos28: no, that's not ok
12:05 samth: at least in general
12:05 samth: consider the expression (ann x Type)
12:05 samth: check-below always returns `Type` here
12:06 samth: so that `ann` controls the type of the expression
12:06 samth: otherwise, it's just a constraint, and thus very frustrating to use
12:06 samth: (SML behaves in the other way, and it's a huge pain)
12:06 mithos28: Ok. I am trying to figure out how to do prototypes from Odersky's paper
12:07 mithos28: And one of the things that it would add would be that expected would no longer always be a valid type
12:07 samth: mithos28: what do you mean by 'prototypes'?
12:08 (quit) anRch: Quit: anRch
12:08 mithos28: I thought that was the terminology used in the paper, let me look it up
12:08 samth: oh, you mean the partial types propagated by inference
12:08 mithos28: yeah
12:08 mithos28: The ? 'type'
12:09 samth: i see two options here
12:09 (join) Znudzon
12:09 samth: 1. change the invariant on check-below so that it always returns an instantiation of the expected type
12:09 samth: 2. add a new version of check-below for inference
12:10 samth: i think probably 1 is better
12:10 mithos28: I agree.
12:10 mithos28: There are just lots of parts of the system that assume expected is a type and not something that needs to be instantiated first
12:11 samth: yes, i bet that's true
12:11 mithos28: is tc-expr the same as tc-expr/check with the type Top?
12:11 samth: mithos28: no
12:11 stamourv: tim-brown: The futures visualizer is available in the latest git. Maybe in the nightlies too, if they're up to date.
12:12 stamourv: tim-brown: And jamesswaine wrote it, not jonrafkind.
12:12 samth: consider what happens with letrec and inference when there's an expected type
12:12 jonrafkind: ill accept rewards in the form of money or food
12:14 (join) nilyaK
12:14 mithos28: samth: not obvious what you mean by that
12:14 samth: mithos28: sorry, that was opaque
12:14 samth: with letrec inference
12:15 samth: if you have (let loop args body), then it uses the expected type as the return type for `loop`
12:15 samth: but if there's no expected type, it just doesn't do it
12:16 samth: but if you introduce odersky's system, there's more opportunity to unify tc-expr and tc-expr/check
12:16 samth: b/c the former would be the latter with ? as the expected
12:17 mithos28: Right. Thanks for the example, I couldn't really see why odersky needed the ?.
12:18 samth: another place where you get ?
12:18 samth: would be:
12:18 samth: (identity arg)
12:18 samth: you'd have an expected ? for checking `arg`
12:19 mithos28: This is why I wanted check below to return a subtype
12:19 mithos28: because then it checked with Top, arg would still return its true value
12:19 samth: yeah
12:19 mithos28: *type
12:20 samth: i think ? is actually the right solution to this problem
12:20 samth: note that there's already something like ? in the system already
12:20 mithos28: What?
12:20 samth: the NoEffect constructor
12:20 (quit) jeremyheiler: Quit: Computer has gone to sleep.
12:21 samth: is for indicating that an annotated/parsed type doesn't have an annotated proposition
12:21 samth: meaning the underlying one should be passed through
12:21 mithos28: ok, but thats a filter not a type
12:21 samth: sorry, NoObject
12:21 samth: yes
12:21 mithos28: Ah, thats why I couldn't find it
12:22 mithos28: Also why does type? include things that aren't types?
12:22 samth: yeah, see https://github.com/plt/racket/blob/master/collects/typed-racket/typecheck/check-below.rkt#L52
12:22 rudybot: http://tinyurl.com/comceb2
12:23 samth: so that they can participate in definition of the type hierarchies
12:23 samth: `type?` could be redefined to be the same as `Type/c`
12:24 mithos28: Ok. As a first pass would only allowing ? at the top level of an expected, and replacing the no check code paths with that, be a good starting approach?
12:25 (quit) djcoin: Quit: WeeChat 0.3.2
12:25 samth: mithos28: yes, that sounds reasonable
12:25 mithos28: Ok, thanks. I've got to run, day job thing.
12:26 samth: mithos28: great!
12:27 DanBurton: samth: typed racket question for you http://pastebin.com/hnkgb5GM
12:28 DanBurton: it seems that and & or under certain circumstances aren't typed correctly
12:28 (join) MayDaniel_
12:29 samth: DanBurton: the first of those is because TR learns something about `x` with the `list?` test
12:30 samth: to work around that, just bind `x` to `y` before the test, and return `y`
12:30 samth: the second one is more surprising
12:30 DanBurton: regarding the first, shouldn't TR know that whatever it learns about x, its type should still be a subtype of its original type?
12:31 DanBurton: I suppose it learns not only about X, but about A in that case
12:31 DanBurton: or it should, rather
12:31 samth: the problem is that the knowledge can't be represented in TR's type system
12:31 DanBurton: aw
12:32 DanBurton: so yeah how 'bout the second
12:33 DanBurton: here's an example closer to the actual code I'm trying to type: http://pastebin.com/8xF50aGD
12:34 (join) jeapostrophe
12:35 samth: ok, checking it out
12:35 tim-brown: stamourv: apologies to jamesswaine - jonrafkind was raving about it last week; i made the assumtion that he'd done it
12:36 tim-brown: thereby making an ass out of me and umption
12:40 (quit) jeapostrophe: Read error: Operation timed out
12:41 tim-brown: night all -- will check the nightlies again (futures browser not in the ones i last looked at; but that was *days* ago)
12:42 (quit) tim-brown: Quit: night all
12:45 samth: DanBurton: your larger example doesn't work b/c it might produce #t if x is `null`
12:51 (join) jacius
12:53 DanBurton: oh right
12:54 DanBurton: it should be unioned with #t not #f for that particular example
12:57 samth: then it will probably typecheck
12:58 DanBurton: it still has the problem that it thinks it "got Boolean"
12:58 DanBurton: I mean sure I could tell it (U Boolean A)
12:58 samth: DanBurton: in my drracket, if I change the return type to (U #t #f A) it works
12:58 samth: or (U Boolean A)
12:59 DanBurton: right, that would make sense, given Boolean <: (U Boolean A)
13:03 (join) acarrico
13:04 DanBurton: I'm dealing with some code whose comments claim its type is "any (-> a) (exn -> a) -> a". I think I can just get away with making it "any (-> #t) (exn -> #f) -> Boolean"
13:11 (join) soegaard
13:15 DanBurton: samth: (U #t #f) should be indistinguishable from Boolean, right?
13:15 samth: DanBurton: yes
13:17 Araq: "Racket is a programming language. " now that's helpful ...
13:17 Araq: how about mentioning that it's a Scheme=
13:17 Araq: ?
13:17 DanBurton: samth: I tried both (U #t #f A) and (U Boolean A), but only the latter works. http://pastebin.com/UVTLyv2G
13:17 DanBurton: Araq: that should be pretty obvious to anyone that know what a Scheme is
13:18 DanBurton: *knows
13:18 samth: DanBurton: yes, that's a bug that's been fixed -- it'll be in the next release
13:18 Araq: well I had to browse a bit around to get it ... ;-)
13:18 DanBurton: samth: oh yay :)
13:18 Araq: I only noticed the syntax is lisp-like from the startpage
13:18 Araq: (and yes I do know what scheme is ...)
13:19 DanBurton: Racket used to be called PLT-Scheme, but then the PLT people decided that it was different enough that they should make a clean break in the name
13:20 (join) snearch
13:20 DanBurton: So even though it's probably fair to say that Racket is a Scheme, it's also somewhat fair to say that it is not.
13:20 Araq: DanBurton: now I know, but it's really missing on the main page
13:20 samth: Araq: we think that the impression you get from the code snippets on the start page is more important than the history
13:21 DanBurton: and to be fair "Racket is a programming language" should be pretty obvious to... well... anyone that finds themself at racket-lang.org so I don't really know why it even says that.
13:22 Araq: samth: I disagree; for me it was just "Lisp-like" with no obvious sign of being a scheme
13:22 (quit) gciolli: Ping timeout: 250 seconds
13:22 Araq: and there are lots of lisp dialects around
13:23 samth: Araq: what crucial information is the Scheme heritage conveying?
13:23 Araq: hygienic macros?
13:23 Araq: nicer names? (car, cdr)
13:24 Araq: oh I see you have those names too
13:24 DanBurton: ugh I'm dealing with some code that makes heavy use of those >,<
13:25 (join) RacketCommitBot
13:25 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/z-1-Cw
13:25 RacketCommitBot: [racket/master] Move racket/future/visualizer, racket/future/trace --> future-visualizer, future-visualizer/trace - James Swaine
13:25 (part) RacketCommitBot
13:25 DanBurton: does regular scheme have structs like racket does?
13:26 Araq: I don't think so, but I don't know really
13:26 Araq: it's said that scheme is much more minimalistic than common lisp
13:26 Araq: so my guess is it doesn't :-)
13:27 DanBurton: structs are *so* much better than this car/cadddadr nonsense xP
13:29 Araq: can racket be compiled to native machine code?
13:30 bitonic: DanBurton: you can easily build your own "struct" library
13:30 bitonic: with lists
13:31 (quit) soegaard: Quit: soegaard
13:31 stamourv: bitonic: That doesn't give you structs. They still answer #t to `list?'.
13:31 stamourv: Araq: Yes, Racket has a JIT compiler.
13:31 stamourv: (that compiles to native code)
13:32 bitonic: stamourv: OK, implement them with vectors then :P
13:32 (quit) MayDaniel_: 
13:32 bitonic: my point is that you don't *need* them in the language
13:33 stamourv: Well, you don't *need* numbers either, lambdas are enough.
13:33 stamourv: But numbers are nice to have, and so are structs.
13:33 DanBurton: church encoding for structs would probably be good enough to satisfy me :)
13:34 bitonic: stamourv: no. you need machine numbers much more than native structs.
13:34 stamourv: Well, neither is strictly necessary. The lambda calculus manages just fine.
13:35 stamourv: My point is that, unless you're playing some mathematical games, there's no reason to avoid useful features.
13:35 Araq: stamourv: and its own GC implementation?
13:35 stamourv: Araq: Yep, custom-build generational GC.
13:36 bitonic: stamourv: no no. machine numbers are strictly necessary to do anything useful with numbers. native structs aren't strictly necessary to do anything.
13:36 Araq: stamourv: precise stack marking?
13:36 stamourv: Araq: I think so, but I'd have to check to be sure.
13:37 Araq: does it transform everything to CPS?
13:37 stamourv: bitonic: http://en.wikipedia.org/wiki/Church_encoding
13:37 bitonic: stamourv: I know, thank you.
13:37 bitonic: but machine numbers are necessary to do anything useful. structs are basically just sugar if you have vector.
13:37 bitonic: *vectors
13:38 stamourv: That's still an encoding. Vectors can't give you information hiding, whereas structs can.
13:38 asumu: No, for the reasons that stamourv already outlined. Vectors answer #t to vector?. It's not an abstraction.
13:39 DanBurton: the downside of encoding structs by using some other container is that they will still respond to operations on that container, which is rather un-struct-like
13:39 bitonic: stamourv: sure, but the gap between church numerals and machine numbers is *much* wider than the one between the one between and "fake" structs and "navtive" structs
13:39 (quit) jonrafkind: Ping timeout: 255 seconds
13:39 stamourv: And structs give you disjoint datatypes. No one can confuse your new thing with a plain old vector, and mistakenly operate on it like a vector, breaking your invariants (e.g. that the head of the vector is the special tag that tells you what it is).
13:39 tewk: Araq: precise stack marking, YES
13:39 stamourv: bitonic: Agreed, but there's no point in avoiding using structs.
13:40 Araq: tewk: how does it handle multi threading? stop the world?
13:40 bitonic: stamourv: I think there is a point in simplicity. It's always going to be a compromise, but it's not that easy - plenty of useful languages lack structs
13:41 asumu: Structs are pretty simple in the hierarchy of complication of language features...
13:41 asumu: (even C has them!)
13:41 asumu: (sort of)
13:41 stamourv: I agree, simplicity is nice, but I don't think that getting rid of structs makes things much simpler, and getting rid of them is painful.
13:41 tewk: Racket threads are user threads or green threads.
13:42 tewk: Places provide true parallelism and each place has its own garbage collector.
13:42 bitonic: asumu: that's not true, people have huge fights about how to have structs/records in the Erlang and Haskell community, where structs are just sugar (but in different ways)
13:42 (part) antithesis
13:42 Araq: tewk: I see; but somebody mentioned a "GIL" in the VM?
13:43 asumu: Haskell already has records. They're just fighting over an improved API for them.
13:43 tewk: There is a master collector that has to stop the world, but it runs very infrequently.
13:43 Araq: tewk: how do you do message passing between places? deep copy?
13:44 bitonic: asumu: haskell records are just and only sugar.
13:44 tewk: Araq: yes, there isn't a GIL.
13:44 bitonic: asumu: they're just functions and patterns over normal datatypes
13:44 asumu: bitonic: Yes, because they have a general ADT mechanism which provides about as much abstraction.
13:44 Araq: tewk: if you do deep copying, why do you even need a master collector?
13:44 bitonic: asumu: not really. it provides no notion of the record's fields
13:45 bitonic: asumu: and no structural records either
13:45 asumu: Yes... so they want more features. So how is that an argument against them?
13:45 bitonic: asumu: it's an argument about the obviousness of records as a feature
13:45 bitonic: since Haskell and Erlang are 2 widely used languages that lack them
13:46 asumu: No, that just shows it's hard to retrofit a feature onto an old language that has its own long-lasting idioms.
13:46 asumu: There have been tons of proposals for how to do "records right" in Haskell. Some even implemented (e.g. TRex in Hugs).
13:46 bitonic: asumu: it's not like everybody in Haskell wants records. Ofc the people who want it make much more noise
13:46 tewk: Araq: the master collector collects place channels and shared atomic value vectors.
13:46 bitonic: I wouldn't even say it's the majority
13:47 tewk: But if you don't kepp allocating a lot of those the master collector quickly reaches a steady state and doesn't need to collect.
13:49 Araq: I see
13:50 Araq: are race conditions completely prevented with this scheme?
13:50 (join) tonyg
13:50 (join) MayDaniel
13:51 samth: Araq: you can always implement a higher-level structure via messages, and then have races there
13:51 (join) yoklov
13:51 samth: but in the conventional sense, places can't have races b/c they don't share any locations
13:54 Araq: samth: how do the "shared atomic value vectors" work?
13:54 Araq: in particular how do you do the ownership transfer of the memory?
13:54 samth: when you send a reference to a shared vector, the other side sees it as a vector
13:54 samth: but no copying takes place
13:55 Araq: but how do you keep a local GC from freeing this memory?
13:56 samth: Araq: shared vectors are allocated in the master heap only
13:56 samth: which is not touched by local GC
13:57 (join) jonrafkind
13:57 (quit) jonrafkind: Changing host
13:57 (join) jonrafkind
13:57 Araq: say if you have N places that need to fill the same shared vector
13:58 Araq: how does that work?
13:58 Araq: say we have a vector of strings
13:58 Araq: each string is allocated per place
14:01 Araq: how is each string put into the vector then?
14:01 Araq: by copy?
14:01 (quit) nilyaK: Quit: You can put whatever you want here.
14:01 tewk: you can only have shared vectors of bytes, fixed integers and float
14:01 Araq: aha alright
14:01 (join) nilyaK
14:01 Araq: hrm, sounds water proof :-)
14:02 DanBurton: Suppose I want to create a macro that generates several (define ) statements. How would I go about doing that?
14:04 asumu: DanBurton: if you're in a top-level context, a `begin` will splice it appropriately.
14:04 Araq: tewk: isn't this too restrictive in practice?
14:04 asumu: (or in a context that allows define in general)
14:05 (join) carleastlund
14:08 (quit) snearch: Quit: Verlassend
14:11 (part) tonyg
14:11 DanBurton: ok, so suppose I want to splice together two names input into a macro. for example, for (struct foo (bar)), I want to generate (define foo-bar ...)
14:11 DanBurton: how do I go about doing that?
14:13 Araq: tewk: if you can only have shared vectors of bytes, fixed integers and floats, why do you even garbage collect these vectors? sounds low level enough to use manual memory management then ;-)
14:14 (quit) nilyaK: Ping timeout: 244 seconds
14:15 (join) nilyaK
14:15 tewk: Araq: manual memory leads to mistakes and crashes.  Places maintain th VM's safety properties.
14:16 Araq: you can get memory safety with manual memory management too
14:16 Araq: you only need a per type allocator
14:16 Araq: but alright
14:18 (join) gciolli
14:19 asumu: rudybot: doc format-id
14:19 rudybot: asumu: your sandbox is ready
14:19 rudybot: asumu: no docs for a current binding, but provided by: racket/syntax
14:19 Araq: tewk: how is the "stop the world" implemented? instruction stream injection in loops and function calls?
14:19 asumu: DanBurton: see `format-id` from racket/syntax.
14:26 (quit) gciolli: Quit: Leaving.
14:35 (quit) yoklov: Quit: computer sleeping
14:39 (join) soegaard
14:44 (join) josdeha_rt
14:55 (quit) josdeha_rt: Remote host closed the connection
15:02 (join) random_malice
15:02 (join) bro_grammer
15:05 (quit) hash_table: Ping timeout: 246 seconds
15:05 (quit) getpwnam: Ping timeout: 246 seconds
15:08 (quit) mithos28: Read error: Connection reset by peer
15:08 (join) mithos28
15:10 (join) anRch
15:13 (join) josdeha_rt
15:30 (quit) soegaard: Quit: soegaard
15:38 (join) soegaard
15:40 (quit) josdeha_rt: Remote host closed the connection
15:42 (quit) mithos28: Read error: Connection reset by peer
15:42 (join) mithos28_
15:49 (quit) mithos28_: Read error: Connection reset by peer
15:49 (join) mithos28
15:51 (quit) mithos28: Read error: Connection reset by peer
15:51 (join) mithos28_
15:53 (quit) mithos28_: Read error: Connection reset by peer
15:53 (join) mithos28
15:55 (join) yoklov
15:59 (join) Fare
16:00 (join) rbarraud
16:01 (quit) soegaard: Quit: soegaard
16:02 (join) josdeha_rt
16:03 (join) soegaard
16:06 (quit) bitonic: Ping timeout: 246 seconds
16:07 (nick) imami|afk -> banseljaj
16:11 (quit) josdeha_rt: Remote host closed the connection
16:14 (join) getpwnam
16:15 (join) hash_table
16:15 (join) axe_wielder
16:16 (quit) bro_grammer: Ping timeout: 250 seconds
16:16 (quit) random_malice: Ping timeout: 250 seconds
16:18 (quit) getpwnam: Ping timeout: 252 seconds
16:22 (quit) soegaard: Quit: soegaard
16:23 (nick) elliottcable -> erparse
16:27 (join) elliottcable
16:28 (quit) kanak: Ping timeout: 250 seconds
16:37 (join) soegaard
16:43 (join) RacketCommitBot
16:43 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/0kz69w
16:43 RacketCommitBot: [racket/master] Use `unstable/lazy-require` to dynamically-load the compile-time of `match`. - Sam Tobin-Hochstadt
16:43 (part) RacketCommitBot
16:44 (quit) axe_wielder: Ping timeout: 248 seconds
16:45 (quit) hash_table: Ping timeout: 248 seconds
16:47 (quit) anRch: Quit: anRch
16:50 (quit) MayDaniel: Read error: Connection reset by peer
16:50 (quit) mithos28: Read error: Connection reset by peer
16:50 (join) mithos28_
16:56 (join) snearch
16:58 (quit) DanBurton: Quit: leaving
17:01 (quit) mithos28_: Read error: Connection reset by peer
17:02 (join) mithos28
17:03 (join) untrusted
17:06 (quit) maahes: Ping timeout: 244 seconds
17:12 (quit) mithos28: Read error: Connection reset by peer
17:12 (join) mithos28_
17:16 (quit) soegaard: Quit: soegaard
17:25 (quit) nilyaK: Quit: Leaving.
17:32 (join) getpwnam
17:33 (join) hash_table
17:37 (join) soegaard
17:42 (join) bitonic
17:47 (quit) jrslepak: Quit: This computer has gone to sleep
17:52 (quit) mceier: Quit: leaving
18:01 (quit) soegaard: Quit: soegaard
18:05 (join) soegaard
18:08 (join) jrslepak
18:11 (quit) untrusted: Remote host closed the connection
18:13 (join) jyc
18:15 (quit) yoklov: Quit: computer sleeping
18:17 (quit) snearch: Quit: Verlassend
18:24 samth: Araq: more info is available in the futures paper: http://dl.acm.org/authorize.cfm?key=389126
18:26 (quit) soegaard: Quit: soegaard
18:26 Araq: thanks; what's incremental parallelization?
18:28 Araq: never mind, got the paper
18:42 (quit) Fare: Ping timeout: 245 seconds
18:48 (part) jtpercon
18:49 (quit) jacius: Remote host closed the connection
18:53 (join) yoklov
18:54 (quit) mithos28_: Quit: mithos28_
19:03 (quit) danl_ndi: Quit: leaving
19:09 (quit) bitonic: Read error: Connection reset by peer
19:36 (nick) samth -> samth_away
19:40 (quit) jonrafkind: Ping timeout: 265 seconds
20:02 (join) dnolen
20:26 (quit) stchang: Read error: Operation timed out
20:28 (join) stchang
20:28 Nisstyre: For mutable fields in structs, how come the guard function doesn't get called when you do set-foo-bar! ?
20:29 Nisstyre: That would be nice if it did
20:30 Nisstyre: I suppose it would have the wrong number of parameters though
20:30 Nisstyre: but it could even just be with the updated one
20:32 (quit) yoklov: Quit: computer sleeping
20:37 (quit) masm: Quit: Leaving.
20:45 (join) ski
20:50 (quit) ski: Ping timeout: 255 seconds
20:55 asumu: Nisstyre: that's what contracts are for.
20:58 asumu: rudybot: (struct point (x y) #:mutable)
20:58 rudybot: asumu: your sandbox is ready
20:58 rudybot: asumu: Done.
20:59 asumu: rudybot: (define/contract (make-pt-ctc x y) (-> integer? integer? (struct/c point integer? integer?)) (point x y))
20:59 rudybot: asumu: Done.
20:59 asumu: rudybot: (set-point-x! (make-pt-ctc 3 2) "bad")
20:59 rudybot: asumu: error: make-pt-ctc: contract violation, expected: integer?, given: "bad" <NEWLINE>   contract from: (function make-pt-ctc), blaming: program <NEWLINE>   contract:  <NEWLINE>     (-> <NEWLINE>      integer? <NEWLINE>      integer? <NEWLINE>      (struct/c point integer? integer?)) <NEWLINE>   at: #<undefined>:1.18
21:00 asumu: Nisstyre: ^ something like that
21:16 Nisstyre: asumu: makes sense, thanks
21:17 (quit) carleastlund: Quit: carleastlund
21:27 (join) RacketCommitBot
21:27 RacketCommitBot: [racket] plt pushed 3 new commits to master: http://git.io/9ToyoQ
21:27 RacketCommitBot: [racket/master] racket/gui/init: module caching in the load handler - Matthew Flatt
21:27 RacketCommitBot: [racket/master] doc and error-message fixes for module name resolver - Matthew Flatt
21:27 RacketCommitBot: [racket/master] FFI repairs for 64-bit big-endian - Matthew Flatt
21:27 (part) RacketCommitBot
21:30 (join) DrDuck
21:32 (quit) DrDuck: Read error: Connection reset by peer
21:50 (join) yoklov
21:52 (join) bjz
22:31 (quit) dnolen: Remote host closed the connection
22:44 (quit) bjz: Ping timeout: 246 seconds
22:53 (join) bjz
22:55 (join) jacius
23:10 (join) mithos28
23:21 mithos28: Does anyone know what the benefit sam is getting with lazy-require in his recent pushes?
23:29 (join) jonrafkind
23:29 (quit) jonrafkind: Changing host
23:29 (join) jonrafkind
23:42 (quit) getpwnam: Ping timeout: 246 seconds
23:43 (quit) hash_table: Ping timeout: 246 seconds