00:07 (join) rekahsoft
00:54 (join) mithos28
00:59 (quit) dnolen: Quit: dnolen
00:59 (quit) mithos28: Quit: mithos28
01:04 (join) mithos28
01:04 (quit) yoklov: Quit: Leaving.
01:20 (quit) Nightwolf: Ping timeout: 250 seconds
01:20 (quit) jonrafkind: Remote host closed the connection
01:21 (quit) mithos28: Quit: mithos28
01:21 (join) erjiang
01:31 (join) mg4001
01:40 (join) sheikra
01:41 sheikra: hello!
01:42 sheikra: I have a question about the structure types
01:42 sheikra: I defined a structure type Arr  with (struct Arr (from to serial))
01:43 sheikra: then defined a helper function for making Arr with a "default" value for serial, that is #f.
01:43 sheikra: (define Arrow (lambda (from to) (Arr from to #f)))
01:44 sheikra: then I redefined Arr with same definition. then Arrow breaks...
01:44 sheikra: (Arr? (Arrow 1 2)) returns #f ...
01:57 (quit) erjiang: Quit: ttfn
03:13 (quit) mg4001: Quit: Leaving
03:24 (quit) sheikra: Read error: Connection reset by peer
03:25 (join) sheikra
03:37 (join) hkBst
03:48 (quit) realitygrill: Ping timeout: 240 seconds
03:50 (join) Nightwolf
03:55 (join) realitygrill
04:00 (join) leo2007
04:09 (quit) leo2007: Ping timeout: 246 seconds
04:44 (quit) ckrailo: Quit: Computer has gone to sleep.
04:53 (quit) rekahsoft: Ping timeout: 276 seconds
04:55 (join) rekahsoft
05:07 (join) sheikra_
05:09 (quit) sheikra: Ping timeout: 276 seconds
05:56 (join) masm
05:57 (quit) sheikra_: Quit: Leaving
05:59 (join) sheikra
06:30 (quit) sheikra: Quit: Leaving
06:43 (join) yoklov
07:40 (join) leo2007
07:47 (quit) yoklov: Quit: Leaving.
08:22 (join) MayDaniel
08:31 (join) yoklov
08:38 yoklov: this might be intentional, but why is there a difference between specifying #lang r5rs and setting drracket to determine language from source versus setting drracket to r5rs mode?
08:40 yoklov: the difference that i've noticed, though there may be others, is that lists are displayed as mcons's.  for example, in determine language from source #lang r5rs, if you type (list 1 2 3) it will display (mcons 1 (mcons 2 (mcons 3 '()))), whereas in r5rs mode it will display (1 2 3)
08:40 (quit) MayDaniel: Read error: Connection reset by peer
08:40 yoklov: i mean, i know why it does it, its because it handles r5rs lists like mutable lists, which they are, but is the difference intentional?
08:47 ohwow: hm, strange D:
08:54 yoklov: yeah, I'm taking a class on scheme and all we use is r5rs
08:54 yoklov: but i program in racket all the time
08:54 yoklov: so it's been bugging me for a while
09:19 (join) ChibaPet
09:20 ChibaPet: Good morning.
09:30 (join) tauntaun
09:41 (join) dnolen
09:41 (join) sheikra
09:59 askhader: hi
10:04 (join) Zorlin
10:07 (quit) tauntaun: Quit: Ex-Chat
10:19 (quit) yoklov: Quit: Leaving.
10:45 (join) tfb
11:07 (quit) masm: Ping timeout: 240 seconds
11:11 (quit) dnolen: Quit: dnolen
11:17 (join) ckrailo
11:23 (join) cmatheson
11:24 (join) mithos28
11:24 (nick) samth_away -> samth
11:25 samth: yoklov, the reason is that the #lang r5rs doesn't change the printer, and the language level does
11:25 samth: the ability for #lang to change the printer is somewhat new
11:26 samth: sheikra, structures are generative, which means that every time you write (struct Arr ...) it creates a *new* version of the struct
11:26 samth: and your `Arrow' function refers to the old version of the struct
11:26 samth: whereas `Arr?' is for the new version
11:27 sheikra: samth: but I think the Arrow function should refer to the new version of the constructor Arr.
11:27 sheikra: samth: the constructor should behave the same as a function
11:27 sheikra: samth: The new constructor will create new records, which are different from the ones already created.
11:28 sheikra: samth: The Arrow function should call this new version
11:28 samth: sheikra, you only need to say my name once :)
11:28 samth: see the docs here: http://docs.racket-lang.org/guide/define-struct.html?q=struct#%28part._.Structure_.Type_.Generativity%29
11:28 rudybot: http://tinyurl.com/3j7xevg
11:28 samth: the `Arrow' function doesn't change when you define a new version of the struct
11:28 sheikra: samth: Arrow should create new version structures. And the new Arr? should return #f for all old ones that are already created, but should return #t for the ones created with Arrow
11:29 sheikra: samth: It doesn't change, but it should call the new Arr
11:29 sheikra: samth: let's make an analogy..
11:29 samth: basically, the new (struct Arr ...) form creates a new function called `Arr', it doesn't change the old one
11:29 sheikra: samth: (define f (lambda (x) (+ 1 x)))
11:30 sheikra: samth: (define g (lambda (x) (f x)))
11:30 samth: and the `Arrow' function calls the old function named `Arr'
11:30 samth: and you didn't do anything to change `Arrow'
11:30 sheikra: samth: when we redefine f, the new calls to g should change their behaviors
11:30 samth: sheikra, i think you'll find that racket doesn't do that either for `f' and `g'
11:30 sheikra: samth: I didn't do anything to change g either in this example
11:31 sheikra: rudybot: hello?
11:31 rudybot: sheikra: Hello, I'm using twitter.el to tweet. Could you please say how to view replies only in it ?
11:31 sheikra: rudybot: init racket
11:31 rudybot: sheikra: your sandbox is ready
11:32 sheikra: rudybot: (define f (lambda (x) (+ 1 x)))
11:32 rudybot: sheikra: Done.
11:32 sheikra: rudybot:  (define g (lambda (x) (f x)))
11:32 rudybot: sheikra: Done.
11:32 sheikra: rudybot: (g 2)
11:32 rudybot: sheikra: ; Value: 3
11:32 sheikra: rudybot: (define f (lambda (x) (+ 2 x)))
11:32 rudybot: sheikra: Done.
11:32 samth: ah, right
11:32 sheikra: rudybot: (g 2)
11:32 rudybot: sheikra: ; Value: 4
11:32 sheikra: samth: see?
11:33 sheikra: samth: I think there shouldn't be any difference here with redefining f.
11:33 samth: the reason that it's doing this differently for structs is because the `Arr' function is really a macro which expands into a reference to a (secretly) named real constructor
11:34 sheikra: samth: That's my guess, and I think that's wrong
11:34 samth: probably if you use `define-struct' instead of `struct', it'll work
11:35 (join) PLT_Notify
11:35 PLT_Notify: racket: master Casey Klein * 4b03ddc (2 files in 2 dirs): Fixes Redex stepper's handling of symbols quoted with || - http://bit.ly/kSxiz4
11:35 (part) PLT_Notify
11:35 samth: but more fundamentally, in Racket we don't mostly care about things like redefintion in the top-level
11:35 samth: in general, these things are pretty hopeless at the top-level
11:36 sheikra: samth: yes, define-struct works correctly
11:36 sheikra: samth: yep. I'm just picky ;-)
11:37 samth: http://www.google.com/search?q=mflatt+the+top-level+is+hopeless+site:racket-lang.org
11:37 rudybot: http://tinyurl.com/3k8bcpb
11:38 samth: sheikra, it's fine to be picky, but there's no real way to "fix" this while maintaining all the other properties we want the `struct' form to have
11:39 sheikra: samth: why is that?
11:39 samth: the (obvious) alternatives are 1. make Arr not be both the constructor and a macro
11:39 sheikra: samth: I don't really know the difference between struct and define-struct
11:40 samth: or 2. re-expand the definition of Arrow every time
11:40 samth: 1 is bad, because we wanted to drop the extra `make-' prefixes, and have symettrey with pattern matching
11:41 samth: that's why we switched to what `struct' does (combine the constructor and the syntactic info) from what define-struct does (separate them)
11:41 sheikra: samth: yes. that's why I like struct more than define-struct
11:41 samth: 2 is really bad
11:42 sheikra: samth: why must Arr be a macro?
11:42 samth: because it has static information
11:42 sheikra: samth: I don't quite understand ..
11:42 samth: like the information that `match' uses for pattern matching on (Arr x y z)
11:43 sheikra: samth: but I guess you can keep a table for that?
11:43 samth: it's not that easy
11:43 sheikra: samth: Arr has the type 'a -> 'b -> 'c -> Arr a b c
11:43 samth: much better for it to be part of the binding `Arr'
11:44 samth: `define-struct' has a binding for the information too
11:44 samth: for example, (define-struct Arr (x y z)) defines `Arr' as the static information, and `make-arr' as the constructor
11:45 samth: (struct Arr (x y z)) is similar, but it hides the name `make-arr'
11:45 sheikra: samth: do you mean by "static information" something like the names of the fields?
11:46 samth: yes, among other things
11:46 samth: see http://docs.racket-lang.org/reference/structinfo.html?q=struct
11:46 sheikra: samth: but Arr in the toplevel gave me #<prodcedure:Arr>
11:47 samth: yes, it expands to a procedure with the same name
11:48 samth: but the name embedded in the procedure isn't related to the names of binding
11:48 samth: it's just for printing
11:48 sheikra: samth: Hmm.. I see .. that problem is that match is implemented as a macro
11:49 samth: there are a bunch of macros that use that information
11:49 sheikra: samth: it only has access to normal data type in the program domain ..
11:49 samth: `shared' and `struct-out' are two other examples
11:49 samth: also `struct-copy'
11:50 sheikra: samth: If match is implemented in the compiler level, then we can store these information inside the type for functions ...
11:50 (join) yoklov
11:50 samth: yoklov, the reason is that the #lang r5rs doesn't change the printer, and the language level does
11:50 yoklov: i see
11:50 samth: sheikra, i don't know what you mean by type
11:51 sheikra: samth: look up a symbol table, get this function object, and take the predicate etc out ..
11:51 samth: and a large part of the point of Racket is that `match' doesn't have to be implemented in the compiler
11:51 (quit) hkBst: Read error: Connection reset by peer
11:51 (join) masm
11:51 sheikra: samth: I mean the object inside the compiler to represent a function
11:52 sheikra: samth: maybe we could use a hashtable in the program space to map the function object Arr to some data structure?
11:52 samth: sheikra, `match' isn't a runtime operation
11:53 samth: and really, if this is the only "problem" with defining `Arr' as a macro, then I don't think anyone is going to worry
11:54 sheikra: samth: I know, but it can generate code to lookup the table
11:54 sheikra: samth: while keep Arr a function
11:55 samth: so, now there's a giant global mutable table that match needs to know about
11:55 samth: and that `shared' and `struct-copy' need to know about
11:55 sheikra: samth: any way ...
11:55 samth: plus, how can match manage the binding structure that way?
11:59 sheikra: samth: probably not worth the trouble with macros. match should really be inside the compiler, then the static info can be stored directly in the function, without any look up..
12:00 sheikra: samth: I think it's worth doing, because match is something very fundamental
12:00 samth: sheikra, i don't see why you think match should be in the compiler
12:00 samth: (and I'm the implementor of match)
12:01 samth: tons of things in Racket aren't in the compiler
12:01 samth: such as classes, units, keyword arguments, optional arguments, pattern matching, type checking, etc
12:01 samth: this is something that we work hard to accomplish, not a drawback
12:01 sheikra: samth: because match is where computation really happens, just like case expressions in haskell
12:02 samth: sheikra, no, that isn't the case
12:02 samth: match is a very useful programming tool, but function calls are where computation happens
12:03 sheikra: samth: we can implement "if", "cond", ... with match
12:03 samth: sheikra, (a) no you can't, and (b) cond is a macro
12:04 sheikra: samth: function calls only bind the names to values, but we need to "look at" this value in order to do useful things
12:04 samth: sheikra, in Racket, to look at values, we use procedures such as `number?' and `first'
12:04 sheikra: samth: and match is the way we destruct the value and operate on it..
12:05 samth: sheikra, unlike in Haskell and ML, pattern matching is not fundamental to computation in Racket
12:06 sheikra: samth: oh, btw. is there a way I can match a symbol or a number without using the pattern (? symbol? x) ?
12:06 samth: instead, pattern matching is a convenient mode of use of the fundamentals of computation, which are procedures
12:06 samth: sheikra, there isn't one built-in
12:06 samth: but you could write one like (symbol: x) pretty easily
12:06 sheikra: samth: I hope it's something like (Number x) or (Symbol x)
12:07 samth: (define-match-expander Symbol (syntax-rules () [(_ p) (? symbol? p)]))
12:07 samth: rudybot, eval (define-match-expander Symbol (syntax-rules () [(_ p) (? symbol? p)]))
12:07 rudybot: samth: your sandbox is ready
12:07 rudybot: samth: Done.
12:07 samth: rudybot, eval (match 'foo [(Symbol x) x])
12:07 rudybot: samth: ; Value: foo
12:08 samth: `Number' would be similarly easy to write, sheikra
12:09 sheikra: samth: that works. Thank you :-)
12:10 (join) dnolen
12:21 (join) jonrafkind
12:24 samth: EM03, Racket is indeed actively developed, and there are lots of plans for it in the future
12:24 samth: what do you want to know about?
12:25 (quit) tfb: Quit: gone
12:28 yoklov: here's somethign i wanted to know about: are there any plans for developing a sound/audio manipulation library/toolkit?
12:29 samth: yoklov, john clements has done some work in that area
12:30 yoklov: yeah, but the library crashes on windows, so I can't use it with my desktop.  i've meant to file a bug report actually, but haven't had the time
12:30 samth: ah, too bad
12:31 samth: i would say that john has plans, and i don't know of other people with such plans
12:31 samth: there are a few other bindings to audio libraries on planet
12:31 samth: synx/ao and evhan/coremidi
12:31 yoklov: his is the most fully developed by far from what i've seen, and does work great on unix systems
12:33 yoklov: yeah, so when a library gets good enough, it gets packaged with racket?
12:33 samth: yoklov, the plan is for more things to be distributed separately in the future
12:33 samth: so that racket isn't quite so monolithic to download/build/install
12:34 yoklov: thats reasonable
12:34 jonrafkind: hopefully a 'racket + everything useful on planet' will be packaged
12:35 yoklov: lol, the only problem with that I see is that planet isn't that great yet
12:35 yoklov: things fail to work a lot
12:35 yoklov: or have for me at least
12:36 yoklov: also the require syntax is awkward
12:38 eli: yoklov: What kind of sound support are you looking for?
12:38 eli: yoklov: Also, what's awkward?
12:39 yoklov: oh, the fact that it's not unified with the built-in libraries and that even if you have the package on your system you have to know its version number and developer
12:39 yoklov: and sound support? anything cross platform, really, even if its just 'play this frequency for this amount of time'
12:39 yoklov: i can abstract away from that
12:40 eli: yoklov: You were talking about the syntax, no?
12:41 yoklov: the syntax of require planet libraries is what i meant
12:41 yoklov: in racket
12:41 yoklov: maybe thats not the right term.
12:41 eli: Can you give an example of the syntax that bugs you?
12:42 yoklov: (require (planet "main.rkt" ("clements" "rsound.plt" 1 10)))
12:43 yoklov: its tough because some libraries have the same name
12:43 eli: yoklov: You should be able to use something like (require planet/clements/rsound) instead.
12:43 yoklov: hm
12:43 yoklov: i guess i wasn't aware of that.
12:45 yoklov: (require (planet clements/rsound)) works
12:45 eli: yoklov: That's what I suspected -- I asked the planet guy to change the web pages.
12:45 eli: And re sound support -- note that John's thing does sound waves, and it's not really something that could be lumped as a core functionality yet.  But if you want to do just windows, then the ffi glue is usually very simple.
12:46 yoklov: i wanted cross platform, ended up sorta postponing my project to a later date when something better exists or I know enough C or whatever to make what i want happen
12:46 yoklov: its not a big deal, i was just wondering what the status of it was
12:47 eli: Do you know about the MCI interface?
12:47 yoklov: no
12:47 eli: It makes it very easy to do most high-level sound things.
12:47 eli: The API is basically just sending strings with media commands.
12:48 (join) mceier
12:51 yoklov: still though, not cross platform (or at least, wikipedia's description of it makes it sound that way) but i guess i could use rsound or something to do it on unix, or maybe even just figure out why rsound crashes racket for me
12:53 eli: Yes, it's not cross platform.
12:53 yoklov: hm, it says it is
12:53 yoklov: i think
12:53 yoklov: or did
12:53 eli: Probably for whatever MS defines as "cross platform"...
12:54 eli: The sad summary: Windows has a really nice sound interface, OSX has something extremely complicated, linux is a mess of libraries and low level systems that is impossible to figure out.
12:54 yoklov: yeah
13:03 (quit) sheikra: Read error: Connection reset by peer
13:08 yoklov: its probably possibel to do bindings for CSound in racket, maybe when I'm further along in my degree if I have an abundance of spare time I'll try to make that work
13:19 (join) lucian
13:34 (quit) swiftkick: Ping timeout: 260 seconds
13:44 (quit) leo2007: Ping timeout: 260 seconds
13:46 (join) carleastlund
13:47 (join) PLT_Notify
13:47 PLT_Notify: racket: master Matthew Flatt * 5e7d1f2 (1 files in 1 dirs): fix `get-char-width' and `get-char-height' in case font match fails
13:47 PLT_Notify: racket: master Matthew Flatt * 34b8dc2 (3 files in 3 dirs): fix documented contracts for path functions ...
13:47 PLT_Notify: racket: master commits 4b03ddc...34b8dc2 - http://bit.ly/lz7akH
13:47 (part) PLT_Notify
13:52 samth: eli, ping
13:52 (join) groovy2shoes
13:55 (join) anRch
14:01 (join) MayDaniel
14:11 (quit) lucian: Remote host closed the connection
14:15 (join) lucian
14:24 (quit) masm: Remote host closed the connection
14:28 (part) Dodek
14:31 (join) masm
14:31 (quit) yoklov: Quit: Leaving.
14:38 (part) jonrafkind
14:39 (join) jonrafkind
14:42 (join) swiftkick
14:50 (quit) lucian: Ping timeout: 252 seconds
14:54 (quit) groovy2shoes: Read error: Connection reset by peer
14:54 eli: samth: ?
14:54 (join) groovy2shoes
14:54 (quit) groovy2shoes: Changing host
14:54 (join) groovy2shoes
14:58 (quit) MayDaniel: Read error: Connection reset by peer
15:00 (join) Counterspell
15:03 (join) lucian
15:03 (join) PLT_Notify
15:03 PLT_Notify: racket: master Robby Findler * 0415bf6 (1 files in 1 dirs): make (many of) the rackunit examples executable ... - http://bit.ly/mvISkR
15:03 (part) PLT_Notify
15:10 (join) entelechy
15:11 (quit) swiftkick: Ping timeout: 240 seconds
15:15 (quit) anRch: Quit: anRch
15:16 entelechy: .
15:16 (nick) entelechy -> swiftkick
15:22 (join) yoklov
15:28 (quit) yoklov: Quit: Leaving.
15:30 (quit) Counterspell: Remote host closed the connection
15:32 (join) Counterspell
16:02 (join) tonyg
16:06 (quit) ChibaPet: Ping timeout: 276 seconds
16:07 (join) ChibaPet
16:36 (join) lucian_
16:37 (quit) lucian: Disconnected by services
16:37 (nick) lucian_ -> lucian
16:38 (quit) groovy2shoes: Quit: groovy2shoes
16:39 (part) ChibaPet
16:50 samth: eli, you may have gotten a bounce mail from launchpad
16:59 (join) MayDaniel
17:02 (quit) rekahsoft: Read error: Operation timed out
17:02 (join) rekahsoft
17:06 (quit) rekahsoft: Ping timeout: 240 seconds
17:11 (quit) lnostdal: Read error: Connection reset by peer
17:12 (join) lnostdal
17:29 eli: samth: I didn't notice anything.  If it's important than tell me what was supposed to be there.
17:29 eli: (But I have to run out now.)
17:36 (join) tauntaun
17:39 (join) chemuduguntar
18:07 bremner_: is there a way to tell raco setup to build only the architecture dependant stuff (e.g. X86_64-linux/* ?)
18:11 (quit) lnostdal: Ping timeout: 248 seconds
18:11 (join) lnostdal
18:16 stamourv: bremner_: afaik, raco setup only builds zos and docs, which are platform-independent, I believe
18:17 bremner_: hmm. But there are a few architecture dependendent bits in collects, or are those gone?
18:18 stamourv: I'm not sure
18:33 (quit) MayDaniel: Read error: Connection reset by peer
18:36 (quit) lucian: Remote host closed the connection
18:49 (join) rekahsoft
19:09 (quit) dnolen: Ping timeout: 252 seconds
19:15 (quit) jonrafkind: Ping timeout: 240 seconds
19:16 (quit) DT``: Ping timeout: 240 seconds
19:20 (quit) carleastlund: Quit: carleastlund
19:21 (quit) mithos28: Quit: mithos28
19:22 bremner_: for example, in 5.0.2, there was collects/sgl/native/x86_64-linux, but in 5.1.1 that doesn't seem to be there anymore (for me)
19:22 bremner_: it looks like commit  557b799 killed that
19:24 bremner_: hmm and same for plot. hurray!
19:29 (join) DT``
19:30 (quit) mceier: Quit: leaving
19:31 (quit) cmatheson: Quit: leaving
19:35 (join) yoklov
19:35 (quit) Zorlin: Disconnected by services
19:35 (join) 50UAANW5T
19:35 eli: bremner_: SGL was the only such racket file, and it's gone; plot was compiled from C inside the collection, but that moved to src so it's gone too.
19:36 (join) yoklov1
19:36 (quit) yoklov: Read error: Connection reset by peer
19:37 bremner_: eli: great. That simplfies things. Is there some concensus that collects will stay arch independent?
19:43 eli: bremner_: Not really, it might return, still.  And of course once things are made more open then nothing will prevent people from, for example, making macros that expand in a way that depends on the platform.
19:46 bremner_: so different platforms will get different .zo files?
19:51 eli: bremner_: In general there's no way to prevent it.
19:52 eli: Any macro that uses (system-type) will likely result in such code.
19:52 eli: Or macros that depend on local files.
19:52 eli: (Like parsing header files.)
19:52 bremner_: blerg
19:52 eli: ?
19:52 (join) mithos28
19:53 eli: Assuming that "blerg" is a general dissatisfied sound, then why are you?
19:53 eli: (Why are you dissatisfied, that is.)
19:53 bremner_: well, if the .zo files are not (in the future) architecture independent, then the current debian packaging will break quite badly.
19:54 bremner_: perhaps more importantly, it means that cross compilation for small systems becomes very hard
19:55 bremner_: neither of which is really racket's problem
19:55 chandler: I'm getting a bit of deja vu from this conversation...
19:56 bremner_: I guess I'm hypothesizing systems that can run useful racket, but not build collects
20:11 (nick) samth -> samth_away
20:19 (quit) mithos28: Quit: mithos28
20:21 (join) dnolen
20:23 Counterspell: I am trying to "Show Profiling" from the View menu, but I get a message that says "There is no profiling information available. Please be sure that profiling is enabled n your language and you have run your program."
20:24 Counterspell: Now I did go into Language > Choose Language ... > selected "Use the language declared in the source) and selected Debugging and profiling under Dynamic Properties.
20:24 Counterspell: But I still get that message.
20:38 (quit) tauntaun: Quit: Ex-Chat
20:41 (quit) swiftkick: Ping timeout: 240 seconds
20:41 (join) mheld_
20:44 (quit) mheld_: Client Quit
20:52 (join) PLT_Notify
20:52 PLT_Notify: racket: master Robby Findler * d3c04ae (2 files in 2 dirs): Adjusted docs-complete to print a different message when there are no docs at all ...
20:52 PLT_Notify: racket: master Robby Findler * d40b43c (4 files in 2 dirs): make the deinprogramm documentation drdr messages be accurate ...
20:52 PLT_Notify: racket: master commits 0415bf6...d40b43c - http://bit.ly/kAFsuC
20:52 (part) PLT_Notify
21:31 (quit) ckrailo: Quit: Computer has gone to sleep.
22:11 (join) EM03_
22:11 (quit) EM03_: Changing host
22:11 (join) EM03_
22:12 (quit) EM03: Ping timeout: 248 seconds
22:12 (nick) EM03_ -> EM03
22:37 (join) jonrafkind
22:52 (join) leo2007
23:16 (quit) dnolen: Quit: dnolen
23:41 (quit) Counterspell: Quit: Ex-Chat