00:03 eli: Fare: So I'm not following what you said...
00:04 eli: If you're just talking about that processor knowing that `foo' is defined as (listof integer?) (or whatever compiles to that), then you can see that in my code -- `foo' is bound to a macro-level thing, and its value has the original expression.
00:05 eli: So that supposed type language compiler would use the same `syntax-local-value' thing to know about type definitions.
00:23 Fare: where is that syntax-local-value defined?
00:23 eli: Fare: It's part of the core language.
00:24 eli: Roughly speaking, its a way to get the value of a syntax binding, which is normally a transformer function which is used only as an expander.
00:24 eli: rudybot: (syntax-local-value #'when)
00:24 rudybot: eli: your sandbox is ready
00:24 rudybot: eli: error: syntax-local-value: not currently transforming
00:24 eli: rudybot: (begin-for-syntax (syntax-local-value #'when))
00:24 rudybot: eli: Done.
00:24 eli: Meh.
00:25 eli: rudybot: (define-syntax (show-stx stx) (syntax-case stx () [(_ id) (eprintf ">>> ~s -> ~s\n" (syntax-e #'id) (syntax-local-value #'id)) #'(void)]))
00:25 rudybot: eli: Done.
00:25 eli: rudybot: (show-stx when)
00:25 rudybot: eli: ; stderr: ">>> when -> #<procedure:...define-et-al.rkt:35:4>\n"
00:25 eli: Fare: There.
00:26 eli: That #<procedure...> thing is the expander for `when'.
00:26 eli: Now, the thing is that you can use `define-syntax' with random values, which is used only as value holders, not as macros.
00:26 eli: rudybot: (define-syntax fooo "what?")
00:26 rudybot: eli: Done.
00:26 eli: rudybot: fooo
00:26 rudybot: eli: error: #<undefined>:1:0: fooo: illegal use of syntax in: fooo
00:27 eli: That's because it's not an expander function.
00:27 eli: rudybot: (show-stx fooo)
00:27 rudybot: eli: ; stderr: ">>> fooo -> \"what?\"\n"
00:27 eli: And that took its value.
00:32 veer: ok silly question follows: so invariant(in context of PL)  means some condition that holds before function call and after function call right? And I can express this using racket contracts right? I can also use normal racket expression to do so for eg.  https://gist.github.com/3022279
00:33 offby1: I don't think you can use contracts to express invariants.
00:34 veer: I see , so what contracts are used for ?
00:35 offby1: I use 'em to check preconditions and postconditions
00:40 (quit) jrslepak: Read error: Operation timed out
00:40 veer: So in my case precondition is same as postcondition , so it this a case of something being invariant ? I'll look into contracts in manual later
00:47 (join) jrslepak
00:49 (quit) jeapostrophe: Ping timeout: 250 seconds
00:54 (quit) siduction: Quit: Leaving
01:22 (quit) mmalorni: Ping timeout: 246 seconds
01:31 (quit) jacius: Quit: Leaving
01:31 (quit) rbarraud: Read error: Connection reset by peer
02:20 (quit) Kaylin: Quit: Leaving.
03:06 (join) mceier
03:09 (join) antithesis
03:18 (join) jesyspa
03:19 (quit) jonrafkind: Ping timeout: 240 seconds
03:32 (join) dalaing
04:12 (join) dalaing1
04:14 (quit) dalaing: Read error: Connection reset by peer
04:16 (join) snearch
04:16 (join) zaamkablam
04:32 (join) chemuduguntar
04:49 (quit) zaamkablam: Ping timeout: 240 seconds
04:49 (quit) dalaing1: Quit: Leaving.
04:55 (quit) chemuduguntar: Read error: No route to host
04:55 (join) dalaing
04:59 (quit) ssbr__: Ping timeout: 246 seconds
05:00 veer: allright! understood how pre and post conditions  helps , example https://gist.github.com/3022699 , without proof. So basically I only need to prove pre-cond => post-cond to say function is correct for post-cond
05:01 veer: I like this idea of specifying pre and post conds and including it in program
05:08 (quit) jesyspa: Quit: Lost terminal
05:10 (quit) antithesis: Quit: antithesis
05:14 (join) neilv
05:14 (join) ssbr__
05:15 neilv: ">sgl/gl.html#(def._((lib._sgl/gl..rkt)._.G.L_.C.O.M.P.R.E.S.S.E.D_.I.N.T.E.N.S.I.T.Y))"
05:21 (join) jesyspa
05:30 (quit) dalaing: Read error: Connection reset by peer
05:35 (join) masm
05:41 (quit) jyc: Read error: Connection reset by peer
05:45 (quit) masm: Quit: Leaving.
05:49 (quit) Fare: Ping timeout: 240 seconds
06:01 (join) rbarraud
06:04 (join) Fare
06:07 (join) bitonic
06:14 (quit) rbarraud: Read error: Connection reset by peer
06:19 (join) rbarraud
06:29 (join) bitonic`
06:30 (quit) bitonic: Ping timeout: 246 seconds
06:34 (join) bitonic``
06:35 (quit) bitonic`: Ping timeout: 246 seconds
06:51 (quit) rbarraud: Ping timeout: 255 seconds
06:54 (quit) ssbr__: Ping timeout: 264 seconds
06:56 (quit) jesyspa: Quit: leaving
06:59 (nick) bitonic`` -> bitonic
07:11 (quit) Fare: Ping timeout: 246 seconds
07:32 (join) antithesis
07:35 (quit) veer: Ping timeout: 246 seconds
07:45 (join) soegaard
07:50 (join) veer
07:58 (quit) Shviller: Ping timeout: 246 seconds
07:59 (join) Shviller
08:03 (quit) snearch: Quit: Verlassend
08:16 (quit) antithesis: Quit: antithesis
08:21 (join) snearch
08:41 (join) Kaylin
08:53 (quit) neilv: Quit: Leaving
08:55 (join) anttih
09:14 (join) jeapostrophe
09:17 (join) bitonic`
09:22 anttih: what are the performace characteristics of Racket's immutable hash tables? Does modification copy the whole table or is the old structure reused?
09:25 (join) anRch
09:27 anttih: I assume it is reused as much as possible.
09:33 (quit) snearch: Quit: Verlassend
09:46 (join) ssbr__
10:00 soegaard: Well, since it is immutable it can't really be modified. However one can make new hash tables and the new hash tables shares some of its internal structure with the old (but this ought not be observable from the outside)
10:01 (join) jao
10:02 (quit) jao: Changing host
10:02 (join) jao
10:03 (quit) anRch: Quit: anRch
10:27 (quit) Kaylin: Quit: Leaving.
10:32 (join) _p4bl0`
10:35 (join) jeremyheiler
10:39 (quit) _p4bl0`: Remote host closed the connection
10:39 (join) _p4bl0``
10:47 (quit) jeapostrophe: Ping timeout: 246 seconds
10:47 (part) _p4bl0``: "ERC Version 5.3 (IRC client for Emacs)"
10:54 (quit) wahjava: Quit: You've got to hide your love away
10:58 (join) wah_java
11:04 (quit) BeLucid_: Ping timeout: 248 seconds
11:05 (nick) wah_java -> wahjava
11:30 (join) jeapostrophe
11:30 (quit) jeapostrophe: Changing host
11:30 (join) jeapostrophe
11:43 (quit) jeapostrophe: Ping timeout: 265 seconds
11:52 (quit) veer: Quit: Leaving
11:56 (join) wah_java
11:59 (join) jonrafkind
11:59 (quit) jonrafkind: Changing host
11:59 (join) jonrafkind
12:06 (join) Fare
12:09 (quit) jonrafkind: Ping timeout: 244 seconds
12:10 (quit) wah_java: Quit: Machines are always better than humans
12:17 (quit) jeremyheiler: Quit: Computer has gone to sleep.
12:50 (quit) lcc: Quit: ERC Version 5.3 (IRC client for Emacs)
12:51 Fare: eli: do you have a protocol so that modules can add their own sauce to the client's whole-module expansion?
13:02 (join) wahjava
13:23 (quit) jao: Ping timeout: 246 seconds
13:31 eli: Fare: What do you mean?
13:34 Fare: do you ever find a need for a module to extend the language in such a way as to require global effects? Can you do it in modular ways, or only as "this module controls the entire language, and others can only have local effects"?
13:42 eli: Fare: There are several answer to that.
13:42 eli: One thing is that each module's toplevel gets to run when the module is instantiated.
13:42 eli: (Which is only once, unless you're doing sandboxing or macros etc.)
13:43 eli: So you could combine toplevel side-effect things by just requiring modules.
13:44 eli: More than that, it used to be the case that the instantiation order was not defined, but recently (IIRC) it was changed so that (require foo bar) will instantiate them in order.
13:44 eli: But since they're used for linking, the required modules get instantiated before the requiring module does.
13:44 eli: rudybot: (module A racket (printf "A\n"))
13:44 rudybot: eli: Done.
13:44 eli: rudybot: (module B racket (printf "B\n"))
13:44 rudybot: eli: Done.
13:45 eli: rudybot: (module C (printf "C\n") (require 'A 'B))
13:45 rudybot: eli: error: #<undefined>:1:10: module: initial import is not a well-formed module path at: (printf "C\n") in: (module C (printf "C\n") (require (quote A) (quote B)))
13:45 eli: rudybot: (module C racket (printf "C\n") (require 'A 'B))
13:45 rudybot: eli: Done.
13:45 eli: rudybot: (require 'C)
13:45 rudybot: eli: ; stdout: "B\nA\nC\n"
13:45 eli: Fare: So in the nightly build version, that would print A and then B for that change,
13:46 eli: and C is always printed last since it needs to finish requiring the other two before it is invoked.
13:46 eli checks
13:46 eli: Yeah.
13:47 eli: So that's one side of the story.
13:47 eli: Another side of the story is when you want a module to behave in two ways, one when it's being used as a library, and another when it's being run directly.
13:47 eli: This is also related to a recent development which we finally have now --
13:47 eli: the new feature is submodules
13:48 eli: So a module can now have submodules, and if it is being run directly (ie, from the command line, or from the racket executable), then it will see if it has a submodule that is called `main'
13:48 eli: and if so, then it will invoke that module.
13:48 (nick) bitonic` -> bitonic
13:49 eli: So code that is put in a `main' submodule usually gets to run only when you run the module directly.
13:49 eli: (Or if you invoke it explicitly in code, of course.)
13:49 eli: Yet another aspect of this, which might be relevant to you, is dealing with languages that need to add some toplevel-ish behavior to modules.
13:50 eli: The classic example for this is a module that provides a `test' macro to write tests,
13:50 eli: and you'd usually want to do some (report-test-results) at the end of the module whos using your test facility.
13:50 eli: For that, there is one of the popular special macros -- `#%module-begin'.
13:50 (join) lcc
13:51 eli: That macro gets implicitly wrapped around the module's body,
13:51 eli: so it can be used to add top-level things to it, before, after, or even process each toplevel expression.
13:51 eli: rudybot: reset
13:51 rudybot: eli: error: reference to an identifier before its definition: reset in module: 'program
13:51 eli: rudybot: init
13:51 rudybot: eli: your sandbox is ready
13:53 eli: rudybot: (module A racket (provide (except-out (all-from-out racket) #%module-begin) (rename-out [module-begin #%module-begin])) (define-syntax-rule (module-begin E ...) (#%module-begin (printf "Starting\n") (begin (printf "Running ~s\n" 'E) E) ... (printf "Done\n"))))
13:53 rudybot: eli: Done.
13:53 eli: rudybot: (module B 'A (define x 1) (printf "x = ~s\n" x) (printf "this is not the last printout\n"))
13:53 rudybot: eli: Done.
13:54 eli: rudybot: (require 'B)
13:54 rudybot: eli: ; stdout: "Starting\nRunning (define x 1)\nRunning (printf \"x = ~s\\n\" x)\nx = 1\nRunning (printf \"this is not the last printout\\n\")\nthis is not the last printout\nDone\n"
13:54 eli: Fare: Like that.
13:54 eli: (Which would make more sense if you type it in and run it as usual...)
13:54 (join) Kaylin
14:02 (join) jacius
14:03 (join) dnolen
14:12 (part) weirdo
14:15 (quit) lcc: Quit: ERC Version 5.3 (IRC client for Emacs)
14:30 eli: Fare: Did that make sense?
14:31 Fare: eli: is #%module-begin shadowed for A? for users of A?
14:32 eli: If you use A like that, (module foo 'A ...blah...)
14:32 eli: then this module uses the `#%module-begin' binding from A.
14:32 Fare: How are such things chained between modules?
14:32 Fare: is there a module-around to do bindings?
14:33 eli: The chaining is simple -- note that A's `#%module-begin' eventually expands to its own `#%module-begin'...
14:33 eli: So it just adds a layer on top of what its own thing is doing.
14:33 eli: BTW, there's one important bit there,
14:34 eli: you can usually write something like this instead:
14:34 eli: (module foo racket (require 'A) ...blah...)
14:34 eli: and that gets you all of A's bindings instead of the ones from racket.
14:34 eli: But it won't work in this case, since `#%module-begin' is used around the body, so by the time it gets to the require it already had used the one from racket.
14:35 Fare: what if I require 'A and require 'B, and each want their new module thingie?
14:35 eli: That's one reason for the special "initial language module" thing in modules.
14:35 eli: Technically, there is no problem...
14:35 eli: Since there is only one initial language position.
14:35 eli: So, if you try (module foo racket (require 'A 'B))
14:36 eli: and both provide a binding by the same name, you get an error.
14:37 eli: If you put it in the initial language thing, then later requires and definitions are allowed to shadow it, but in the case of `#%module-begin' that doesn't make any difference since it has already been used by the time it gets to these requires and definitions.
14:37 eli: So if you really want to combine `#%module-begin' from two modules, the best is to write a third one that uses both.
14:37 eli: Ie, something like
14:37 eli: (require (prefix-in A: 'A) (prefix-in B: 'B))
14:37 eli: and then
14:38 eli: (define-syntax-rule (module-begin E ...) (A:#%module-begin (B:#%module-begin E ...)))
14:38 eli: (provide (rename-out [module-begin #%module-begin]))
14:38 eli: Like that.
14:40 Fare: ok, so the initial language is special and is the only one to do global stuff
14:40 Fare: what about special bindings?
14:40 Fare: how do I introduce a parameterize around the whole module?
14:40 eli: Roughly, yes.  Global in the sense of can affect how the module expands.
14:40 eli: WHat do you mean?
14:41 eli: `parameterize' -- ?
14:41 Fare: yes
14:41 eli: It doesn't make sense to `parameterize' around a module's body, because it's a dynamic-extent thing...
14:41 Fare: and?
14:41 eli: So I'm not following you...
14:42 Fare: in lisp, I often bind special variables around the compilation of a source file.
14:42 eli: IOW, the parameterize would have an effect only on toplevel expressions, but it would be ignored for functions, since they're used after the module was involed.
14:43 eli: Wait, you mean special variable that control compilation?
14:43 Fare: yes
14:43 eli: Oh, that would be like using parameters at the syntax level.
14:43 eli: And there you can knock youself out...
14:43 Fare: but around the whole module
14:43 eli: Specifically, you *don't* need special variables in the parameters sense.
14:44 Fare: ?
14:44 eli: You can just have some global bindinging and `set!' it.
14:44 eli: The resulting mutation is localized to each use, since when your module is used to compile a new one, it will be re-instantiated, with a new state.
14:47 (quit) mceier: Remote host closed the connection
14:48 (join) mceier
14:59 (join) jeapostrophe
14:59 (quit) jeapostrophe: Changing host
14:59 (join) jeapostrophe
15:05 (quit) dnolen: Ping timeout: 252 seconds
15:14 (join) RacketCommitBot
15:14 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/dJrRPw
15:14 RacketCommitBot: [racket/master] racket/class: doc fixes and improvements - Asumu Takikawa
15:14 (part) RacketCommitBot
15:20 (quit) Kaylin: Quit: Leaving.
15:24 (quit) maahes: Ping timeout: 240 seconds
15:27 (quit) soegaard: Quit: soegaard
15:41 (join) jeapostr1phe
15:41 (quit) jeapostrophe: Read error: Connection reset by peer
15:44 (join) maahes
15:58 (join) jsnikeris
16:11 (join) antithesis
16:31 (quit) antithesis: Quit: antithesis
16:32 (join) soegaard
16:44 (join) jonrafkind
16:52 (join) MayDaniel
17:07 (join) jao
17:07 (quit) jao: Changing host
17:07 (join) jao
17:40 (quit) wahjava: Quit: Machines are always better than humans
17:43 (quit) MayDaniel: Read error: Connection reset by peer
17:47 (quit) jeapostr1phe: Ping timeout: 246 seconds
18:00 (join) BeLucid
18:08 (quit) jacius: Ping timeout: 276 seconds
18:20 (join) jacius
18:36 (join) anRch
18:37 (quit) Myk267: Remote host closed the connection
18:53 (join) jeapostrophe
19:01 (quit) soegaard: Quit: soegaard
19:03 (quit) jeapostrophe: Ping timeout: 252 seconds
19:15 (quit) Fare: Ping timeout: 244 seconds
19:26 (quit) anRch: Quit: anRch
20:07 (join) jyc
20:15 (join) rbarraud
20:15 (quit) stchang: Read error: Operation timed out
20:18 (join) stchang
20:34 (join) asdfhjkl
20:39 (join) adu
20:43 (quit) jonrafkind: Ping timeout: 240 seconds
20:44 (quit) stchang: Read error: Operation timed out
20:46 (join) stchang
20:56 (quit) jao: Remote host closed the connection
20:59 (join) jao
20:59 (quit) jao: Changing host
20:59 (join) jao
21:10 (join) jeapostrophe
21:10 (quit) jeapostrophe: Changing host
21:10 (join) jeapostrophe
21:10 (join) Myk267
21:10 (quit) adu: Quit: adu
21:30 (quit) asdfhjkl: Quit: Leaving
21:31 (join) asdfhjkl
21:56 (quit) jeapostrophe: Ping timeout: 244 seconds
22:16 (join) Kaylin
22:19 (quit) vu3rdd: Read error: Connection reset by peer
22:20 (join) vu3rdd
22:20 (quit) vu3rdd: Changing host
22:20 (join) vu3rdd
22:58 (quit) jao: Remote host closed the connection
23:02 (join) chimeracoder
23:03 chimeracoder: For typed/racket, since round and truncate don't return Integer, what's the best way to create an integer after division?
23:03 chimeracoder: (I'm getting type errors even on an Exact Integer)
23:04 (quit) asdfhjkl: Quit: Leaving
23:06 offby1: (inexect->exact) maybe
23:06 offby1: don't really know; I don't use typed racket
23:14 chimeracoder: offby1: well, no, unfortunately the code that's causing the error is (inexact->exact (round (* foo 1.05946))) bar)
23:24 (quit) ivan\: Quit: ERC Version 5.3 (IRC client for Emacs)
23:24 (join) adu
23:28 chimeracoder: Ah, got it - for the record, you have to convert to an exact form *before* rounding
23:28 chimeracoder: http://www.digipedia.pl/usenet/thread/14667/1904/
23:28 chimeracoder: ie,  (round (inexact->exact (truncate (* 43 1.05946))))
23:30 asumu: chimeracoder: in the next version of Racket, there's an exact-round function built-in which will give you an Integer too.
23:30 chimeracoder: asumu: ooh, thanks for letting me know
23:30 (join) jeapostrophe
23:30 (quit) jeapostrophe: Changing host
23:30 (join) jeapostrophe
23:52 (join) Fare
23:56 (join) ivan\