00:21 (join) RacketCommitBot
00:21 RacketCommitBot: [racket] plt pushed 4 new commits to master: http://git.io/RY8_lg
00:21 RacketCommitBot: [racket/master] Fix requires to avoid circular dependencies. - Asumu Takikawa
00:21 RacketCommitBot: [racket/master] Class contracts for racket/draw. - Asumu Takikawa
00:21 RacketCommitBot: [racket/master] Class contracts for racket/snip. - Asumu Takikawa
00:21 (part) RacketCommitBot
00:45 (join) RacketCommitBot
00:45 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/eR65yQ
00:45 RacketCommitBot: [racket/master] Fix docs for `get-selected-text-color` method. - Asumu Takikawa
00:45 (part) RacketCommitBot
01:04 (join) realitygrill
01:15 (quit) ssbr__: Ping timeout: 252 seconds
01:15 cky: So, apparently, with the version of Racket that rudybot is on, vector->list is busted when srfi/43 is loaded. (It works correctly in Racket 5.2.1, however.)
01:15 cky: rudybot: init
01:15 rudybot: cky: your sandbox is ready
01:15 cky: rudybot: (require srfi/43)
01:15 rudybot: cky: Done.
01:15 cky: rudybot: eval vector->list
01:15 rudybot: cky: error: car: expects argument of type <pair>; given: #f
01:15 cky: rudybot: eval (vector->list '#(1 2 3 4 5))
01:15 rudybot: cky: error: car: expects argument of type <pair>; given: #f
01:16 mithos28: rudybot: (version)
01:16 rudybot: mithos28: your sandbox is ready
01:16 rudybot: mithos28: ; Value: "5.2.900.1"
01:17 mithos28: rudybot: vector->list
01:17 rudybot: mithos28: ; Value: #<procedure:vector->list>
01:17 mithos28: rudybot: (require srfi/43)
01:17 rudybot: mithos28: Done.
01:17 mithos28: rudybot: vector->list
01:17 rudybot: mithos28: error: car: expects argument of type <pair>; given: #f
01:17 mithos28: rudybot: (version)
01:17 rudybot: mithos28: ; Value: "5.2.900.1"
01:17 mithos28: iteresting
01:17 cky: Yep.
01:18 mithos28: rudybot: vector?
01:18 rudybot: mithos28: I've got a vector of N^2 length. How can I print a 'square' of N values per line? I know it's trivial, but I haven't done scheme in awhile.
01:18 cky: rudybot: eval vector?
01:18 rudybot: cky: ; Value: #<procedure:vector?>
01:18 mithos28: rudybot: eval vector?
01:18 rudybot: mithos28: ; Value: #<procedure:vector?>
01:18 mithos28: it is not even applying vector->list
01:18 cky: Indeed.
01:19 mithos28: is it a macro?
01:19 cky: No.
01:19 (join) jeapostrophe
01:19 mithos28: positive? many things are macros but act like procedures
01:19 cky: > vector->list
01:19 cky: #<procedure:|my-vector->list|>
01:19 cky: ^--- 5.2.1
01:19 mithos28: that doesn't mean that the binding is not a macro
01:20 cky: Well, I can't speak about that. :-)
01:20 mithos28: almost all contracted exports are macros
01:20 cky: Right. That's one of those, it seems.
01:21 mithos28: yep it is contracted
01:22 mithos28: https://github.com/plt/racket/blob/master/collects/srfi/43/vector-lib.rkt
01:22 mithos28: who runs rudybot?
01:23 mithos28: looks like offby1
01:23 mithos28: offby1: ping
01:55 (join) RacketCommitBot
01:55 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/h18I3Q
01:55 RacketCommitBot: [racket/master] Moving tests re: Sam - Jay McCarthy
01:55 (part) RacketCommitBot
01:55 (join) Kaylin
02:33 (quit) jeapostrophe: Ping timeout: 245 seconds
02:41 (quit) realitygrill: Quit: realitygrill
02:44 (quit) Skola: Quit: Lost terminal
03:15 (join) fftb
03:23 lewis1711: man some of those racket bench makrs on the shootout are cheating
03:23 mithos28: which ones?
03:23 lewis1711: surely. I mean one of them just opens up a C library with ffi-lib
03:24 lewis1711: hold up
03:24 lewis1711: http://shootout.alioth.debian.org/u32/program.php?test=pidigits&lang=racket&id=2
03:24 rudybot: http://tinyurl.com/3pen4sc
03:26 mithos28: Did you read the comment? It is based on the perl one
03:26 mithos28: http://shootout.alioth.debian.org/u32q/program.php?test=pidigits&lang=perl&id=2
03:26 rudybot: http://tinyurl.com/c57przc
03:26 mithos28: also using gmp
03:26 mithos28: http://shootout.alioth.debian.org/u32q/program.php?test=pidigits&lang=python3&id=2
03:26 rudybot: http://tinyurl.com/bq77ex9
03:26 mithos28: python using gmp
03:27 mithos28: So if its 'cheating', then everyone is doing it
03:27 lewis1711: oh dear
03:27 lewis1711: no wonder that shootout is worthless :D
03:27 lewis1711: the ruby one doesn't do it, and is of course very slow
03:28 (quit) Kaylin: Quit: Leaving.
03:28 mithos28: which isn't a great benchmark, because if you needed to do that you would use gmp
03:28 lewis1711: using foreign libs in language shootouts are a bit like using peroformance enhancing drugs in sport. It's wrong but everyone else is doing it and you have to stay competitive!
03:29 mithos28: I have also made racket ones for mandelbrot and nbody that perform at C speed
03:29 lewis1711: mithos28: yeah but I figured the point of the benchmark was to test numerical speed, not see which language can called gmp the fastest
03:29 mithos28: but invoking the LLVM compiler might be a little over the top
03:31 mithos28: but that requires arbitrary precision arithmetic
03:31 (quit) cdidd: Remote host closed the connection
03:31 mithos28: so you either have a language which has it, use gmp, or implement it in the language
03:32 mithos28: implementing it in the language isn't testing numerical speed
03:32 mithos28: and most languages that have it built in, will be using gmp as the backend
03:32 lewis1711: yeah but still, it then becomes part of the language and fair game IMO
03:33 mithos28: what is part of the language in the racket ecosystem?
03:33 mithos28: only #%kernel?
03:37 mrcarrot: somebody should make a better benchmark site with strict rules
03:37 mithos28: what rules would you want to see?
03:38 mrcarrot: very clear rules telling what is allowed... like this with using foreign libraries or not
03:39 mithos28: In addition to language specific multiprecision arithmetic, we will accept programs that use GMP.
03:39 mithos28: http://shootout.alioth.debian.org/u64q/performance.php?test=pidigits
03:39 mithos28: what is unclear about that
03:39 mrcarrot: ah, okay
03:42 mrcarrot: hmm. haskell is not doing it
03:42 mrcarrot: and it is still pretty fast
03:43 mithos28: what is it not doing?
03:44 mithos28: The Integer type is now provided by a separate integer package, which provides an API that hides the implementation details. By default this is integer-gmp.
03:44 mithos28: http://hackage.haskell.org/trac/ghc/wiki/ReplacingGMPNotes
03:44 mithos28: It is using gmp, just the same
03:45 mrcarrot: yeah, but not calling an outside library from the code
03:46 mithos28: I don't understand the distinction
03:46 mithos28: It is calling the library that is baked into the compiler, which by default is gmp
03:58 (quit) fftb: Remote host closed the connection
04:30 (quit) mithos28: Quit: mithos28
04:39 (join) bas_
04:39 (nick) bas_ -> Skola
04:43 (quit) dous: Remote host closed the connection
05:21 (join) mceier
05:47 (join) gciolli
05:50 (join) masm
07:04 (join) bitonic
07:16 (join) dous
07:50 (quit) dous: Remote host closed the connection
08:02 (part) gciolli
08:05 (part) lewis1711
08:06 (join) RacketCommitBot
08:06 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/G7KZrA
08:06 RacketCommitBot: [racket/master] Synch German string constants with latest. - Mike Sperber
08:06 (part) RacketCommitBot
08:18 (join) fftb
08:24 (join) samth
08:36 (join) dous
08:45 (join) cdidd
09:26 (join) jeapostrophe
09:27 em: for a struct do you do it like this: (struct foo (bar baz #:mutable))
09:27 em: does that just make the baz mutable or does it make the bar and baz both mutable?
09:27 em: can you add more fields later?
09:31 (quit) noam: Ping timeout: 260 seconds
09:37 (quit) samth: Ping timeout: 246 seconds
09:41 (join) noam
09:51 (join) RacketCommitBot
09:51 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/jY9Shg
09:51 RacketCommitBot: [racket/master] try to improve docs for `make-keyword-procedure' - Matthew Flatt
09:51 (part) RacketCommitBot
10:02 asumu: em: you can't add fields to an existing struct, but you can make sub-structs that inherit.
10:05 em: too hard. I'll just use hashes.
10:06 em: asumu: is it okay to 'initialise' a hash by doing something like (begin (hash-set! .... ) (hash-set! ....) (hash-set! .... ) ) ?
10:06 em: is using (begin ..) frowned on ?
10:15 asumu: em: It's more straightforward to initialize a hash by using #hash(('a . "foo) ...) or something.
10:16 asumu: Also, yes, if you are putting a variable number of keys use a hash.
10:16 asumu: If you have a fixed number of keys, use a struct.
10:18 (quit) jeapostrophe: Ping timeout: 246 seconds
10:19 em: asumu: but #hash(...) makes an immutable hash and i need mutable ones.
10:22 em: asumu: here' something kind of interesting, I tried a (begin (hash-set!...) (hash-set! ....) ) as i described above, and it worked, but the resulting hash does not have the keys in the same order as I included them in the (begin... ) sequence.
10:22 em: its like begin did those sets in some other order than the way I wrote them
10:28 (join) gridaphobe
10:30 em: Im trying to see how do you change an immutable hash into a mutable hash. It isn't in the guide or the reference where it should be.
10:33 (quit) gridaphobe: Ping timeout: 264 seconds
10:33 chandler: em: Hashes are unordered; they'll be printed out in some arbitrary order, but any ordering is equivalent.
10:34 chandler: em: To get a mutable copy of an immutable hash, use `hash-copy'.
10:36 em: chandler: oh and then i can give that copy a new name to work with?
10:36 em: chandler: doesn't a hash have to be mutable if you want to add keys to it, or change the values of existing keys? Otherwise what does mutable and immutable mean?
10:39 (join) gridaphobe
10:43 chandler: em: In a mutable hash, you can mutate the values of keys within the hash object itself. In an immtable hash, you can use the functional update procedures such as `hash-set' and `hash-remove'; these procedures return a new immutable hash table which is like the old one with the update applied.
10:44 chandler: em: Here's an example of how that works:
10:44 chandler: rudybot: eval (define h (make-hash '((a . 1) (b . 2))))
10:44 rudybot: chandler: your sandbox is ready
10:44 rudybot: chandler: Done.
10:44 chandler: rudybot: eval (hash-ref h 'a)
10:44 rudybot: chandler: ; Value: 1
10:44 chandler: rudybot: eval (hash-set! h 'a 3)
10:44 rudybot: chandler: Done.
10:44 chandler: rudybot: eval (hash-ref h 'a)
10:44 rudybot: chandler: ; Value: 3
10:45 chandler: So, that's a mutable hash. `h' still points to the same hash object, and that hash object has been mutated.
10:45 chandler: Here's an immutable hash:
10:45 chandler: rudybot: eval (define i (make-immutable-hash '((a . 1) (b . 2))))
10:45 rudybot: chandler: Done.
10:45 chandler: rudybot: eval (hash-ref i 'a)
10:45 rudybot: chandler: ; Value: 1
10:45 chandler: rudybot: eval (hash-set i 'a 3)
10:45 rudybot: chandler: ; Value: #hash((b . 2) (a . 3))
10:46 chandler: em: Here's a difference: `hash-set!' on a mutable hash returns void; `hash-set' (no exclamation point) on a mutable hash returns a value. Now let's see if the original hash has changed:
10:46 chandler: rudybot: eval (hash-ref i 'a)
10:46 rudybot: chandler: ; Value: 1
10:46 chandler: Nope; it's still the same.
10:47 em: hm.
10:47 em: yeah you didn't change a to 3.
10:48 chandler: Not in the original immutable hash, but `hash-set' returned a new hash where `a' was associated with 3.
10:48 em: yes i see.
10:48 em: i think i probably want a mutable hash then
10:49 asumu: em: what are you using a hash for?
10:49 asumu: (whether you use mutable/immutable depends on what you're doing...)
10:50 asumu: But generally avoiding mutation is nice.
10:50 chandler: Unless you're just exchanging mutation of the hash for `set!' of a global or its equivalent, in which case it makes more sense to use the mutable hash.
10:55 em: im building an application that i can use to keep track certain facts and details about my clients. Like "number of visits", "money spent" etc.
10:56 em: so I have one giant master hash that looks like #hash( ("name" . #hash())
10:57 em: for each client, and then the hash for that client has stuff like, #hash(("name" . name) ("phone" . phone) ("visits" . visits) ("spent" . hash-of-spent-by-date) ) etc
10:58 em: so then eventually what my app will do is let me enter in stuff like this (visit+ "name")  and update the visits part of "name"'s hash.
11:00 em: or (spent "name" 320) and it will (hash-set! (hash-ref (hash-ref master name) spent) "date" 320)
11:00 em: do you get what i have in mind.
11:00 em: it will be a big hash of data that i can quickly update and analyse
11:03 (join) dnolen
11:05 asumu: em: Why not use a struct in this case? Or are you concerned that your representation will change?
11:06 em: asumu: im concerned that since this is new I will not anticipate data i want to start keeping in the future.
11:08 em: so if I want to create a literal mutal hash i would do (hash-copy #hash(("first" . "firstname") ("last" . "lastname"))) ?
11:08 asumu: I see. Well, it's hard to say what you want in this case. It sounds like it's really a DB.
11:08 asumu: What you probably want is make-hash
11:08 asumu: rudybot: doc make-hash
11:08 rudybot: asumu: your sandbox is ready
11:08 rudybot: asumu: http://docs.racket-lang.org/reference/hashtables.html#(def._((quote._~23~25kernel)._make-hash))
11:09 em: can you just initialise (make-hash) in the first place somehow? like (make-hash ("first" . "firstname") ("last" . "lastname"))
11:09 asumu: rudybot: (define h (make-hash '((a . "Banana") (b . "Grape")))
11:09 rudybot: asumu: Is there a difference between (make-hash) and (hash) ?
11:09 asumu: rudybot: (define h (make-hash '((a . "Banana") (b . "Grape"))))
11:09 rudybot: asumu: Done.
11:09 asumu: rudybot: (dict-ref h 'a)
11:09 rudybot: asumu: error: car: expects argument of type <pair>; given: #f
11:09 em: asumu: I agree it is probably a database. Im basically making a database from hashes. But i don't have time to figure out databases :)
11:10 asumu: rudybot: (hash-ref h 'a)
11:10 rudybot: asumu: ; Value: "Banana"
11:10 asumu: rudybot: (hash-set! h 'a "Apple")
11:10 rudybot: asumu: Done.
11:10 asumu: rudybot: (hash-ref h 'a)
11:10 rudybot: asumu: ; Value: "Apple"
11:11 asumu: `hash` makes an immutable hash, `make-hash` mutable.
11:11 em: okay
11:11 em: and you just initialise it by using a quote?
11:11 asumu: You initialize it with some list of pairs.
11:11 asumu: Which can be a quoted literal list of pairs or just some list of pairs you have somewhere (e.g. in a variable).
11:11 em: do you have to use cons for that or can you type out (key . "value") ?
11:12 asumu: '(key . value) is a literal pair.
11:12 asumu: So yes.
11:13 em: so you have to put a quote then
11:13 em: i don't get what you said yes to given the way i asked the question, sorry
11:14 em: asumu: but it looked like your key was not a string.
11:14 em: when you did (make-hash '((a . "banana") .. )
11:15 em: you just typed (a . "banana") and it recognised that as a key-value pair.
11:15 em: and the 'a' was not a string.
11:20 asumu: em: that's because the whole list of pairs is quoted
11:21 asumu: rudybot: '((a . "blueberry") (b . "raspberry"))
11:21 rudybot: asumu: ; Value: ((a . "blueberry") (b . "raspberry"))
11:21 asumu: rudybot: '(a . "blueberry")
11:21 rudybot: asumu: ; Value: (a . "blueberry")
11:21 asumu: rudybot: (a . "fruit")
11:21 rudybot: asumu: error: #<undefined>:1:0: application: bad syntax in: (a . "fruit")
11:21 em: okay
11:22 em: asumu: is there a way to create a field like '((future . place-holder).. )
11:22 em: where place-holder is some null value thing that i will set later?
11:22 em: like say i don't know a phone number.
11:22 em: '((phone . ?) .. )
11:23 em: or maybe just give it a value like 000000000 and overwrite it huh?
11:23 asumu: em: there is no special null value in Racket.
11:23 asumu: Most people use #f for that.
11:23 em: oh okay.
11:23 asumu: Unless #f is something that is otherwise meaningful.
11:23 asumu: In which case there are some tricks you can play to use an <undefined> value, but it's better to avoid that.
11:24 em: so it would be okay to put like '((phone . #f) .. ) but later on overwrite that with (phone . 212-343-3452 ) etc
11:24 em: just as an initial value
11:24 em: when you do (make-hash) with no extra stuff it seems to make a null hash.
11:25 asumu: Well as long as you design your program with that data definition in mind, that should work.
11:25 asumu: Since #f isn't a valid phone number ever.
11:54 em: Im on Ubuntu, what is (current-seconds) measuring from?
11:54 bremner: probably Jan. 1, 1970 or so.
11:55 em: does the key part of a key-value pair in a hash have to be a string?
11:55 em: I want to make a hash of comments where it is like '(time-of-comment . "my comment goes here it is one big string")
11:57 em: sorry i just realised i can answer some of these questions just by trying it in the repl
11:59 (join) jeapostrophe
12:03 asumu: Yea, REPLs are great.
12:03 bremner: rudybot: init racket
12:03 rudybot: bremner: your sandbox is ready
12:03 bremner: rudybot: eval (- 2012 (/ (current-seconds) (* 24 3600.0 365.25)))
12:03 rudybot: bremner: ; Value: 1969.7531302443786
12:03 bremner <- too lazy to leap years properly
12:04 bremner: hmm. kindof better without "do"
12:05 jamessan: personally, I only leap seconds :)
12:05 em: i look before i leap years
12:06 bremner: good plan. Otherwise you may end up getting old.
12:07 (quit) asdfhjkl: Quit: Leaving
12:09 bremner: jamessan: I have a sneaking suspicion that the problem with libracket3m for $USER is that they are trying to link it to a shared library. In which case it might make more sense to provide libracket3m.so as a private library. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=666201
12:10 em: rudybot: init racket
12:10 rudybot: em: your sandbox is ready
12:10 em: rudybot: eval (let ((time-of-post (current-seconds)))  (make-hash '((time-of-post . "this is my comment"))))
12:10 rudybot: em: ; Value: #hash((time-of-post . "this is my comment"))
12:10 em: i'm trying to get the key to be the time in seconds when the comment was made
12:10 em: but it is just saying 'time-of-post'
12:14 em: how do you get the key of a hash to be the time/date ?
12:14 em: at the moment you are setting it
12:15 asumu: em: it's because `time-of-post` is a quoted symbol, not the value stored in the variable.
12:15 asumu: You want to use quasiquote and unquote.
12:15 asumu: rudybot: (let ((time-of-post (current-seconds)))  (make-hash `((,time-of-post . "this is my comment"))))
12:15 rudybot: asumu: ; Value: #hash((1333210544 . "this is my comment"))
12:16 asumu: see http://docs.racket-lang.org/guide/qq.html
12:16 (join) mithos28
12:16 em: asumu: holy cow
12:19 (join) HBu
12:20 HBu: hi
12:22 HBu: can anyone tell me how to change the user interface language in the drracket ide? it's always in german but I want the interface in english
12:22 mithos28: are you running linux, windows, or mac?
12:22 HBu: starting from the terminal with LANG=en_US.UTF-8 set doesn't work either
12:22 HBu: mac
12:23 mithos28: is osx set to german?
12:23 asumu: HBu: you should be able to change it from the help menu
12:24 em: is it safe for me to assume that if I leave it as (current-seconds) that is giving me some integer. and there will be some built in racket procdure for turning that integer into a human readable date?
12:25 HBu: @asumu thanks, i didn't think of that
12:25 HBu: *facepal*
12:25 HBu: m
12:25 HBu: bye
12:25 (part) HBu
12:27 jamessan: bremner: I guess I'm not sure how that'd help, but it's also not clear to me what upstream's stance on linking against the lib is.  I do know that Vim provides mzscheme bindings, but I've never enabled them in my packaging due to my uncertainty about the lib
12:28 bremner: I guess it's documented: http://docs.racket-lang.org/inside/index.html
12:28 jamessan: yeah, just found that
12:29 em: rudybot: init racket
12:29 rudybot: em: your sandbox is ready
12:29 em: rudybot: eval (date->string (seconds->date 1333210982))
12:29 rudybot: em: error: reference to an identifier before its definition: date->string in module: 'program
12:29 em: eval (require date/time)
12:29 em: rudybot: eval (require date/time)
12:29 rudybot: em: error: #<undefined>:1:9: date/time: standard-module-name-resolver: collection not found: "date" in any of: (#<path:/mnt/racket-5.2.900.1/collects>) in: date/time
12:30 em: rudybot: evalu (require racket/date)
12:30 rudybot: em: eval (require racket/date) (date->string (seconds->date (current-seconds)))
12:30 (quit) jeapostrophe: Ping timeout: 252 seconds
12:30 em: rudybot: eval (require racket/date)
12:30 rudybot: em: Done.
12:30 em: rudybot: eval (date->string (seconds->date 1333210982))
12:30 rudybot: em: error: car: expects argument of type <pair>; given: #f
12:30 em: that should have worked
12:34 (join) anRch
12:34 (join) RacketCommitBot
12:34 RacketCommitBot: [racket] plt pushed 3 new commits to master: http://git.io/ie_JLQ
12:34 RacketCommitBot: [racket/master] Uncomment & fix racket/draw contracts - Asumu Takikawa
12:34 RacketCommitBot: [racket/master] Remove dynamic checks from racket/draw. - Asumu Takikawa
12:34 RacketCommitBot: [racket/master] Update racket/draw contracts. - Asumu Takikawa
12:34 (part) RacketCommitBot
12:36 (join) realitygrill
12:39 rapacity: looks like a bug
12:39 em: offby1: we found a bug
12:39 em: rapacity: but what i really am trying to get is, how do i convert the given seconds into not just the year-month-day but also the time of the day.
12:39 mithos28: I think it is with the contract library
12:40 rapacity: rudybot: (version)
12:40 rudybot: rapacity: ; Value: "5.2.900.1"
12:40 mithos28: cky showed a similar bug last night with vector->list from srfi/43
12:40 em: (date->string (seconds-date 1333210982)) is only returning 2012-03-31
12:41 em: i want to get 2012-03-31 hour:minute
12:41 em: anyone know how ?
12:41 em: it doesn't have to be that order, just to include year, month, day, hour, minute at least
12:42 rapacity: rudybot: (seconds->date 1333210982)
12:42 rudybot: rapacity: ; Value: #(struct:date* 2 23 16 31 3 2012 6 90 #f 0 0 "UTC")
12:42 (join) mmajchrzak
12:42 rapacity: rudybot: (match (seconds->date 1333210982) [(struct* ([hour h])) h])
12:42 rudybot: rapacity: error: #<undefined>:1:35: struct*: bad syntax in: (struct* ((hour h)))
12:42 em: rapacity: so date->string loses the hours and minutes?
12:43 rapacity: rudybot: (match (seconds->date 1333210982) [(struct* date ([hour h])) h])
12:43 rudybot: rapacity: ; Value: 16
12:43 rapacity: I don't know how to make date->string outpu the hours/minutes
12:43 mmajchrzak: Hi . I am a newbie. Can any one tell me how i can generate list with element from 50 to 50000 (50 100 150 200 ...) by lambda and build-list ?
12:44 rapacity: rudybot: (match (seconds->date 1333210982) [(struct* date ([hour h] [minute m])) (format "~a:~a" h m)])
12:44 rudybot: rapacity: ; Value: "16:23"
12:44 em: i dont know what you did there.
12:45 asumu: mmajchrzak: the `range` function might be easier to use.
12:45 asumu: rudybot: doc range
12:45 rudybot: asumu: http://docs.racket-lang.org/reference/pairs.html#(def._((lib._racket%2Flist..rkt)._range))
12:45 rapacity: ah
12:45 asumu: mmajchrzak: Oh shoot, I think it's not in the latest release version...
12:46 rapacity: date->string takes an optional parameter
12:46 em: rapacity: anyway if my program is saving the seconds then i can always figure out how to format that data later right>
12:46 em: ?
12:46 rapacity: (date->string (seconds->date x) #t)
12:46 (quit) dented42: Quit: Computer has gone to sleep.
12:46 rapacity: that outputs the time
12:47 asumu: rudybot: (build-list (lambda (n) (* n 50)) (/ 50000 50))
12:47 rudybot: asumu: error: build-list: expected argument of type <exact-nonnegative-integer>; given: #<procedure>
12:47 asumu: rudybot: (build-list (/ 50000 50) (lambda (n) (* n 50)))
12:47 rudybot: asumu: ; Value: (0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900 1950 2000 2050 2100 2150 2200 2250 2300 2350 2400 2450 2500 2550 2600 2650 2700 2750 2800 2850 2900 2950 3000 3050 3100 3150 3200 3250 3300 3350 3400 3450 3500 3550 3600 3650 3700 3750 3800 3850 3900 3950 4000 4050 4100 4150 4200 4250 4300
12:47 mmajchrzak: thank you . Its great :)
12:47 em: rapacity: oh i saw that but i put the #t on the inside of the (seconds->date ..) procedure and not the (date->string ..) procedure!
12:47 em: rapacity: thanks.
12:48 em: rapacity: do you know if there is a way, in Dr. Scheme to 'get back' something you wrote in a previous line, sort of like using the up-down arrows in bash?
12:48 em: Dr. Racket that is.
12:48 rapacity: alt+p maybe? though I'm using emacs-style keybindings
12:50 em: hey alt-p worked!
12:54 rapacity: great
13:01 mmajchrzak: If i would like to make a plot with numbers from my list then i make something like that (plot (lambda (x) x) (mylist)) #:x-min 0 #:x-max 50000) but it doesnt work . Why ?
13:09 (join) Kaylin
13:13 (join) RacketCommitBot
13:13 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/uHPxPw
13:13 RacketCommitBot: [racket/master] select the contents of the search string in the 'search in files' dialog - Robby Findler
13:13 (part) RacketCommitBot
13:18 (quit) gridaphobe: Remote host closed the connection
13:18 mmajchrzak: I read a doc from plot but I still dont know how to plot my numbers from my list .
13:23 mithos28: Is it possible to create mutually recursive contracted structs? I am having issues because they are very eager in evaluating the contract binding.
13:23 mithos28: https://gist.github.com/2266898
13:25 em: for "write-to-file" i want to replace old data with new data. Should I use #:exists 'truncate  or #:exists 'replace ?
13:26 mithos28: depends on whether or not you want to keep the same file (i.e. in ode)
13:26 mithos28: inode*
13:26 (quit) Skola: Quit: leaving
13:26 em: i am just trying to save a big hash so i can close the program and start again when i open it.
13:26 mithos28: Either should work, I would go with replace
13:27 mithos28: Also they require slightly different permissions, but I don't think you will need to deal with that
13:28 em: I will have some hash saved as text in a file. Then when I start my program it will read that file to define the master hash. I will work on the master hash inside the program, and along the way the master hash with change values, at the end (or ideally if i can figure out how, after so many minutes) it will write the new (current) version of the master hash back to the same file name.
13:28 (quit) anRch: Read error: Connection reset by peer
13:29 mithos28: are you manipulating the file from multiple programs or reading it while it is being written to? If no, than either will give the same behaviour
13:30 em: yes oka, Imn not doing those things.
13:31 em: it is just one file. I think the way to 'write it' is with (write-to-file master-hash /path/to/file/master.txt #:exists 'replace)
13:31 em: do you happen to know how to read it back in? do i have to open ports and all that?
13:32 mithos28: (call-with-input-file "/path" read)
13:33 rapacity: (file->value "/path/to/file/master.txt") works as well
13:33 mithos28: yeah that is better
13:36 em: rapacity: does that automatically take care of the opening and closing ports and all that?
13:36 mithos28: yep
13:37 rapacity: yeah
13:38 em: the "/path/to/file/master.txt" is with respect to your home directory right?
13:38 em: so you would not do /home/em/data/master.txt  you could just do /data/master.txt ?
13:39 mithos28: no
13:39 em: you put in the full path?
13:39 mithos28: or use ~
13:39 em: oh i didn't realise racket understood what ~ meant
13:40 mithos28: It shouldn't need to, just pass it to the os which understands
13:40 mithos28: unless it is a shell thing
13:41 mithos28: there is expand-user-path aswell
13:41 mithos28: which does the expansion if it doesn't work without it
13:41 (join) samth
13:43 mithos28: even adding recursive-contract to my example doesn't fix it
13:46 mithos28: this seems like an error because recursive-contract, says that it delays evaluation till the contract is checked
13:46 mithos28: and I'm never checking the contract
13:59 mithos28: but eta expansion works as expected, that is weird
14:01 asumu: mmajchrzak: you probably want to use the `points` renderer in plot.
14:03 (quit) mmajchrzak: Ping timeout: 265 seconds
14:05 (quit) rudybot: Remote host closed the connection
14:05 (join) rudybot
14:06 offby1: em: thanks, noted: http://goo.gl/zDIRm
14:07 em: mithos28: is it possible to build a string in your program, and give it a name like my-string and then use that in write-to-file like: (write-to-file value ~/my-string)
14:07 (join) jeapostrophe
14:07 em: in other words have the name of the file be a variable that you determine in your program
14:08 mithos28: use string-append
14:12 em: mithos28: thanks
14:12 em: offby1: im kind of getting my program to work now. It's going to be great.
14:13 offby1: \o/
14:15 (quit) Kaylin: Quit: Leaving.
14:30 (join) mmajchrzak
14:35 (join) dented42
14:36 (join) asdfhjkl
14:37 (join) yoklov
14:45 (join) Shvillr_
14:45 (quit) Shviller: Disconnected by services
14:45 (nick) Shvillr_ -> Shviller
14:47 (quit) Shvillr: Ping timeout: 265 seconds
14:47 (join) Shvillr
14:57 (quit) jeapostrophe: Ping timeout: 252 seconds
14:58 mmajchrzak: Anyone here use plot with list of numbers ?
15:03 (join) jeapostrophe
15:08 (quit) realitygrill: Quit: realitygrill
15:17 (quit) dnolen: Ping timeout: 272 seconds
15:20 (part) scyrmion
15:55 (join) realitygrill
15:55 (quit) realitygrill: Client Quit
16:17 (join) bremner`
16:26 bremner: mmajchrzak: I use it to plot histograms
16:27 mmajchrzak: bremner Could you tell me how to plot a list of numbers ?
16:27 bremner: mmajchrzak: I'm not sure what that means.
16:28 bremner: here is what I know, pretty much http://www.cs.unb.ca/~bremner/blog/posts/racket-histogram/
16:29 mmajchrzak: bremner i have list of numbers like (list 10 20 30 40 50 ) and i would like to plot them like function f(x)=(list-ref (my-list) x)
16:31 bremner`: so just (define (f x) (list-ref my-list x)) and then plot it?
16:32 bremner: hmm. Not sure how you tell plot to only plot integers.
16:32 (quit) bremner`: Quit: Coyote finally caught me
16:32 (join) bremner`
16:33 mmajchrzak: bremner maybe plot only points and after that connect them by a line ?
16:35 (quit) fftb: Read error: Connection reset by peer
16:35 (join) dee
16:36 (join) fftb
16:37 dee: what's the most idiomatic way to draw Mandelbrot's fractal in racket?
16:39 mmajchrzak: dee i have a example...give me a one minute
16:39 dee: ok, thanx
16:40 mmajchrzak: dee http://pastebin.com/AD5UY8SU  i found that on a roseta code repository.
16:42 dee: this is ok, but how to draw it using 2htdp drawing lib?
16:46 dee: beside, i love for/fold construct and got the feeling that is possible to code mandelbrot elegantly with for/fold...
16:48 dee: but, because of "sound of silence" here, the most probable outcome will be to code it in the nicest language ever invented: in Shen! :)
16:48 dee: Shen rules!
16:48 dee: HAHAHAHAHA
16:49 dee: rudybot: shen the
16:49 rudybot: dee: shen, web doc sucks etc, and ban on that
16:49 dee: oh, how clever idea, rudybot!
16:50 mmajchrzak: dee I told you that this isn't mine. But if you write your version with for/fold then add alternative version to http://rosettacode.org/wiki/Mandelbrot_set#Racket
16:50 dee: i sure will, maybe in Shen
16:51 mithos28: mmajchrzak: ignore him
16:51 dee: hi mithos
16:51 dee: why he should ignore me??
16:51 dee: i'm chater just like you
16:52 (quit) fftb: Remote host closed the connection
16:53 dee: hahaha, samth has banned all my country from this channel, all because of your blind hate directed on me!
16:53 dee: oh god, hehehehe
16:53 (join) tommc
16:54 dee: congratulations, community! that's the job! :)))
16:56 dee: rudybot: racketnoob
16:56 rudybot: dee: For all we know, Racketnoob is trolling #shen and saying that Racket is better!
16:56 dee: rudybot: racketnoob the
16:56 rudybot: dee: you should drink at a pub, not in other people's workplaces
16:56 dee: hahahahaha, poor boys!
16:56 (quit) dous: Remote host closed the connection
16:57 dee: eli: hi, dude!
16:57 dee: chandler: hi, pal!
16:57 dee: bremner: hi, stud!
16:58 dee: offby1: hy, bro! :)
16:58 dee: noam: long live Scala!
16:59 dee: samth: ola, you artificial ban-loving scientist! :)))
16:59 (join) dous
17:00 dee: stamourv: hating isn't good for health, dude! :)
17:00 dee: asumu: hi, IP address genius!
17:02 (quit) jeapostrophe: Ping timeout: 244 seconds
17:04 dee: good night, racket community! Good day, Shen!
17:04 (part) dee
17:15 mmajchrzak: How to run executable racket file on linux ? Program works fine but when i create executable file then nothing happend
17:16 mithos28: mmajchrzak: It should just work. I assuming you are using the create executable from DrRacket
17:17 mmajchrzak: mithos28 yes i created stand-alone file with DrRacket.
17:18 mithos28: what happens when you try to run it?
17:18 mithos28: nothing, and the prompt comes back?
17:19 mmajchrzak: mithos28 nothing
17:19 mithos28: what is the error code?
17:20 mithos28: run 'echo $?' right afterwords, before any other command
17:20 mithos28: also what is your program supposed to do?
17:20 chandler: mmajchrzak: Does the program do anything if you run it through "racket foo.rkt" (where foo.rkt is the filename of your program)?
17:21 (quit) bremner`: Quit: Coyote finally caught me
17:21 (join) bremner`
17:22 (quit) dous: Remote host closed the connection
17:22 mmajchrzak: chandler if i run him from terminal he works fine .
17:25 mmajchrzak: mithos28 My program make a plot and after that saves this plot with a png file in the same place where is source.
17:26 mithos28: is it using a gui at all? That sometimes is an issue, and you need to select that when build the executable
17:28 mmajchrzak: mithos28 aaa i see now . Thank you . That was the issue . Thank you and sorry for problem . I am a tottaly newbie in scheme and DrRocket
17:29 mithos28: No problem
17:29 mithos28: I hang out here to help people out
17:38 mmajchrzak: mithos28 Are you using racket in your work or this is only a hobby ?
17:38 mithos28: I wish I used it in my work
17:39 mithos28: Its just a hobby
17:40 mithos28: I also TAd many classes that were taught in Racket
17:41 mithos28: but the real world doesn't appreciate good languages
17:41 mmajchrzak: mithos28 I am a newbie but for now i prefer racket from lisp . Its more... hmm ...clearer. But I think that this is a only academic use language. For now of course. Maybe some one make a nice web framework like rails for ruby
17:43 mithos28: The problem is that you need a community and good standard libraries. And racket has these but they are just a lot smaller than other languages
17:43 mithos28: Racket is also still trying to figure out their story on distributing packages
17:44 mithos28: Hopefully planet v2 will be out soon
17:44 mmajchrzak: So you dont see a bright future for racket ?
17:44 mithos28: I don't see it becoming as popular as say ruby or python
17:45 mithos28: but I wouldn't say that the future doesn't look good for racket
17:48 mmajchrzak: I think that ruby is a good example for be popular in  the reason of only one technology .
17:49 mithos28: I don't get the craze of ruby, rails is not that great a framework
17:49 mithos28: I worked with it for two summers
17:52 mmajchrzak: I prefer php frameworks . But on my university i have Django
17:55 mmajchrzak: And i think that django hava a one big fault...his documentation . But if python  or ruby or even Smalltalk have theirs own frameworks then why not racket ?
18:00 mithos28: It does have a web framework
18:00 mithos28: Why it hasn't seen high adoption rates I don't know
18:01 mmajchrzak: Do you know name of this framework ?
18:02 samth: mmajchrzak, it doesn't really have a name -- it's just built in to the racket web server
18:02 mithos28: It is just the web-server libraries in racket
18:02 mithos28: http://docs.racket-lang.org/web-server/index.html
18:02 mithos28: as samth said
18:03 (quit) mceier: Quit: leaving
18:04 mmajchrzak: i see . It looks that it is very basic web-server .
18:07 (join) jeapostrophe
18:09 samth: mmajchrzak, in what sense is it basic?
18:09 mithos28: samth: I think it is that it doesn't do any orm stuff that most other web frameworks do
18:13 (quit) jeapostrophe: Ping timeout: 264 seconds
18:15 mmajchrzak: But from those basic stuff someone can create a nice and useful web framework
18:18 samth: ah yes, it's basic as a web development framework like django or rails
18:19 mithos28: does anyone use rails without activerecord?
18:33 samth: i don't think so
18:37 (quit) Nisstyre: Read error: Connection reset by peer
18:39 (join) Nisstyre
18:49 (quit) Nisstyre: Quit: Leaving
18:52 (join) Nisstyre
18:59 (quit) bremner`: Quit: Coyote finally caught me
18:59 (join) bremner`
19:07 (join) dnolen
19:08 (join) lewis1711
19:15 (join) dous
19:15 (quit) tommc: Remote host closed the connection
19:19 (quit) dous: Ping timeout: 246 seconds
19:27 (join) startling
19:36 (quit) mmajchrzak: Ping timeout: 246 seconds
19:38 (quit) dnolen: Ping timeout: 252 seconds
19:39 (join) Lunaqus
19:43 Lunaqus: so whats special about racket?
19:44 lewis1711: Lunaqus: it uses lots of round brackets. it has a contract system.
19:45 Lunaqus: a contract system?
19:45 asumu: Lunaqus: It's a general-purpose & kitchen-sink Scheme/Lisp that comes with lots of libraries and language features.
19:46 asumu: And for the so-inclined, it's a "programming language language".
19:46 Lunaqus: would it be any good for doing IRC bots?
19:46 asumu: rudybot: (banner)
19:46 rudybot: asumu: your sandbox is ready
19:46 rudybot: asumu: ; Value: "Welcome to Racket v5.2.900.1.\n"
19:46 asumu: ^ as you can see, yes.
19:47 asumu: offby1 is your man.
19:47 Lunaqus: thats a full sandbox?
19:47 asumu: Yep, try it out.
19:47 mithos28: rudybot: (let loop (()) (loop))
19:47 rudybot: mithos28: your sandbox is ready
19:47 rudybot: mithos28: error: #<undefined>:1:11: let: bad syntax (not an identifier and expression for a binding) at: () in: (let loop (()) (loop))
19:47 mithos28: rudybot: (let loop () (loop))
19:48 rudybot: mithos28: error: with-limit: out of time
19:48 mithos28: prevents you from crashing it
19:50 rudybot: Also my fearsome countenance.
19:50 offby1: rudybot: will you help Lunaqus in his quest?
19:50 rudybot: *offby1: it uses lots of round brackets. it has a contract system.
19:50 offby1: rudybot: we already know that
19:50 rudybot: *offby1: Seems, that someone has already struggled with free-vars and gave up, according to these thread and piece of code: http://list.cs.brown.edu/pipermail/plt-scheme/2008-March/023857.html and http://socrates-lang.cvs.sourceforge.net/socrates-lang/socrates/free-vars.ss?revision=1.5
19:53 lewis1711: it's full featured unlike many schemes, and cross platform unlike many common lisps. that's the shallow reason I like it. also I prefer scheme style syntax
19:59 offby1: also, it has rudybot
19:59 rudybot beams
19:59 offby1: APRIL FOOL
20:08 (quit) samth: Ping timeout: 246 seconds
20:10 Lunaqus: how many of lines of code does it take to write a client that can win at idlerpg?
20:10 Lunaqus: not time out atleast ;)
20:12 lewis1711: is there anyway to have something in a procedure that keeps state? like a static field?
20:12 offby1: lewis1711: sure
20:12 offby1: lewis1711: easiest way is to just whack a global
20:13 offby1: hopefully you have enough taste to find that disagreeable :)
20:13 mithos28: rudybot: (define counter (let ((v 0)) (lambda () (begin0 counter (set! counter (add1 counter))))))
20:13 rudybot: mithos28: Done.
20:13 lewis1711: lol
20:13 mithos28: rudybot: (counter)
20:13 rudybot: mithos28: error: add1: expects argument of type <number>; given: #<procedure:counter>
20:13 offby1: lewis1711: yeah, mithos28's way is the traditional not-so-horrible way
20:13 mithos28: if you use the right variable names
20:13 offby1: once he fixes the spelling errors
20:13 mithos28: rudybot: (define counter (let ((v 0)) (lambda () (begin0 counter (set! counter (add1 v))))))
20:13 rudybot: mithos28: Done.
20:14 mithos28: rudybot: counter
20:14 rudybot: mithos28: ; Value: #<procedure:counter>
20:14 offby1: lewis1711: paul graham has a whole essay about judging languages by how well they handle this problem
20:14 mithos28: rudybot: (counter)
20:14 rudybot: mithos28: ; Value: #<procedure:counter>
20:14 offby1: lewis1711: http://paulgraham.com/accgen.html
20:14 lewis1711: begin0 eh..
20:14 lewis1711: hmm
20:14 mithos28: ugh
20:14 mithos28: rudybot: (define counter (let ((v 0)) (lambda () (begin0 v (set! counter (add1 v))))))
20:14 rudybot: mithos28: Done.
20:14 mithos28: rudybot: (counter)
20:14 rudybot: mithos28: ; Value: 0
20:14 offby1: \o/
20:14 mithos28: rudybot: (counter)
20:14 rudybot: mithos28: error: procedure application: expected procedure, given: 1 (no arguments)
20:15 offby1: *snicker*
20:15 mithos28: lol
20:15 mithos28: rudybot: (define counter (let ((v 0)) (lambda () (begin0 v (set! v (add1 v))))))
20:15 rudybot: mithos28: Done.
20:15 offby1: mithos28: once you get this working, do it again without using set! (or calls to other processes, or writes to the file system)
20:15 mithos28: rudybot: (counter)
20:15 rudybot: mithos28: ; Value: 0
20:15 mithos28: rudybot: (counter)
20:15 rudybot: mithos28: ; Value: 1
20:15 asumu: offby1: I find it odd that the Scheme solution uses set!.
20:16 lewis1711: does the "static variable" have to be a let field?
20:16 asumu: Or is it supposed to be mutable to solve the exercise?
20:16 mithos28: rudybot: (define counter (let ((v (box 0))) (lambda () (begin0 (unbox v) (set-box! v (add1 (unbox v)))))))
20:16 rudybot: mithos28: Done.
20:16 mithos28: rudybot: (counter)
20:16 rudybot: mithos28: ; Value: 0
20:16 mithos28: rudybot: (counter)
20:16 rudybot: mithos28: ; Value: 1
20:16 mithos28: offby1: there you go
20:17 lewis1711: https://gist.github.com/2269946 not sure if this explains the problem
20:17 lewis1711 looks up begin0
20:17 mithos28: it is begin, but returns the first value in the sequence
20:18 mithos28: offby1: were you originally thinking of call/cc way to do it?
20:20 offby1: rudybot: (define counter (thread (thunk (let loop ([count 0]) (let ([client (thread-receive)]) (thread-send client count)) (loop (add1 count))))))
20:20 rudybot: *offby1: your sandbox is ready
20:20 rudybot: *offby1: Done.
20:20 lewis1711: wait so the point of begin is that every expression but the last is a sideffect?
20:20 lewis1711: rudybot: (+ 1 (begin 3 49))
20:20 rudybot: lewis1711: your sandbox is ready
20:20 rudybot: lewis1711: ; Value: 50
20:20 offby1: mithos28: technically correct but not in the spirit I meant
20:20 offby1: rudybot: (begin (thread-send counter (current-thread)) (thread-receive))
20:20 rudybot: *offby1: ; Value: 0
20:20 offby1: rudybot: (begin (thread-send counter (current-thread)) (thread-receive))
20:20 rudybot: *offby1: ; Value: 1
20:20 offby1: rudybot: (begin (thread-send counter (current-thread)) (thread-receive))
20:20 rudybot: *offby1: ; Value: 2
20:20 offby1: that's the Erlang way to do it
20:20 offby1: no set-! anywhere
20:21 offby1: lewis1711: yes, that is exactly the point of "begin"
20:21 offby1: us die-hard functional-programming aficionados don't use it much
20:22 lewis1711: the exercise? I must be one of the few people that comes here not for school :D
20:23 offby1 raises his hand
20:23 offby1: long done with school, thank god
20:23 offby1: still have nightmares about it :-p
20:24 mithos28: i'm done with school as well, at least for the time being
20:24 asumu: I guess technically I am in school...
20:25 lewis1711: I mean in terms of noobs that ask questions, like me
20:25 mithos28: oh, I still ask questions
20:25 lewis1711: rudybot: (define counter (let ((v 0)) (lambda () (begin0 v (set! v (add1 v))))))
20:25 rudybot: lewis1711: Done.
20:25 lewis1711: I don't get how begin, which seems to have to do with letting some part of a group of sexpressions be side effects, keeps the state of v for each call
20:26 offby1: "begin" doesn't keep the state.
20:26 offby1: the "let" keeps the state.
20:26 offby1: or rather -- the "closure" keeps the state.
20:26 offby1: "begin" simply does one thing after the other, returning the value of the last thing.
20:26 Lunaqus: anyone up for coding a poker dealing bot? I'd be up for paying a comission if anyones interested
20:27 offby1: that should be very easy
20:27 offby1: rudybot: (define *deck* (build-list 52 values))
20:27 rudybot: *offby1: Done.
20:27 offby1: rudybot: (shuffle *deck*)
20:27 rudybot: *offby1: ; Value: (25 1 26 23 37 47 15 48 39 6 36 42 43 16 44 8 17 3 7 49 13 50 19 24 29 35 30 0 31 21 38 20 28 45 34 14 41 46 32 18 5 12 22 51 33 4 2 10 9 27 40 11)
20:27 offby1: now map the integers to "cards", find a way to distribute "cards" to "hands" ... pretty straightforward
20:27 offby1: next find a way to evaluate hands -- also ought to be easy
20:28 mithos28: offby1: that step is a little more tricky
20:28 offby1: rudybot: give *deck* to Penn
20:28 rudybot: *offby1: error: reference to an identifier before its definition: to in module: 'program
20:28 mithos28: especially with wilds
20:28 offby1: rudybot: he's the taller louder of the two
20:28 rudybot: *offby1: talks louder
20:28 lewis1711: offby1: I had no idea that let keeps state
20:29 mithos28: lewis1711: that is how set! works
20:29 mithos28: it just modifies what value is bound to the variable from let
20:29 offby1: lewis1711: it does -- what you've got there is the classic example of a closure.  Study it well; it comes up all over the place (in real languages)
20:29 mithos28: and in bad languages, but they have to fake it
20:30 offby1: I was about to say
20:31 lewis1711: ok right so begin is just a tricky smoke screen
20:31 lewis1711: I'm not sure how to do it in my example, because the end procedure may or may not mutate the state
20:31 lewis1711: so I dunno whether to use begin or not
20:31 mithos28: (begin e) is the same as e
20:32 mithos28: so there shouldn't be any harm in using begin
20:32 offby1: lewis1711: it's not tricky, and it's not a smoke screen.
20:32 offby1: It's a way to ensure that expressions get evaluated in order.
20:33 offby1: and that's all it is.
20:33 asumu: lewis1711: re: let & state, the reference page on the evaluation model might help http://docs.racket-lang.org/reference/eval-model.html
20:33 offby1: you could always use (let () this that the-other); in fact it's possible that (begin this that the-other) macroexpands into that.
20:34 asumu: In particular, the section on imperative update explains how mutation on function arguments actually works.
20:34 lewis1711: offby1: ok that makes sense
20:34 lewis1711: but here
20:34 asumu: (though it's kinda dense too)
20:34 lewis1711: https://gist.github.com/2270048
20:34 lewis1711: "msg" may or may not mutate state
20:34 lewis1711: ie it only will if it is sent 'make-slot
20:36 offby1: from what I can see, "msg" will never mutate state
20:37 (join) dnolen
20:37 offby1: but I think I can guess what you meant.
20:38 lewis1711: offby1: what if msg is 'make-slot
20:38 offby1: that's what I meant by guess.
20:38 offby1: except I tried it and it didn't do what I expected.
20:39 offby1: http://ix.io/2gm
20:39 (join) lebro
20:39 offby1: see if you can guess what I expected from that transcript.
20:39 lewis1711: offby1: that's what I am expecting as well
20:39 lewis1711: that's the whole issue
20:40 offby1: OH, so you want me to DEBUG it, EH?!
20:40 lewis1711: lol
20:40 lewis1711: so are you saying, in your eyes, that it SHOULD keep state?
20:40 lewis1711: slots I mean
20:42 offby1: this sorta works -- http://ix.io/2gn
20:43 offby1: but I'm not sure it's what you ultimately want
20:43 offby1: yes, I think _something_ should keep state.
20:43 offby1: otherwise, "make-slot" has no real effect anywhere
20:43 offby1: ultimately I don't know what "root" is.  Perhaps if the function were named "new-object" or something.
20:44 offby1: you should write your tests first, that way you have a design in view while you code.
20:44 offby1: otherwise (if you're like me, anyway) you keep forgetting whether you're writing code for a single object, as opposed to for the entire system ...
20:45 lewis1711: offby1: it's for a single object. for a prototype system
20:46 lewis1711: also that lambda thing means any other message returns a procedure
20:47 offby1: you're the one who put all those lambda in there
20:47 (quit) masm: Quit: Leaving.
20:49 lewis1711: no I mean now any message returns a procedure of a procedure..
20:56 (quit) asdfhjkl: Quit: Leaving
20:57 offby1: wasn't that the case originally?
20:58 lewis1711: no
20:58 offby1: ok ok so http://ix.io/2go
20:58 lewis1711: before I could do ((root 'responds to)) and it would return a quoted list
20:58 offby1: whiner
21:00 lewis1711: i don't get that at all
21:00 lewis1711: how does 'responds-to work with that
21:00 lewis1711 tries to figure it out
21:00 lewis1711: I'm not whining, just useless
21:01 offby1: Try It And See™.
21:01 offby1: rudybot: (define root (let ([slots (make-hash)]) (hash-set! slots 'responds-to (λ() (hash-keys slots))) (hash-set! slots 'make-slot (λ (key value) (hash-set! slots key value))) (hash-set! slots 'clone (λ() ("implement me!"))) (lambda (msg . args) (apply (hash-ref slots msg) args))))
21:01 rudybot: *offby1: Done.
21:01 lewis1711: I am trying it
21:01 offby1: rudybot: (root 'make-slot 'poop (lambda () "Here I go!"))
21:01 rudybot: *offby1: Done.
21:01 offby1: rudybot: (root 'responds-to)
21:01 rudybot: *offby1: ; Value: (poop make-slot responds-to clone)
21:01 offby1: *shrug*
21:01 lewis1711: ((root 'responds-to) 'responds-to)
21:02 lewis1711: what how did you do that
21:02 lewis1711: ok I've missed something
21:02 offby1: you saw everything I did.
21:02 offby1: it's all out in the open.
21:02 offby1: albeit poorly formatted.
21:02 lewis1711: yeah you changed something I missed I mean
21:02 offby1: diff mine against yours.
21:04 lewis1711: oh right
21:04 lewis1711: ok that's not quite how I wanted it, because now the call is different but I will figure it out. thanks for your help
21:10 lewis1711: aha! got it. thanks offby1
21:11 offby1: \o/
21:42 offby1: lewis1711: "clone", although I'm not sure I'm happy with it: http://ix.io/2gp
21:50 (join) wtetzner
21:51 lewis1711: offby1: to be clear "root" is kind of an internal definition, I have a defobject macro that does the nice syntax transformation (with a bit of help from someone here of course)
21:51 lewis1711: but I will see what I can understand from your clone
21:52 offby1: I'd probably make objects be structs, and keep state more or less explicitly
21:52 offby1: might not use closures, or at least, might not use them for everything
21:53 lewis1711: why structs over hashtables? I thought about that too btw
21:54 offby1: dunno really.
21:54 offby1: just seems more "right"
21:54 lewis1711: javascript and self use something like hashtables IIRC. which is why i went with it
21:54 offby1: sure
21:54 lewis1711: not a very good reason, but I guess neither is yours :D
21:56 offby1: actually the only struct member I can think of would be "slots", which would be ... a hash table.
21:56 offby1: sorta makes the struct pointless.
21:57 offby1: I wonder if you could make a new type that is mostly ordinary hash table ... but "hash?" returns #f for it, and it prints like #<your-new-type-name>, but all the existing functions like "hash-set" etc still work on it
22:00 (join) Kaylin
22:03 (join) cmadd
22:04 (quit) cdidd: Remote host closed the connection
22:05 lewis1711: offby1: something like luas metatables?
22:07 offby1: dunno
22:07 offby1: dunno lua
22:07 lewis1711: https://gist.github.com/2270048 my attempt at clone
22:07 offby1: dude.  Where's your tests?!
22:08 offby1: How'm I supposed to tell if it works?
22:08 lewis1711: good thinking
22:08 (quit) bitonic: Quit: WeeChat 0.3.5
22:08 offby1: for that matter ... how are _you_ supposed to tell?
22:09 lewis1711: I've been meaning to get into unit testing. especially for dynamically typed languages
22:09 lewis1711: guess now is as good a time as any
22:13 offby1: never too late etc
22:13 lewis1711: wait, should two cloned objects be equal? or eq? or eqv..? hmm I guess objects are functions so..
22:13 offby1: cosmetic update: http://ix.io/2gq
22:13 lewis1711: (or I have defined them as)
22:15 offby1: my hunch is: they should not be eqv? or eq? or equal?
22:15 offby1: they're different.
22:16 offby1: maaaaaybe equal?
22:16 offby1: probably not possible to actually implement though
22:18 lewis1711: I guess what you did
22:18 lewis1711: which is check to see they respond to the same methods
22:19 lewis1711: (define new (root 'clone))
22:19 lewis1711: (check-equal? ([root 'responds-to]) ([new 'responds-to]))
22:19 offby1: but you'd also want to check that the respond _in the same way_, and that's not possible
22:19 offby1: I suspect that'd be equivalent to solving the Halting Problem
22:20 offby1: s/that the/that they/
22:21 lewis1711: yeah i guess once you clone it you're on your own
22:23 cmadd: I got a quick question about macros
22:23 lewis1711: what does racket use for nil?
22:24 cmadd: is there a reason (syntax-test (a c) b) doesn't match the pattern in this macro:
22:24 cmadd: (define-syntax-rule (syntax-test (a ...) b ...+ ) (b ...+ a ...))
22:31 offby1: rudybot: (null? ' ())
22:31 rudybot: *offby1: ; Value: #t
22:37 (join) dous
22:39 (quit) lewis1711: Ping timeout: 246 seconds
22:41 (quit) dous: Ping timeout: 246 seconds
22:59 (join) realitygrill
23:07 (join) dyoo
23:18 (join) lewis1711
23:19 (quit) dyoo: Ping timeout: 245 seconds
23:22 asumu: cmadd: ...+ isn't something built into syntax-rules, so it's matching it as an identifier.
23:23 mithos28: use b0 b ...
23:23 mithos28: eh we were an hour late
23:24 (join) gridaphobe
23:25 rapacity: there's syntax-parse that has ...+
23:26 asumu: Too bad dyoo isn't here right now.
23:26 asumu: He deserves a virtual high-five for his april fools joke.
23:28 (quit) gridaphobe: Ping timeout: 265 seconds
23:40 asumu: Hmm, it appears Rosettacode used to consider Racket a separate language and now does not. http://rosettacode.org/wiki/Racket
23:43 (quit) lebro: Ping timeout: 246 seconds
23:44 lewis1711: god dman it stat eisn't being kept again
23:44 (join) lebro
23:46 offby1: say what now?
23:47 (join) RacketCommitBot
23:47 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/km_WuA
23:47 RacketCommitBot: [racket/master] improve the gui for specifying the submodules to run - Robby Findler
23:47 (part) RacketCommitBot
23:48 lewis1711: offby1: once again the changes to slot aren't kept between method calls. I have no clue how I got it working in the first place. or why it isn't now.
23:52 (join) dous
23:53 lewis1711: https://gist.github.com/2271048 line 24 check passes, but line 27 fails. can't figure that out