00:06 (quit) Demosthenex: Ping timeout: 250 seconds
00:08 (join) RacketCommitBot
00:08 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/x_Vwtg
00:08 RacketCommitBot: racket/master bc5e811 Matthew Flatt: dist-spec fix...
00:08 (part) RacketCommitBot
00:21 (join) Demosthenex
00:27 (quit) Demosthenex: Remote host closed the connection
00:27 (quit) mithos28: Quit: mithos28
00:28 (join) Demosthenex
00:29 (join) mithos28
00:30 lewis1711: (struct internal (parent slots [depth #:auto])
00:30 lewis1711: #:auto-value (add1 [parent '_depth]) ...
00:30 lewis1711: getting an error because "parent" is unbound here. what would be away around this?
00:31 mithos28: auto values are the same for all structure values of a given structure type
00:31 mithos28: I don't think they are what you want
00:31 mithos28: you probably want a smart constructor
00:31 lewis1711: ah
00:31 mithos28: just write a function internal* which calculates depth for you
00:31 mithos28: and then on export rename it
00:32 lewis1711: that seems like a sane solution. thanks
00:34 (quit) Demosthenex: Ping timeout: 252 seconds
00:36 (join) Demosthenex
00:44 (join) sw2wolf
00:59 (join) mizu_no_oto
01:13 (quit) jonrafkind: Ping timeout: 252 seconds
01:19 mithos28: is there a way to make xrepl not hide the backtrace of error messages
01:42 (join) jonrafkind
01:45 (quit) jeapostrophe: Read error: Operation timed out
02:02 lewis1711: is there a way to bind a symbol with a define?
02:02 mithos28: what do you mean by that?
02:02 lewis1711: given 'a, give me an object a equal to 3
02:03 mithos28: eval?
02:03 mithos28: eval is usually not a good answer though
02:04 lewis1711: some macro that takes 'a
02:04 lewis1711: and gives  me (define a 3)
02:04 lewis1711: or does that, with a begin form maybe
02:04 mithos28: rudybot: (define-syntax-rule (f x) (define x 3))
02:04 rudybot: mithos28: Done.
02:04 mithos28: rudybot: (f y)
02:04 rudybot: mithos28: Done.
02:04 mithos28: rudybot: y
02:04 rudybot: mithos28: ; Value: 3
02:05 mithos28: is that what you want?
02:05 lewis1711: that wouldn't work with 'y
02:06 lewis1711: what is the type of y? that's probably where I am getting confused
02:07 mithos28: rudybot: (define-syntax (g stx) (syntax-case stx (quote) ((_ 'x) (define x 3))))
02:07 rudybot: mithos28: error: begin (possibly implicit): no expression after a sequence of internal definitions in: ((define x 3))
02:07 mithos28: rudybot:  (define-syntax (g stx) (syntax-case stx (quote) ((_ 'x) #'(define x 3))))
02:07 rudybot: mithos28: Done.
02:07 mithos28: rudybot: (g 'a)
02:07 rudybot: mithos28: Done.
02:07 mithos28: a
02:07 mithos28: rudybot: a
02:07 rudybot: mithos28: ; Value: 3
02:07 mithos28: Is that what you want
02:08 lewis1711: yes! alright that's a tiny enough macro that I should figure out the nuts and bolts of it
02:08 mithos28: so 'y is read as (quote y)
02:09 mithos28: the second part of syntax-case says which words are literals
02:09 mithos28: ie quote is used for matching against specificly an identfier with the binding of quote
02:10 mithos28: rudybot: (g (quote b))
02:10 rudybot: mithos28: Done.
02:10 mithos28: rudybot: b
02:10 rudybot: mithos28: ; Value: 3
02:11 lewis1711: so... y is the syntax
02:11 lewis1711: and (quote) tells syntax case that it's coming in quoted form?
02:11 mithos28: no
02:11 mithos28: quote is not magic at all
02:12 mithos28: rudybot: (define-syntax (h stx) (syntax-case stx (list) ((_ (list x)) #'(define x 3))))
02:12 rudybot: mithos28: Done.
02:12 mithos28: rudybot: (h (list q))
02:12 rudybot: mithos28: Done.
02:12 sw2wolf: What's corresponding of #'multi-value-bind in racket ?
02:12 mithos28: rudybot: q
02:12 rudybot: mithos28: ; Value: 3
02:12 mithos28: sw2wolf: I don't know common lisp. Can you explain what that does?
02:13 sw2wolf: bind multiple value
02:14 mithos28: rudybot: (let-values (((a b) (values 1 2))) (+ a b))
02:14 rudybot: mithos28: ; Value: 3
02:14 mithos28: rudybot: (define-values (x y) (values 3 4))
02:14 rudybot: mithos28: Done.
02:14 mithos28: rudybot: (+ x y)
02:14 rudybot: mithos28: ; Value: 7
02:14 lewis1711: rudybot: (define-syntax (g stx) (syntax-case stx (quote list) [(_ 'x) #'(define x 3)] ((_ (list x)) #'(define x 4))))
02:14 rudybot: lewis1711: your sandbox is ready
02:14 rudybot: lewis1711: Done.
02:14 lewis1711: rudybot: (g (list lol))
02:14 rudybot: lewis1711: Done.
02:14 lewis1711: rudybot: lol
02:14 rudybot: lewis1711: ; Value: 4
02:15 lewis1711: hm
02:15 (quit) dnolen: Quit: ERC Version 5.3 (IRC client for Emacs)
02:15 mithos28: sw2wolf: Is that what you wanted?
02:16 sw2wolf: (multiple-value-bind (x y) (xlib:query-pointer (window-xwin win)) ...)
02:16 sw2wolf: i want to translate it into racket code ?
02:16 mithos28: i had two examples with let-values and define-values
02:17 sw2wolf: yes
02:17 sw2wolf: mind show me ?
02:18 mithos28: I sent them to rudybot a minute ago, did you miss that?
02:18 sw2wolf: i just come back
02:18 mithos28: rudybot: (let-values (((a b) (values 1 2))) (+ a b))
02:18 rudybot: mithos28: ; Value: 3
02:19 mithos28: rudybot: (define-values (x y) (values 3 4))
02:19 rudybot: mithos28: Done.
02:19 mithos28: rudybot: (+ x y)
02:19 rudybot: mithos28: ; Value: 7
02:19 sw2wolf: let-values will be my choice, thx
02:27 (quit) jonrafkind: Ping timeout: 265 seconds
02:29 lewis1711: rudybot: (define-syntax (g do-nothing stx) (syntax-case stx (quote list) [(_ 'x) #'(define x 3)] [(_ (list x)) #'(define x 3)]))
02:29 rudybot: lewis1711: Done.
02:29 lewis1711: rudybot: (g null 'a)
02:29 rudybot: lewis1711: error: #<undefined>:1:0: g: illegal use of syntax in: (g null (quote a))
02:30 lewis1711: rudybot: (g a 'a)
02:30 rudybot: lewis1711: error: #<undefined>:1:0: g: illegal use of syntax in: (g a (quote a))
02:30 mithos28: lewis1711: thats not going to work
02:30 mithos28: you want g to only have one argument
02:31 mithos28: if you call g as (g null 'a), stx is bound to the syntax object #'(g null 'a)
02:31 mithos28: the (_ 'x) part destructures it
02:31 mithos28: the underscore matches the macro part because you usually don't care about it
02:32 lewis1711: hmm
02:32 mithos28: so if you want a macro that takes two arguments change that to (_ x y)
02:32 mithos28: where x and y are any patterns you want
02:32 lewis1711: but syntax-case only takes "stx" as an argument
02:32 lewis1711: so how does it even 'know' about do-nothing?
02:32 mithos28: it doesn't even get that far
02:32 mithos28: for a macro to work it needs to take one argument
02:33 mithos28: which is the whole form
02:33 lewis1711: rudybot: (define-syntax (g val stx) (syntax-case stx (quote) [(_ 'x t) #'(define x t)]))
02:33 rudybot: lewis1711: Done.
02:34 lewis1711: (g 'a 3)
02:34 lewis1711: rudybot: (g 'a 3)
02:34 rudybot: lewis1711: error: #<undefined>:1:0: g: illegal use of syntax in: (g (quote a) 3)
02:34 mithos28: get rid of the val
02:34 (join) asvil
02:34 lewis1711: rudybot: (define-syntax (g stx) (syntax-case stx (quote) [(_ 'x t) #'(define x t)]))
02:34 rudybot: lewis1711: Done.
02:34 lewis1711: (g 'a 3)
02:34 lewis1711: rudybot: (g 'a 3)
02:34 rudybot: lewis1711: Done.
02:34 lewis1711: rudybot: a
02:34 rudybot: lewis1711: ; Value: 3
02:35 (quit) Demosthenex: Ping timeout: 250 seconds
02:35 lewis1711: so macros of the form "define-syntax" always take one argument? which is the whole form?
02:36 mithos28: all macros are of the form define-syntax, all other macros expand out to them
02:36 mithos28: *macros -> macro definitions
02:36 mithos28: and yes they always take the whole form
02:37 (join) Demosthenex
02:38 lewis1711: I see
02:46 (join) cdidd
02:59 (quit) Demosthenex: Ping timeout: 255 seconds
02:59 (quit) mizu_no_oto: Quit: ["Textual IRC Client: www.textualapp.com"]
03:00 (join) Demosthenex
03:01 (quit) ambrosebs: Ping timeout: 246 seconds
03:04 (quit) veer: Quit: Leaving
03:05 (join) RacketCommitBot
03:05 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/Py2lZg
03:05 RacketCommitBot: racket/master 777a6cd Eli Barzilay: New Racket version 5.3.1.9.
03:05 (part) RacketCommitBot
03:08 (quit) Demosthenex: Ping timeout: 244 seconds
03:11 (quit) tsion: Quit: Leaving
03:15 (join) ambrosebs
03:17 (join) Demosthenex
03:22 (join) momo-reina
03:28 (quit) nilyaK: Quit: Leaving.
03:33 (quit) momo-reina: Remote host closed the connection
03:45 (quit) Nisstyre-laptop: Read error: Operation timed out
03:59 (part) sw2wolf: "ERC Version 5.3 (IRC client for Emacs)"
04:10 (join) neilv
04:24 (join) bitonic
04:33 (quit) Demosthenex: Ping timeout: 252 seconds
04:35 (join) Demosthenex
04:54 (quit) Demosthenex: Read error: Operation timed out
04:58 (join) Demosthenex
05:04 (quit) mithos28: Quit: mithos28
05:04 (quit) Demosthenex: Ping timeout: 264 seconds
05:06 (join) Demosthenex
05:16 (quit) ambrosebs: Ping timeout: 264 seconds
05:18 neilv: (canonicalize-path "y/b1/s") ;==> #<path:/tmp/a/b2/c/f>
05:20 (quit) Demosthenex: Ping timeout: 245 seconds
05:28 (join) Demosthenex
05:42 (join) MayDaniel
05:44 (quit) bitonic: Ping timeout: 240 seconds
05:49 (join) soegaard
05:50 lewis1711: woo, I have coded basic prototype objects. they're pretty ugly to use though
05:52 neilv: cool
05:53 neilv: i made a simple prototype object library in scheme a while ago. http://www.neilvandyke.org/protobj/
05:57 lewis1711: neilv: nice. did you use hashtables internally?
05:58 lewis1711: much better looking than mine lol
06:01 (join) bitonic
06:03 neilv: i used old scheme that didn't have hash tables
06:04 lewis1711: ah I see. racket 5.3 is kicking up a bit of fuss installing it :)
06:07 neilv: this is strange. drracket is trying to install neil/parted when i require neil/protobj
06:08 lewis1711: hmm still seems to work a treat
06:09 neilv: looks like i just had an aborted planet install, although i don't recall that
06:10 neilv: btw, i would not use mutable pairs nowadays
06:12 neilv: but you might look at using assq rather than hash-ref, and whether that's faster when you don't have that many slots.  and/or do optimizations of slot lookup, such as at compile time
06:13 lewis1711: I will indeed
06:14 lewis1711: do you have any tips on coding "self"? mine is sort of stupid in that I can't send messages to self in any of the methods, just to an instance of an object, which of course remains fixed when its cloned
06:15 neilv: you can make self be an express argument to each method closure
06:16 lewis1711:   (! a xplus (lambda ($ n) (+ (? $ x) n)))
06:16 lewis1711: ohhh of course I see what you've done
06:16 neilv: if you want to have self be more magical, you could to it with syntax extension (there is a feature to make macros for identifiers that are not surrounded by parens), or you could make it a parameter (in the sense of make-parameter)
06:16 lewis1711: @ just passes in the object into the $ position
06:17 neilv: oh yeah, i guess $ was short for $elf
06:21 lewis1711: I tried to copy how lua did it
06:22 (join) sw2wolf
06:39 lewis1711: (define-syntax-rule (method args body ...)
06:39 lewis1711: (lambda args body ...))
06:40 lewis1711: if I wanted to append or cons "self" onto the arguments list, how would I do it? In the context of define-syntax-rule, what type is args?
06:40 (quit) Demosthenex: Read error: Operation timed out
06:43 (join) Demosthenex
06:45 neilv: args is a list
06:46 (join) masm
06:48 sw2wolf: it is a list
06:48 lewis1711: rudybot: (define-syntax-rule (method args body ...) (lambda (cons self args) body ...))
06:48 rudybot: lewis1711: Done.
06:48 lewis1711: rudybot: (method (x) x)
06:48 rudybot: lewis1711: error: #<undefined>:1:8: lambda: not an identifier, identifier with default, or keyword at: (x) in: (lambda (cons self (x)) x)
06:49 lewis1711: oh
06:53 lewis1711: ok how would I do this? I mean I sort of get why its done that. it's just shuffling around tokens, it's not actually doing anything to them
07:05 (quit) sw2wolf: Quit: Leaving
07:19 (join) mye
07:28 (join) RacketCommitBot
07:28 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/wzbnKQ
07:28 RacketCommitBot: racket/master 19d7519 Matthew Flatt: don't fail when there's no user-specific docindex
07:28 (part) RacketCommitBot
07:30 lewis1711: http://pastebin.com/Nq1701UK I got this far. self isn't in the context.. hmm
07:37 neilv: the problem you're running into is called "hygiene"
07:37 neilv: or, rather, you're running into the soap
07:38 lewis1711: I see
07:39 lewis1711: so I should write a common lisp type macro?
07:39 neilv: introducing "self" as a variable like that is intentionally made difficult
07:39 neilv: http://doc.racket-lang.org/guide/pattern-macros.html#%28part._.Identifier_.Macros%29
07:39 rudybot: http://tinyurl.com/blwlxbt
07:39 neilv: this is one way that might work
07:40 neilv: it's an advanced thing in macros. you almost never want to do things that way. but in the case of "self", you could make a good argument for it
07:46 lewis1711: (syntax-id-rules  ()
07:46 lewis1711: [(_ () body) (lambda (self) body)]
07:46 lewis1711: [(_ (args ...) body) (lambda (self args ...) body)]))
07:46 lewis1711: I am not sure if syntax-id-rules needs an argument
07:48 (quit) bitonic: Ping timeout: 240 seconds
07:50 (quit) mye: Quit: mye
07:58 (join) mye
08:02 (join) RacketCommitBot
08:02 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/2vCcHA
08:02 RacketCommitBot: racket/master 7ef4bec Jens Axel Søgaard: Added more margin notes
08:02 (part) RacketCommitBot
08:04 (quit) neilv: Quit: Leaving
08:25 (join) jeapostrophe
08:25 (quit) jeapostrophe: Changing host
08:25 (join) jeapostrophe
08:37 (quit) bjz: Quit: Leaving...
09:00 (join) karswell
09:07 (part) lewis1711
09:08 (quit) mye: Quit: mye
09:09 (join) RacketCommitBot
09:09 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/a1Lp5Q
09:09 RacketCommitBot: racket/master 124b5ab Mike Sperber: Synch German string constants with latest.
09:09 (part) RacketCommitBot
09:09 (join) em
09:18 (join) anRch
09:38 (quit) jschuster: Quit: Coyote finally caught me
09:42 (quit) soegaard: Quit: soegaard
09:53 (join) ambrosebs
10:06 (join) dnolen
10:10 (join) soegaard
10:14 (quit) anRch: Quit: anRch
10:16 (join) RacketCommitBot
10:16 RacketCommitBot: [racket] plt pushed 4 new commits to master: http://git.io/4HlwNA
10:16 RacketCommitBot: racket/master c12a129 Matthew Flatt: remove unused locking layer in the doc database manager
10:16 RacketCommitBot: racket/master 3f08da6 Matthew Flatt: remove 'always-run style from the doc that list documents...
10:16 RacketCommitBot: racket/master 4ca6e3c Matthew Flatt: fix exception handling during collection-table read...
10:16 (part) RacketCommitBot
11:00 (quit) MayDaniel: Read error: Connection reset by peer
11:22 (join) RacketCommitBot
11:22 RacketCommitBot: [racket] plt pushed 7 new commits to master: http://git.io/0BepcA
11:22 RacketCommitBot: racket/master ecb88af Jay McCarthy: Fixing some ugly line-widths in DrDr
11:22 RacketCommitBot: racket/master fd86928 Jay McCarthy: Changing the timeout for this file for DrDr...
11:22 RacketCommitBot: racket/master cf7e54b Jay McCarthy: Set a maximum timeout for all tests in DrDr
11:22 (part) RacketCommitBot
11:29 ambrosebs: I've been thinking about this problem for a while.
11:29 ambrosebs: How do runtime contracts generated for Typed Racket interact with parametric polymorphism?
11:31 ambrosebs: It seems like we could add runtime casts at any point in a Typed Racket but it's not always clear what to test for with type variables in the types.
11:31 ambrosebs: I'm being vague because I guess I don't really know if it's a problem.
11:32 ambrosebs: haha please call me out if I'm not being clear enough.
11:40 (quit) walter: Ping timeout: 265 seconds
11:40 (join) walter|r
12:00 (quit) karswell: Ping timeout: 265 seconds
12:22 (join) MayDaniel
12:25 (quit) Shviller: Read error: Connection reset by peer
12:30 (join) Shviller
12:33 (join) torog
12:34 torog: if, cond etc. are evaluated lazily in racket, aren't they?
12:41 soegaard: What do you mean by lazily?
12:45 (join) mithos28
12:51 (join) muraiki
12:52 muraiki: Hi all! In Racket, what would be the idiomatic way to traverse a hash? For instance, I have a hash that maps names to structs and I want to call some functions to use the data in the structs (but not modify the hash). In Python I'd use a list comprehension, but I'm not quite sure what to do in Racket.
12:53 muraiki: I thought about using the hash-iterate functions in a recursive function but I feel like there's probably a more straightforward way to do it.
12:54 mithos28: rudybot: (for (((key value) (hash 'a 1 'b 2))) (printf "Key ~a, value ~a" key value))
12:54 rudybot: mithos28: i guess i don't see what makes a simple printf falsely lazy
12:54 muraiki: oh, it looks like I might have stumbled in on a relevant discussion
12:54 mithos28: muraiki: That should have done it
12:54 muraiki: haha
12:54 mithos28: but I think I messed up a paren
12:55 mithos28: rudybot: eval (for (((key value) (hash 'a 1 'b 2))) (printf "Key ~a, value ~a" key value))
12:55 rudybot: mithos28: ; stdout: "Key b, value 2Key a, value 1"
12:56 muraiki: Thanks!
12:56 mithos28: hash iterate is way too low level, many things like for comprehensions are built on top of it
12:57 muraiki: Yeah, I felt like there would be a higher level function for it. I'll look up how for works.
12:58 (join) Nisstyre-laptop
13:02 (join) mizu_no_oto
13:07 (quit) masm: Quit: Leaving.
13:09 (quit) torog: Quit: leaving
13:20 (join) mye
13:25 offby1: any of the "for" forms will traverse a hash
13:29 (quit) Demosthenex: Quit: leaving
13:30 (join) Demosthenex
13:30 (quit) ambrosebs: Ping timeout: 240 seconds
13:35 (join) RacketCommitBot
13:35 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/wH6Pdg
13:35 RacketCommitBot: racket/master 4b094f8 Matthew Flatt: raco setup: fix for non-parallel doc build in a fresh installation
13:35 (part) RacketCommitBot
13:54 (join) myx
13:55 (quit) Nisstyre-laptop: Quit: Leaving
13:57 (join) Nisstyre-laptop
14:32 (quit) mye: Quit: mye
14:40 (quit) mithos28: Quit: mithos28
14:43 (join) mye
14:50 (quit) snorble_: Ping timeout: 240 seconds
14:53 (join) mye_
14:53 (join) bitonic
14:57 (quit) mye: Ping timeout: 252 seconds
14:57 (nick) mye_ -> mye
15:02 (quit) mizu_no_oto: Quit: ["Textual IRC Client: www.textualapp.com"]
15:06 (join) RacketCommitBot
15:06 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/VGxaIg
15:06 RacketCommitBot: racket/master 793ee71 Robby Findler: fix a coloring bug...
15:06 (part) RacketCommitBot
15:06 (quit) jrslepak: Quit: What happened to Systems A through E?
15:11 mye: If I want to write what a macro returns to an output port, what do I do? (except, obviously, come here and ask :-)
15:11 mye: I suppose it will involve manipulating phases or something
15:12 mye: (I'm generating (trying anyway) ffi bindings from a json input file and want to write the bindings as a single file)
15:13 mye: now I'm doing all the macros that create the ffi syntax, but I'm not sure how to create the file
15:17 (join) mithos28
15:19 (quit) muraiki: Quit: Leaving
15:23 mye: does the macro stepper only do phase 1 ?  :(
15:24 (quit) Nisstyre-laptop: Ping timeout: 244 seconds
15:30 soegaard: mye: In the body of the macro use:  (let () (define out <the expression returned>) (display out) (newline) out)
15:31 soegaard: This will print out the expansion every time the macro is used.
15:31 soegaard: Then in DrRacket click the small triangles to fold out the syntax object printed.
15:37 mye: soegaard: yeah that's a good way to do it thanks. (I was changing the macros to return strings to write that)
15:42 (join) snorble_
15:42 (join) Nisstyre-laptop
16:05 (join) jschuster
16:22 (join) nathanpc
16:36 (join) jonrafkind
16:37 (join) RacketCommitBot
16:37 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/kmrM_A
16:37 RacketCommitBot: racket/master dd9d85f Eli Barzilay: Cleanup junk files in a bunch of racket tests.
16:37 (part) RacketCommitBot
17:12 (quit) mithos28: Quit: mithos28
17:16 (quit) jschuster: Quit: Coyote finally caught me
17:34 (join) anRch
17:44 (join) walter|rtn
17:47 (quit) walter|r: Ping timeout: 256 seconds
17:51 (join) jschuster
17:53 (join) mithos28
18:18 (join) bjz
18:26 mithos28: can planet 2 packages run code on installation?
18:37 soegaard: yes. Planet 1 can too.
18:37 mithos28: is there documentation on how to do it for planet 2?
18:37 soegaard: Unless I have misunderstood something.
18:38 soegaard: Well, when you require a module, code runs.
18:38 mithos28: thats at runtime
18:39 mithos28: I want code that runs once per installation, it is compiling a c++ file into a dynamic library
18:39 soegaard: Well, (begin-for-syntax …) will make it run at compile time.
18:39 mithos28: thats not installation time
18:39 soegaard: Just check whetner the resulting file is present, and compile only of it is missing.
18:40 mithos28: So it seems like the answer is no then, but I can hack it as you suggest
18:42 soegaard: I guess it depends on the point of view whether it is a hack or not :-)
18:44 mithos28: It will work, but I wish there was a more principled way
18:47 (join) francisl
18:49 (quit) cdidd: Remote host closed the connection
18:56 (quit) MayDaniel: Read error: Connection reset by peer
18:57 (quit) mithos28: Quit: mithos28
18:57 (quit) anRch: Quit: anRch
19:01 (join) mithos28
19:10 (quit) mithos28: Quit: mithos28
19:12 (quit) Demosthenex: Ping timeout: 240 seconds
19:14 (join) Demosthenex
19:22 (quit) francisl: Quit: francisl
19:32 (quit) bitonic: Ping timeout: 260 seconds
19:43 (join) tsion
19:43 (quit) tsion: Changing host
19:43 (join) tsion
19:51 (quit) jonrafkind: Ping timeout: 246 seconds
19:58 (quit) jeapostrophe: Ping timeout: 265 seconds
20:03 (join) mizu_no_oto
20:03 (join) aezx
20:05 (join) mithos28
20:07 (join) lewis1711
20:10 aezx: i just tried sublime for the second time and I'm switching to it. it has paredit now, and it's gorgeous
20:11 lewis1711: what's that? a text editor?
20:13 aezx: yeah
20:13 aezx: sublime text 2
20:13 aezx: very graphical
20:13 aezx: old school lispers seem to dislike it
20:13 (join) jonrafkind
20:13 aezx: "emacs!!!!11!!!one!!!!"
20:14 aezx: emacs gives you arthritis after half an hour, no thanks
20:15 lewis1711: LOL
20:15 jonrafkind: what does vi give you?
20:15 lewis1711: yeah one of the reasons I avoid common lisp is that if you try and use anything without emacs you are out of luck
20:15 aezx: vi is great
20:16 aezx: i dont have to use weird multi key combos
20:16 aezx: i can type commands
20:16 lewis1711: emacs is probably the worlds best editor but I can't figure it all out. and I spent days and days on it. of course some people here may know I'm not the sharpest tool in the shed but still, vim is plenty complex enough for me
20:16 aezx: eh
20:17 aezx: i can make do with whatever i have to
20:17 offby1: lewis1711: #emacs for occasional actual useful help, and constant off-topic randomness :)
20:17 offby1: aezx: if I didn't have Emacs, it'd be like someone cut off both my arms.
20:17 aezx: i dont like arguing about text editors. if people have time to argue about personal preference, they aren't coding enough
20:17 aezx: offby1, do hand stretches a lot or you will get arthritis for real
20:18 offby1: aezx: I've been using it for almost 30 years; no hint of arthritis.
20:18 aezx: cyborg detected
20:18 offby1: I do insist on "Microsoft Natural"-style keyboards for desktops, though.
20:19 jonrafkind: real hackers use ibm model M
20:19 aezx: i like those keyboards that are slightly curved, it feels more natural because I have huge hands
20:19 aezx: real hackers arent defining themselves through keyboards and editors
20:20 jonrafkind: are you sure?
20:21 aezx: 100%
20:22 aezx: that being said, it's futile to deny that Model M was the rosetta stone of keyboards
20:22 aezx: we've studied and studied it, but it's true power remains locked
20:23 aezx: aliens
20:26 (join) eMBee
20:27 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
20:28 offby1: oh, the true power is well-understood: it's the clackety-clackety sound.
20:28 offby1: That sounds summons up the hacker daemons.
20:28 offby1: Also the ghosts of typists past.
20:29 lewis1711: I have hands like a little girl. maybe that's why I don't mind netbook keyboards
20:43 aezx: i love the way scheme teaches me simplicity through exemplifying my own stupidity
20:43 aezx: the best teacher is truly oneself
20:45 (join) mizu_no_oto
21:10 (quit) mithos28: Quit: mithos28
21:16 offby1: I wish :-|
21:27 lewis1711: oh yeah, I should ask my syntax-id-rules question again
21:28 (join) mithos28
21:28 offby1: oh yeah.
21:28 lewis1711: http://pastebin.com/kkCG0DYX how can I capture self in this?
21:29 jonrafkind: you need to syntax-local-introduce it
21:29 jonrafkind: or create the self syntax using the lexical context of body
21:29 jonrafkind: i think both things you cannot do in syntax-id-rules
21:29 jonrafkind: you probably have to do 'make-set!-transformer' yourself
21:29 jonrafkind: so you can use syntax-case
21:30 lewis1711: I had a syntax-case version before, but make-set!-transformer is new to me
21:31 lewis1711: why do I need something that cooperates with set! ?
21:31 lewis1711: http://docs.racket-lang.org/reference/syntax-model.html#%28tech._assignment._transformer%29
21:31 rudybot: http://tinyurl.com/crozwrh
21:32 jonrafkind: i dunno why were you using syntax-id-rules in the first place?
21:32 jonrafkind: i almost never use set! transformers
21:32 lewis1711: jonrafkind: someone else here suggested it because appareantly the issue was I was running into variable capture hygiene problems
21:33 lewis1711: because the self i write in method becomes different to the self inside the macro
21:33 jonrafkind: variable capture is orthogonal to set! transformers
21:33 jonrafkind: you just need to get the right lexical context
21:34 jonrafkind: so either do (with-syntax ([self (datum->syntax #'body 'self #'body)]) #'(lambda (self blah...) body))
21:34 jonrafkind: or something with syntax-local-introduce that i would have to read the docs to figure out
21:35 (join) sw2wolf
21:36 lewis1711: i will look at with-syntax
21:37 jonrafkind: the deal with syntax-local-introduce is it adds a mark for the current macro transform so it will end up not being alpha renamed
21:37 lewis1711: ok these docs make no sense to me
21:38 jonrafkind: all the syntax passed to the macro is marked with a mark like A on input and then remarked with A on output, so two marks cancel
21:39 lewis1711: thanks, but I don't understand any of your explinations :) I think I need to do a macro tutorial or something, it doesn't seem to be something you can learn by doing
21:40 jonrafkind: http://www.cs.indiana.edu/~dyb/pubs/bc-syntax-case.pdf
21:40 jonrafkind: this explains marking
21:41 offby1: hm, my cat marks occasionally.
21:41 offby1: I wish he wouldn't.
21:43 lewis1711: oh so even offby1s cat is a better scheme programmer than me ;)
21:43 lewis1711: what is a binding, exactly?
21:44 lewis1711: "The transformation above works fine in most cases, but it breaks down if the identifier t appears free in e2 (i.e., outside of any binding for t in e2),"
21:46 offby1: it's what "lambda" creates for the parameters.
21:46 offby1: A region of your code where a variable is ... uh ... bound.
21:47 jonrafkind: a binding is a mapping between an identifier and a location in the environment basically
21:47 jonrafkind: to look up an identifier in the environment you use its name, like its actual symbol value, and its "lexical context'
21:47 jonrafkind: lexical context is an opaque thing, you can't really print it
21:48 jonrafkind: you can think of it as just a simple number though
21:48 jonrafkind: (lambda (x) (lambda (x) x))
21:48 jonrafkind: the first x in the lambda has lexical context 1, and the second has lexical context 2
21:48 jonrafkind: so both have the same symbol 'x, but different lexical contexts
21:48 lewis1711: so it sort of represents its "distance" from the top level?
21:49 jonrafkind: well i think thats putting too much meaning into what the number is
21:49 jonrafkind: i just chose numbers because they are a simple thing
21:49 lewis1711: fair enough
21:49 jonrafkind: i could have chosen strings, like "a" for the first x and "b" for the second
21:50 jonrafkind: so when dealing with macros you have to be aware of a) lexical context and b) marks on syntax
21:50 sw2wolf: what's "marks on syntax' ?
21:50 jonrafkind: although I think the lexical context incorporates marks so its kind of the same thing
21:51 jonrafkind: when you invoke a macro the syntax object passed to the macro is "marked" so that its differentiated from syntax built inside the macro
21:51 lewis1711: is a generated identifier like "gensym" ?
21:51 jonrafkind: gensym generates symbols that are not interned
21:51 lewis1711: jonrafkind: I thought only some macros took in syntax objects
21:51 jonrafkind: all macros take in syntax objects
21:51 lewis1711: others took in lists, like define-syntax-rule
21:52 jonrafkind: define-syntax-rule is hiding the true macro api from you
21:52 jonrafkind: so its easier to use
21:52 jonrafkind: define-syntax-rule is a macro that is implemented in terms of define-syntax and syntax-case or something
21:52 lewis1711: hmm
21:52 lewis1711: maybe i should just use define-syntax macros until I understand what's going on
21:53 jonrafkind: as soon as you try to do unhygienic things life becomes very hard
21:53 lewis1711: variable capture? like my "self" attempt?
21:54 jonrafkind: yea
21:55 lewis1711: I see
21:55 lewis1711: aren't there old style common lisp macros when you need to do what I'm doing?
21:56 lewis1711: IIRC CL ones don't have hygience, they just capture variables any old way, and you use gensym when you want to avoid it
21:56 jonrafkind: no in lisp there is no such idea as lexical context. macros in lisp are just unhyienic so when you want to capturer 'self' you just bind it
21:56 jonrafkind: i dont use lisp much but i guess there are people in the lisp community who think hygiene isn't really necessary
22:00 (quit) myx: Ping timeout: 265 seconds
22:13 (join) jeapostrophe
22:13 (quit) jeapostrophe: Changing host
22:13 (join) jeapostrophe
22:16 offby1: jonrafkind: when you say "lisp", do you mean Common Lisp?
22:16 jonrafkind: i dunno i guess
22:16 jonrafkind: are there any lisps that do hygiene?
22:16 offby1: well, Scheme :)
22:17 jonrafkind: gimme a break
22:17 offby1: dunno if Clojure does
22:17 jonrafkind: i think clojure is hygienic
22:18 lewis1711: I use LISP when talking about the family and cl/scheme/clojure/racket.
22:18 jonrafkind: what about racket without s-expressions? would yous till call it a lisp?
22:19 lewis1711: hmm. i guess not
22:19 jonrafkind: i guess s-expression is not a trivial thing but its not really enough to bind those things together with the same language family name
22:20 lewis1711: I can't use cl after getting slightly used to racket. which is a shame is CL has far more resources docs/etc
22:20 lewis1711: yeah, perhaps not
22:20 jonrafkind: https://github.com/kazzmir/Honu/blob/master/asteroids/asteroids.honu i mean this is racket but with non-s-expression syntax (and a different macro system)
22:20 lewis1711: but I mean they're all dynamically typed and  have macros and eval and such..
22:20 jonrafkind: so what about python
22:20 jonrafkind: if it had macros it would be a lisp?
22:20 lewis1711: needs macros and s-expressions
22:20 lewis1711: or something :/
22:20 jonrafkind: what about just macros?
22:20 lewis1711: nah then C would be a lisp :D
22:21 jonrafkind: C macros aren't even comparable :p
22:21 lewis1711: they're the only ones I know :(
22:22 sw2wolf: i feel it seems CL macro simpler than racket's ?
22:22 lewis1711: speaking of macros - when declaring a new struct you do (struct whatever (field-1 field-2)) - does this break hygiene? since it captures "whatever", "field-1" and "field-2"
22:22 jonrafkind: sw2wolf, yes because it doesnt have hygiene
22:22 jonrafkind: so you are just manipulating lists which is pretty easy
22:23 sw2wolf: sorry ! i donot understand why we need hygiene macro ?
22:23 lewis1711: woah, honu looks like javascript
22:23 jonrafkind: to prevent unwanted variable capture
22:23 sw2wolf: unhygien seems work well in CL ?
22:24 jonrafkind: i assume you mean gensym. i dont remember offhand common situations where it doesn't work, but anyway its easy to write a broken macro if you mis-use/dont use genym
22:24 sw2wolf: would you mind showing me a example if hygiene usage ?
22:24 offby1: jonrafkind: don't we have a bot with factoids or something?  I'd think this is a prime candidate for a channel FAQ
22:25 offby1: _once_ I read an article about hygienic macros that I almost understood.  It might have been by Shriram or Matthew
22:25 jonrafkind: (define-syntax-rule (or a b) (let ([t a]) (if t t b))) (let ([t 2]) (or #f t)) will be #f instead of 2
22:25 offby1: can't find it now though
22:29 lewis1711: can anyone show me an example of syntax-local-introduce?
22:29 lewis1711: a really basic one
22:30 (join) Kaylin
22:30 jonrafkind: hm you might just be able to do (with-syntax ([self (syntax-local-introduce (datum->syntax #f 'self))]) ...)
22:33 lewis1711: what does that look like in the context of define-syntax ?
22:33 jonrafkind: sw2wolf, http://www.ccs.neu.edu/racket/pubs/dissertation-kohlbecker.pdf kohlbeckers dissertation will hopefully answer some questions about hygiene and why its needed
22:33 offby1: "if", I'd like you to meet "lambda".
22:33 jonrafkind: (define-syntax (foo stx) (syntax-case stx [(whatever) (with-syntax ...)]))
22:33 (quit) Shviller: Ping timeout: 240 seconds
22:33 (join) Shviller
22:34 Kaylin introduces if to cond first.
22:34 sw2wolf: (defmacro or (a b) `(let ((t ,a)) (if t t ,b)))
22:34 sw2wolf: (let ((t 2)) (or nil t)) will report error
22:34 sw2wolf: so CL needs gensym
22:35 jonrafkind: you asked for an example, not a particularly great one
22:35 jonrafkind: i know you can use gensym
22:35 jonrafkind: the point is for very complicated macros its easy to mess things up with gensym
22:35 sw2wolf: yes
22:35 jonrafkind: in a hygienic macro system you have to try very hard to mess things up
22:38 sw2wolf: doesnot racket macro need gensym ?
22:38 jonrafkind: you only need gensym as a way to generate new unique identifiers, but even then its just a convenience, not really necessary
22:39 jonrafkind: you don't need to do anything special to prevent identifiers created in the macro (like that 't') from conflicting with identifiers outside the macro
22:39 jonrafkind: hygiene will do it for you
22:40 sw2wolf: so in (let ([t a]) ... ) ,  then binding doesnot occur ?
22:40 jonrafkind: binding does occur, but hygiene will alpha rename that 't' so it does not conflict with the users t
22:41 sw2wolf: alpha rename seems like gensym ?
22:41 jonrafkind: yes but its automatic
22:41 jonrafkind: gensym is manual
22:41 sw2wolf: oh, i see now. thx
22:41 jonrafkind: k
22:41 lewis1711: is this stuff supposed to be very hard to learn or am i doing it wrong? :/
22:41 jonrafkind: its not the easiest thing..
22:42 offby1: we deliberately made it hard to keep people like you from penetrating the inner sanctum
22:44 lewis1711: SUCESSS (maybe)
22:45 sw2wolf: yea
22:45 offby1: damn, now I have to find some other secret society to join
22:45 (quit) mizu_no_oto: Quit: ["Textual IRC Client: www.textualapp.com"]
22:46 lewis1711: http://pastebin.com/0mVqt2zM
22:46 jonrafkind: local-get-shadower??
22:46 offby1: jonrafkind: it's for private detectives who are in an unfamiliar city.
22:46 lewis1711: is that a bad idea for some reason?
22:46 jonrafkind: i barely have any idea what it does.. most likely you have none :lp
22:47 lewis1711: you'd be right, i saw it on a blog :(
22:47 jonrafkind: ok i read the docs and i still have no idea
22:47 jonrafkind: its definitely the wrong thing
22:49 lewis1711: "ok i read the docs and i still have no idea" I know that feeling :D
22:50 jonrafkind: yes the lack of examples for a lot of the macro api makes it very hard
22:50 jonrafkind: take out the shadower thing and just do (datum->syntax #f 'self)
22:50 jonrafkind: as the argument to syntax-local-introduce
22:51 lewis1711: that also works
22:51 offby1: lewis1711: by the way, what you're doing -- making an ultra-simple object system in Racket -- has been done many times, so it should be easy to find examples on the web
22:53 lewis1711: that's true. I sorta wanted to forge ahead on my own for educational purposes though
22:54 offby1: fair enough
22:54 lewis1711: http://pastebin.com/ankUn6PP is there anyway to avoid having to write that out twice?
22:54 offby1: a _little_ cheating can help, though :)
22:54 (join) francisl
22:54 jonrafkind: just do args ...
22:54 jonrafkind: you dont need the () pattern
22:54 lewis1711: haha, true. if I can;t finish it this weekend i will then. but I'm almost done
22:55 jonrafkind: if there are no args then args ... will match the empty list
22:55 offby1: if there is no allah but allah ...
22:55 lewis1711: oh, I didn't know ... could do that. I thought it was 1 or more
22:56 lewis1711: cool, done
22:59 (join) jackhammer2022
23:03 lewis1711: thanks alot for your help jonrafkind
23:05 jonrafkind: sure
23:06 (quit) mithos28: Quit: mithos28
23:17 (quit) Nisstyre-laptop: Remote host closed the connection
23:18 (join) Nisstyre-laptop
23:23 (join) ambrosebs
23:33 (quit) francisl: Quit: francisl
23:42 (quit) sw2wolf: Quit: ERC Version 5.3 (IRC client for Emacs)
23:47 (quit) nathanpc: Quit: Computer has gone to sleep.