00:00 (quit) mithos28: Quit: mithos28
00:50 (join) Sgeo
00:51 Sgeo: Which is generally used more/preferred? swindle/setf, or srfi-17?
00:51 (join) PLT_Notify
00:51 PLT_Notify: racket: master Stephen Bloch * 79778e0 (3 files in 3 dirs): Added some illegal-argument tests to map-image.rkt. ... - http://bit.ly/mreQ0S
00:51 (part) PLT_Notify
01:07 eli: Sgeo: They're pretty different -- the swindle thing is like CL's `setf', and does the dispatch syntactically.  As in CL, it's faster (no runtime lookup), but more limited (can only be used with syntactically known identifiers).
01:28 (join) completenoob
01:28 completenoob: what does { } mean?
01:29 jonrafkind: nothing special, just alternate syntax for ()
01:29 jonrafkind: rudybot: {+ 1 1}
01:29 rudybot: jonrafkind: your sandbox is ready
01:29 rudybot: jonrafkind: ; Value: 2
01:29 completenoob: can i redefine it?
01:29 jonrafkind: you want to redefine what {+ 1 1} means? using {} as an application?
01:30 completenoob: i mean just using syntax-case or something
01:30 jonrafkind: are you saying you have  macro where one of the sub-expressions uses {} and you want to detect that?
01:30 completenoob: nope I want to redefine what expressions inside of { } mean
01:31 completenoob: although
01:31 completenoob: can i detect it?
01:31 jonrafkind: you can but its not trivial, you to inspect the syntax object, detect that it uses {} using (syntax-property expr 'paren-shape) and then transform it
01:31 completenoob: whoa
01:32 completenoob: what about the first case, where I just want to redefine the syntax to look for a { } pattern and transform the code between them?
01:32 jonrafkind: rudybot: (syntax-property (with-input-from-string "{1}" read-syntax) 'paren-shape)
01:32 rudybot: jonrafkind: ; Value: #\{
01:32 jonrafkind: rudybot: (syntax-property (with-input-from-string "(1)" read-syntax) 'paren-shape)
01:32 rudybot: jonrafkind: ; Value: #f
01:33 jonrafkind: so just run that over each syntax object that you get in your macro
01:33 completenoob: regular parens are false?
01:33 jonrafkind: (syntax-case expr () [(_ stuff ...) (map (lambda (s) (syntax-property ... s ...)) (syntax->list #'(stuff ...)))
01:33 jonrafkind: right, regular parens are #f
01:34 jonrafkind: brackets are  #\[, and { are #\{
01:37 completenoob: I'm kind of new to Scheme in general
01:37 completenoob: is there a way to introduce new bindings with a macro?
01:37 completenoob: (by 'kind of new' i mean 'an hour ago')
01:39 jonrafkind: yes
01:39 jonrafkind: this stuff is sort of advanced for someone new to it..
01:39 jonrafkind: but anyway (define-syntax-rule (foo name value) (define name value))
01:39 jonrafkind: (foo x 2) is the same as (define x 2)
01:40 completenoob: (I'm not new to Lisp, just Scheme)
01:40 jonrafkind: oh ok
01:40 jonrafkind: yes, in scheme macros are hygienic by default, i gather in lisp thats not the case
01:41 completenoob: right, in Lisp you can make a version of "when" that lexically binds the identifier 'it' to the result of the test, expression, for instance
01:41 jonrafkind: so using the above example, if you had (define-syntax-rule (foo value) (define x value)), that would define a name 'x' that no one can use because its defined by the macro
01:41 jonrafkind: which is why i had to pass the 'name' variable in
01:41 jonrafkind: yes, that sort of binding 'it' is unhygienic, which you can do in scheme as well
01:42 jonrafkind: (with-syntax ([it (datum->syntax stx 'it stx)]) #'(... it ...))
01:42 jonrafkind: datum->syntax imbues the syntax object with whatever lexical properties you want
01:42 jonrafkind: actually maybe using syntax-local-introduce is better, i forget..
01:43 eli: Not.
01:43 jonrafkind: how is it any different?
01:44 completenoob: I noticed that the "Swindle" language has a 'defmacro' that is supposedly totally unhygienic, but I tried using it and it didn't seem to work the way I was used to...
01:44 (quit) dnolen: Quit: dnolen
01:44 jonrafkind: I suppose if the source expression has compltely different lexical context than the current macro usage then it might matter
01:44 eli: To break hygiene, use this: http://blog.racket-lang.org/2008/02/dirty-looking-hygiene.html
01:45 jonrafkind: 404
01:45 completenoob: worked for me...
01:45 jonrafkind: oh now it works, very wierd
01:45 eli: completenoob: the `defmacro' in Swindle is almost the same as the `mzlib/defmacro' -- they're both pretty bad.
01:46 eli: jonrafkind: Some blogspot problem.
01:46 completenoob: are [] and {} a part of the Scheme standard or are they just Racket?
01:47 eli: completenoob: In R6RS []s are also specified to behave the same as ()s, not {}s though.
01:47 completenoob: hmm...
01:48 jonrafkind: the last paragraph in that blogpost refers to 'if' when it looks like it should refer to 'if*'
01:48 jonrafkind: im confused..
01:49 completenoob: what's the most idiomatic way to do partial application? just writing a lambda inline?
01:49 jonrafkind: oh its simalteanously talking about if* -- the implementation, and 'if' the rename provided thing from the implementing module
01:50 jonrafkind: damn eli, thats the most confusing thing ive ever seen
01:50 jonrafkind: completenoob, theres a curry library if you want
01:51 jonrafkind: http://docs.racket-lang.org/reference/procedures.html?q=curry#(def._((lib._racket/function..rkt)._curry))
01:51 eli: jonrafkind: No, I think that it's a typo.
01:51 jonrafkind: oh ok
01:52 eli: Uh, actually maybe you're right, and I did talk about the provided renamed `if', since the point was to make an arc-like language.
01:52 completenoob: the example in that article was notreally satisfactory. I WANT 'it' to be lexical.
01:52 eli: completenoob: It is.
01:53 completenoob: at the end of the article he claims it's global, so any lexical definition above the if will take priority
01:53 jonrafkind: if*'s `it' is in an outer scope compared to the let-bound `it'
01:53 completenoob: hmmm
01:53 completenoob: it's a minor issue for sure
01:53 jonrafkind: I suppose if you really want to overwrite the scope of `it' no matter what then use syntax-local-introduce or datum->syntax, but in general using syntax-parameterize is probably the 'right' solution
01:53 eli: completenoob: `it' is globally bound to a special kind of macro, which is rebound by `if*'s.
01:54 completenoob: yeah now that I'm sort of trying it on in my head, the hygienic solution does feel... better.
01:55 eli: completenoob: If you're just coming from Lisp, then this might work out better: http://blog.racket-lang.org/2011/04/writing-syntax-case-macros.html
01:56 completenoob: I'm getting the image of "CL = buddha is fat with hairy armpits and laughs, Scheme = buddha is thin and serious"
01:56 eli: completenoob: You've probably spent time on c.l.l
01:56 eli just hates all of these vague analogies
01:57 completenoob: haha
01:57 jonrafkind: supposedly racket is sufficiently different from scheme that you should just call it racket
01:57 eli: Things always go downhill when they're pulled out.
01:57 jonrafkind: namely, (equal? (- Racket Scheme) (- Scheme Lisp)) => #t
01:58 eli: (I've seen so many arguments on c.l.l degrade to discussions about latin.)
01:58 completenoob: hmmm, then Racket might be a good starting place
01:59 completenoob: my problem with them is you say "I have an idea for an object system what do you thi-" "CLOS is superior."
01:59 jonrafkind: ive never felt a need to modify my object protocol, do people do that in CLOS a lot?
01:59 completenoob: well they use the MOP to change the object protocol sometimes
01:59 completenoob: but see you said "my object protocol"
01:59 eli: completenoob: (FWIW, that budha argument is usually used when comparing CL the standard to Scheme the standard; if you consider racket specifically, then it's way larger than CL.)
02:00 completenoob: not "the object protocol"
02:00 jonrafkind: sorry, i mean the object protocol
02:00 eli: jonrafkind: IME, rarely -- as much as I love the whole concept...
02:00 completenoob: well, I've got an idea for an object system :p
02:00 eli: completenoob: CLOS has a bunch of big problems, idolizing it is ... questionable, IMO.
02:01 jonrafkind: I suppose synchronous method dispatch is sort of nice sometimes
02:01 eli: What's that?
02:01 completenoob: eli: I ahven't used it extensively. If it really is the best thing ever, I'll be glad to reinvent it and then accept it. I like to learn by reinventing
02:02 jonrafkind: where all the arguments are checked to find the most specific method, instead of just checking the first one
02:02 eli: jonrafkind: I've never heard that name for what you're talking about...
02:02 jonrafkind: i read it in some OO papers
02:02 eli: completenoob: I have.  I worked hard for my opinion...
02:03 (join) Nican_
02:03 jonrafkind: anyway my eyes are bleeding, cya
02:03 eli: jonrafkind: So early?
02:03 completenoob: eli: good for you! :D I mean, I'd still be using C++/Java if I didn't have this attitude, right?
02:04 eli: completenoob: Yeah, well, looking at C++ makes everything looks ideal...
02:04 (quit) chemuduguntar: Remote host closed the connection
02:05 eli: completenoob: BTW, I made swindle since I loved the idea so much...  But some of the problems are impossible to iron out.
02:05 completenoob: eli: yep, but it wasn't until I accidentally reinvented a bizarre smalltalk-lisp hybrid with YAML files that I decided to abandon it
02:05 eli: Abandon C++?
02:06 completenoob: yeah
02:06 completenoob: not completely, but as the language I think in
02:06 (quit) Nican: Ping timeout: 240 seconds
02:06 eli: Heh, in that case it does sound like a good case of someone who learns through his feet.
02:07 completenoob: so you made Swindle?
02:07 eli: Yes.
02:07 (quit) Nican_: Ping timeout: 240 seconds
02:07 (quit) jonrafkind: Ping timeout: 252 seconds
02:08 completenoob: cool! that's a lot of stuff in there. I didn't even realize that the lets didn't work the way yours do >.>
02:09 eli: You mean the extended `let' syntax?
02:09 completenoob: yeah
02:09 eli: Heh, I was just having fun playing with the freedom of making my own language...
02:09 completenoob: do you write in Swindle?
02:10 eli: Not any more.
02:10 eli: The CLOS problems make it too bad.
02:10 completenoob: what would you say are the biggest weaknesses of CLOS?
02:11 eli: The biggest problem IMO is that it inherently relies on side-effects when you add methods to generic functions.
02:12 eli: It's just a mess, and there's no way to guarantee any kind of safety.
02:12 eli: But of course CLOS advocates put that lack of safety on a shrine...
02:12 eli: (There's some quote on CLOS being for nice people and C++ for thieves, or something like that.)
02:13 completenoob: erik naggum?
02:13 eli: Yeah, he made it.
02:13 completenoob: well doesn't 'defun' have the same sort of problem? really any non-lexical definition is a side-effect.
02:13 eli: Yes, using a toplevel to develop code is bad in a similar way.
02:14 eli: Same way that `load' is much worse than using modules.
02:14 eli: But CLOS takes the whole thing to another level.
02:14 completenoob: but it's idiomatic in CL to use the toplevel for so much. That's one of the least attractive things abotu it to me
02:14 eli: It's not idiomatic to do that -- it's the only way to work in CL.
02:15 eli: And yes, there's a whole bunch of problem that that leads to...
02:15 completenoob: I tend to use 'labels' but then I haven't really written a whole lot of CL.
02:15 eli: This is a nice horror story: http://fare.livejournal.com/146698.html
02:15 completenoob: sweet
02:16 eli: `labels' won't help that much -- CL makes it too awkward to write large chunks of code this way.
02:17 completenoob: (I've written a lot more Lua than Lisp, btw.)
02:18 eli: (I have almost zero experience with it.)
02:20 completenoob: Lua is just an adorable little language...
02:20 eli: Can you specify some reasons that make it cuter than JS?
02:21 Sgeo: The other day (yesterday I think), I asked something about Racket, someone said that Racket isn't good for long-running programs where things get redefined. Can I have some more information about that, and pointers to Schemes where it's not a bad thing?
02:22 completenoob: metatables, which allow for programmable dispatch like CLOS, smaller language definition, unification of all data structures as associative arrays (which are optimized under the hood with a vector and hash part)
02:22 completenoob: also semi-continuations with coroutines
02:22 completenoob: (basically you could implement call/cc with something like 20 lines of code in the default implementation)
02:23 completenoob: the big win for Lua though is the C_API is very very easy to use
02:23 completenoob: (over JS or Python)
02:24 eli: completenoob: The first sounds like the usual complaint about the prototype classes, the second will probably be gone when JS gets continuations, and the third is probably the deal breaker...
02:24 eli: Sgeo: Racket should be fine with redefining stuff, provided that you know what you're doing...
02:25 Sgeo: Define "know what you're doing". What issues could I come across?
02:25 eli: What are you trying to do, exactly?
02:26 Sgeo: Want to use Racket as a substitute for Smalltalk, I guess
02:26 eli: (To clarify, "know what you're doing" holds in any language that allows runtime evaluation of new code.)
02:26 Sgeo: Ah
02:27 eli: What aspects of smalltalk do you need?
02:28 Sgeo: Not really "need", just thinking about. I think just the comfortable environment, the ability to modify things as it's running
02:29 eli: Sgeo: You mean a continuous REPL, right?
02:29 Sgeo: Yeah
02:29 Sgeo: Maybe the ability to connect to a REPL on a Racket process running remotely
02:29 eli: That's perfectly doable.
02:30 eli: Did you ask a question about it recently on SO?
02:30 completenoob: I forget, does JS have mutable strings? .
02:31 completenoob: (because one thing I think is neat about Lua is that it does not)
02:31 Sgeo: eli, no
02:31 eli: completenoob: I use them so rarely I don't even know...
02:32 eli: Sgeo: Then see this: http://stackoverflow.com/questions/5946380/racket-repl-over-tcp
02:32 Sgeo: ty
02:45 (join) leo2007
02:53 (quit) realitygrill: Quit: realitygrill
02:56 (join) hkBst
03:28 (join) Tabemasu
04:28 (part) Tabemasu
04:50 (join) Andrew_
04:50 (nick) Andrew_ -> Guest97951
04:51 (join) Andrew2_
04:51 (nick) Andrew2_ -> Andrew55
04:52 Andrew55: hi
04:53 Andrew55: I'm having a problem with regex, this line fails to return a result : (regexp-match* "[\\s]+" " ")
04:53 Andrew55: I think it might be a bug in the library
04:56 (quit) Guest97951: Ping timeout: 252 seconds
04:58 (quit) Andrew55: Ping timeout: 252 seconds
05:04 ohwow: ah
05:04 ohwow: he quit :(
05:56 (quit) Nightwolf: Ping timeout: 240 seconds
05:58 (join) Nightwolf
06:37 (join) Demosthenes
06:46 (join) masm
07:05 (quit) leo2007: Quit: rcirc on GNU Emacs 23.3.50.1
07:13 (quit) Demosthenes: Ping timeout: 276 seconds
07:14 (join) Demosthenes
07:15 (quit) masm: Ping timeout: 260 seconds
07:31 (join) masm
07:38 (join) leo2007
08:02 (quit) tildedave: Read error: Connection reset by peer
08:04 (quit) masm: Ping timeout: 252 seconds
08:13 (join) lucian
08:14 (quit) Nightwolf: Ping timeout: 240 seconds
08:15 (join) Nightwolf
08:26 (quit) Nightwolf: Ping timeout: 248 seconds
08:28 (join) MayDaniel
08:34 (quit) MayDaniel: Read error: Connection reset by peer
08:43 (quit) Demosthenes: Quit: leaving
09:05 (join) masm
09:39 (join) Nightwolf
10:26 (join) realitygrill
10:31 (quit) lucian: Remote host closed the connection
10:37 (quit) realitygrill: Read error: Connection reset by peer
10:37 (join) realitygrill
10:54 (join) PLT_Notify
10:54 PLT_Notify: racket: master Kevin Tew * c9c02f3 (2 files in 2 dirs): duplicate and cycle detection for places_deserialize_worker - http://bit.ly/jxYrVv
10:54 (part) PLT_Notify
10:56 (join) mithos28
11:09 (join) ckrailo
11:10 (join) jonrafkind
11:15 (join) lucian
11:17 lucian: is there a racket foundation/org or somesuch? i need to reference something in the racket docs in my dissertation and i'm not sure what entity to ref it to
11:18 jonrafkind: PLT
11:20 (join) jessopher
11:23 (join) anRch
11:28 lucian: jonrafkind: PLT foundation? thanks
11:28 jonrafkind: its not foundation, i forget what its supposd to be. Probably something like PLT Inc.
11:29 (nick) elliottcable -> ec|detached
11:29 bremner_: still called PLT Scheme Inc. according the web site
11:29 (join) dnolen
11:30 lucian: bremner_: i see. thanks
11:37 (quit) Sgeo: Read error: Connection reset by peer
11:39 (quit) realitygrill: Read error: Connection reset by peer
11:39 (join) realitygrill
11:40 lucian: btw if anyone is interested, i'm referencing racket's reader docs as an example of very nice lisp reader docs
11:43 (quit) realitygrill: Read error: Connection reset by peer
11:43 (join) realitygrill_
11:53 (quit) anRch: Quit: anRch
11:53 (nick) samth_away -> samth
11:57 samth: lucian, if you need to cite Racket, see here: http://racket-lang.org/tr1
11:58 lucian: samth: does that go for racket's docs too?
11:59 (join) anRch
11:59 samth: yes
11:59 lucian: i see. thanks
11:59 samth: although if you're citing some specific bit of research, a citation to the relevant paper is appropriate
12:00 samth: that page has instructions on the right way to cite a specific section in the racket reference
12:01 samth: lucian, what's your dissertation on?
12:02 lucian: samth: building mobile applications directly from mobile devices
12:02 samth: cool
12:03 lucian: meh, it ended up being rather lame :)
12:03 lucian: think app inventor, on html&js and very simple
12:05 lucian: heh, 64 references
12:08 lucian: any ideas on articles treating exploratory programming? i think my markers might not be familiar with it
12:21 (quit) mithos28: Quit: mithos28
12:21 (quit) hkBst: Remote host closed the connection
12:29 (quit) anRch: Quit: anRch
12:38 (join) sstrickl
12:41 samth: hmm
12:42 samth: not sure
12:42 samth: might look at johnathan edwards
13:03 (quit) EM03: Quit: EM03
13:03 (quit) completenoob: Quit: Page closed
13:12 (join) Fare
13:15 (join) mithos28
13:24 (quit) sstrickl: Quit: sstrickl
13:24 (join) sstrickl
13:25 (join) EM03
13:25 (quit) EM03: Changing host
13:25 (join) EM03
13:47 (join) anRch
13:49 (join) PLT_Notify
13:49 PLT_Notify: racket: master Stephen Bloch * 7597d2c (1 files in 1 dirs): Fixed a typo that crashed the nightly build :-( - http://bit.ly/mdXpd6
13:49 (part) PLT_Notify
14:20 lucian: samth: dodged it in the end. thanks for all the help
14:21 samth: happy to help
14:21 samth: i spent a bunch of time on the bib for my dissertation, so i know the feeling
14:54 (quit) anRch: Quit: anRch
14:58 (quit) jessopher: Remote host closed the connection
14:58 (join) jessopher
15:01 (quit) jessopher: Remote host closed the connection
15:01 (join) jessopher
16:23 (quit) leo2007: Ping timeout: 260 seconds
16:32 (join) Sgeo
16:32 Sgeo: Racket's printf/format doesn't have an equivalent for Common Lisp's format's ~{ and ~}?
16:33 jonrafkind: nah
16:36 Sgeo: Um
16:36 Sgeo: Does Racket's Make Executable generally just include the Racket environment?
16:36 Sgeo: http://pastie.org/1894499 resulted in a 4.5 MB .exe
16:36 (join) tauntaun
16:39 Sgeo: And the version for distribution results in a .zip file, which contains a .exe and some .dlls
16:40 jonrafkind: yea
16:40 jonrafkind: the exe is the racket binary + some collects + your bytecode
16:41 Sgeo: collects?
16:41 jonrafkind: the racket libraries
16:41 jonrafkind: the directory that contains them is called 'collects'
16:41 jonrafkind: its a 'collection' of code
16:44 (join) yoklov
16:45 (join) yoklov1
16:45 (quit) yoklov: Read error: Connection reset by peer
16:54 (quit) jessopher: Ping timeout: 252 seconds
17:25 (quit) Sgeo: Ping timeout: 246 seconds
17:48 (quit) sstrickl: Quit: sstrickl
17:57 EM03: Hello everyone
18:03 (quit) yoklov1: Quit: Leaving.
18:11 (quit) dnolen: Ping timeout: 252 seconds
18:12 (quit) realitygrill_: Quit: realitygrill_
18:12 (join) yoklov
18:18 (quit) Fare: Remote host closed the connection
18:29 (join) Fare
18:37 (quit) mithos28: Quit: mithos28
18:40 (quit) masm: *.net *.split
18:44 (join) masm
18:46 (join) mithos28
18:47 (quit) lucian: Remote host closed the connection
18:50 (join) lucian
18:51 (quit) mithos28: Client Quit
18:53 (quit) Fare: Quit: Leaving
19:13 (quit) masm: Quit: Leaving.
19:18 (quit) DT``: Ping timeout: 240 seconds
19:19 (join) DT``
19:19 (join) dnolen
19:23 (nick) samth -> samth_away
19:25 (join) realitygrill
19:30 (quit) lucian: Remote host closed the connection
19:33 EM03: Documentation question: blog : (listof post?) Would this be legal syntax / some sort of definition or is that just the docs breaking down what the result of blog should be?
19:35 DT``: it's valid, it returns a contract that accepts a list of `post?'.
19:38 DT``: (if you're talking about the infix :, it's not legal syntax, but you can enforce types with contracts)
19:40 EM03: (struct post (title body))
19:40 EM03: title : string?
19:40 EM03: body : string?
19:40 EM03: doing this returns errors
19:41 DT``: yes, I misunderstood the question.
19:42 EM03: http://docs.racket-lang.org/continue/index.html?q=file/gif
19:42 EM03: those examples in blue
19:42 EM03: what is the purpose just to show what you want ....or is that actual legal syntax
19:42 DT``: just to show what you want.
19:43 EM03: its almost like its being presented as legal syntax
19:43 DT``: I thought you were talking about `(listof post?)'.
19:43 EM03: well thats a blue example as well
19:43 EM03: and it looks like title and body should be a string but if that is not legal syntax how is the program as they are showing it doing just that?
19:45 DT``: well, there are contracts to enforce types on procedures (and constructors).
19:45 DT``: but no, the `x : t' thing is not legal.
19:46 EM03: I'm curious what the purpose was
19:46 EM03: it seems like its being presented as legal syntax
19:47 DT``: well, probably to just say that that should be a string.
19:47 EM03: hmm i guess so
19:47 EM03: but to a new user it easily looks like thats some sort of code to enforce it....which it should be enforced but if that is not doing it then yea it can be anything as is
19:49 DT``: those signatures are almost everywhere in the docs, they are there just to show the types.
19:49 DT``: runnable code never appears inside the blue boxes afaik.
19:49 EM03: thanks for the helpful note
19:49 EM03: I assumed so but then I was like ok maybe its trying to enforce it as a string here
20:03 (join) mithos28
20:10 (join) PLT_Notify
20:10 PLT_Notify: racket: master Eli Barzilay * 387b6c5 (3 files in 3 dirs): Added Mac OS X build for x86_64. ... - http://bit.ly/mcBFTt
20:10 (part) PLT_Notify
20:12 EM03: DT``: how long have you been using racket yourself?
20:14 DT``: dunno, 3 months?
20:16 (quit) ckrailo: Quit: Computer has gone to sleep.
20:22 (join) rpr
20:31 EM03: DT``: seems like most people have either used scheme for a very short time or a long time. I was on a campus today and the teacher was teaching scheme with MIT scheme using edwin .....racket probably would have been a better idea
20:33 (quit) jonrafkind: Ping timeout: 248 seconds
20:35 DT``: EM03, using SICP? if so, it was (IMHO) a good idea, since it has everything SICP requires built-in (e.g. cons-stream). of course, any Scheme{,-like} system would do too, but I'm of the idea that starting with a small Scheme with just conses, lists, symbols and basics will teach you better habits (to not abuse of features).
20:36 DT``: of course then you would move to something more complete when doing real-world stuff.
20:36 EM03: but i bet most of those people will never even find out about racket
20:37 DT``: if they'll get interested in Scheme, they surely will.
20:45 EM03: hopefully
20:48 EM03: some professors seem to teach scheme as a language that you will never use again after they are done with it in the class
20:52 DT``: well, it's not exactly as widespread as java or c#.
20:52 DT``: but it has the potential (just like, you know, any other language other than java and c#)
20:53 EM03: the non standard functional nature of the langauge makes picking it up much more difficult.
20:54 (quit) mithos28: Quit: mithos28
20:54 DT``: it all reduces to parens, EM03. there's CL for a non-strictly-functional-style Lisp.
20:56 (join) mithos28
21:05 (quit) mithos28: Quit: mithos28
21:23 bremner_: DT``: scheme/racket is hardly strictly function. Mutation is quite natural.
21:25 bremner_: but, I guess it is true that CL is even less strictly functional in style
21:26 bremner_: although that is just hearsay for me.
21:27 DT``: bremner_, FP in Scheme is natural enough that you can switch to immutable conses without having to rewrite *everything*.
21:27 DT``: but, of course, this is not haskell, we have set-*!
21:27 bremner_: so what is missing from CL that makes it harder? or is it just a culture thing?
21:28 DT``: I guess it's a culture thing, and CL's funcall.
21:28 bremner_: yeah, I guess haskell is my reference when people talk about purely functional languages (although of course it has unsafeLaunchMissiles)
21:29 bremner_: right, I guess lisp-2 makes it a bit more clunky, but it's surprising that makes such a difference
21:30 bremner_: I guess libraries can also make a big difference
21:30 DT``: well, compare: (λ (f) (λ (x) (f (f (f x))))) vs (λ (f) (λ (x) (funcall f (funcall f (funcal f x))))).
21:30 DT``: yeah, libraries too.
21:30 DT``: *funcall
21:50 (join) Sgeo
21:51 Sgeo: Does DrRacket not have any indication of the arguments for the function currently being typed in?
21:51 Sgeo: The way SLIME does for Common Lisp?
22:03 rpr: geiser does.  geiser is slime for racket/guile.  Its really pretty darn good.
22:11 Sgeo has been relying on LispBox and ClojureBox. I should probably just install a full emacs at this point
22:12 (quit) tauntaun: Quit: Ex-Chat
22:15 Sgeo: Does Racket have a way to distinguish between optional arguments that were provided with a value and optional arguments that weren't, the way CL does?
22:17 (join) Demosthenes
22:19 Demosthenes: yoklov: hey! i did it ;]
22:20 bremner_: Sgeo: maybe you want case-lambda?
22:20 Sgeo: That wouldn't work with keyword arguments
22:20 bremner_: nope.
22:21 yoklov: yeah?
22:22 yoklov: what'd you end up doing?
22:23 Sgeo: Well, Practical Common Lisp has an example where if you provide some searching function with a true or false, then it will look for records that have that property, but if you don't provide it, it won't care
22:25 Sgeo: "Note that you need to use a three-item list to specify the keyword parameter ripped because you need to know whether the caller actually passed :ripped nil, meaning, "Select CDs whose ripped field is nil," or whether they left out :ripped altogether, meaning "I don't care what the value of the ripped field is.""
22:25 Sgeo: http://www.gigamonkeys.com/book/practical-a-simple-database.html
22:28 Demosthenes: yoklov: got that parser going
22:28 yoklov: yeah?
22:28 yoklov: right
22:28 Demosthenes: was trying to figure out if i ought to try and modularize it and share
22:28 DT``: Sgeo, I'd do (define my-secret-optarg (gensym)) and compare to that.
22:29 DT``: with eq?.
22:29 Sgeo: Ah
22:29 Sgeo: Wait, what? How would you get that entirely inside one define?
22:29 yoklov: i mean if you think it's useful and good then yeah i'd say you probably should
22:30 DT``: (define my-function (let ((optarg (gensym))) (lambda ((opt optarg)) (unless (eq? opt optarg) (display "and nothing of value was passed here"]
22:31 DT``: s/unless/when/
22:31 Sgeo: Ah
22:31 Sgeo: ty
22:32 Demosthenes: yoklov: http://bazaar.launchpad.net/~rladams/+junk/RacketParser/files
22:33 Demosthenes: the unit tests roc
22:33 Demosthenes: k
22:33 Sgeo: "; you need to get all your code in order, make sure it'll run, then execute. And that's it. If you need to make changes (like, while developing web apps) you need to tweak the code, then restart the server, then re-navigate to the page you were just on because it auto-generates new urls each time."
22:33 Sgeo: [about Racket's web stuff]
22:34 Demosthenes: yoklov: if you have suggestions, i'm open to criticism ;]
22:34 Sgeo: Is there no way around that? That's the sort of thing I was trying to get at a while ago: Can Geiser prevent you from needing to do that?]
22:36 yoklov: it looks very good
22:36 yoklov: no suggestions I can think of, to be honest
22:36 Demosthenes: the tests rocked.
22:36 Demosthenes: found some great edge cases
22:36 yoklov: yeah?
22:37 yoklov: awesome
22:37 (quit) rpr: Read error: Connection reset by peer
22:37 yoklov: i like rackunit a lot
22:37 Demosthenes: i'm planning on adding validators and pre/post processors per line next, yeehaw.
22:39 Demosthenes: wish i could find a better logging solution
22:39 Demosthenes: i don't like the built in logger, even when you can change it dynamically between runlevels
22:40 Sgeo pokes
22:41 yoklov: wait, for what?
22:41 Demosthenes: logging, for everything racket
22:42 yoklov: i see, it's never actually been a thing i've done a ton of
22:42 yoklov: to be honest
22:43 Demosthenes: i'd really like a logger with multiple logging levels, function specific logging, dynamic level changes, output options...
22:44 yoklov: hm, yeah that does sound pretty resonable, I've just never come into a situation like yours where i'd need/want to log that much
22:55 Demosthenes: logging is essential... thats one of the issues with parsing with regexp, its hard to debug why something didnt work
22:55 Demosthenes: perl you can turn on a debug mode where every regexp operation is verbosely dumped, but i've never found that in racket
23:10 ASau: eli: I remember it, but I've got no time to deal with it.
23:11 (quit) ASau: Quit: off
23:16 (join) leo2007
23:19 (nick) ec|detached -> elliottcable
23:36 (join) mithos28
23:38 EM03: blog : (listof post?) how would you write this? this is obviously not legal syntax but wouldn't you need to pass blog to listof as well?
23:39 EM03: or does it only check if the list derives from a struct in this case post?
23:40 DT``: EM03, http://docs.racket-lang.org/guide/contracts-struct.html?q=contract%20struct&q=contract
23:40 rudybot: http://tinyurl.com/3qysedh
23:41 EM03: its funny that the docs are implying that the code should have these contracts but yet I don't see any of them
23:43 DT``: it's more a simple annotation (just for clarity) than something enforced there.
23:44 EM03: I keep thinking that with those blue examples that since its so in depth that it should function that way as well
23:44 EM03: oh well maybe I'm over analyzing
23:53 (quit) mithos28: Quit: mithos28