00:07 (quit) ynniv: Quit: ynniv
00:17 (quit) jeapostrophe: Ping timeout: 250 seconds
00:29 (join) sw2wolf
00:59 (join) RacketCommitBot
00:59 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/WaddgA
00:59 RacketCommitBot: racket/master c818eb3 Robby Findler: Fix the new lexer interaction when some of the lexer states were...
00:59 (part) RacketCommitBot
01:12 Sgeo: Is there no (define ((constantly foo) bar) foo) built into Racket?
01:12 Sgeo: I keep seeing examples using lambda directly for that use case
01:13 mithos28: rudybot: doc const
01:13 rudybot: mithos28: your sandbox is ready
01:13 rudybot: mithos28: http://docs.racket-lang.org/reference/procedures.html#(def._((lib._racket%2Ffunction..rkt)._const))
01:13 Sgeo: Oh, ok
01:14 Sgeo: So why do I keep seeing examples like (lambda (stx) (syntax "I am foo")))
01:15 jonrafkind: where
01:15 Sgeo: http://www.greghendershott.com/fear-of-macros/Transform_.html
01:15 Sgeo: Pretty sure I've seen similar in official Racket documentation but don't remember where
01:15 vyz: are there any good libraries for event driven programming in racket?
01:16 mithos28: vyz: What do you mean event driven programming?
01:16 jonrafkind: it would be pointless to abstract a small function to some constant function
01:16 mithos28: there is thunk*
01:16 vyz: something like nodejs's libev, setinterval function to make something happen every n milliseconds
01:18 mithos28: vyz: (thread (thunk (let loop () (sync (alarm-evt (+ n (current-inexact-milliseconds)))) (something) (loop))))
01:18 vyz: i dont think you're human, you are clearly more intelligent
01:19 vyz: mithos, do you write lisp for a living?
01:20 mithos28: sadly no
01:20 vyz: thats surprising
01:20 vyz: although most people dont like the lack of maintainability that lisp codebases introduce
01:20 Sgeo: mithos28, hmm, is it idiomatic in Racket to write code in a synchronous style and just use the light-weight threads for concurrency, rather than explicit CPS node.js style?
01:21 mithos28: Sgeo: Yes, because explicit cps is a pain
01:21 Sgeo: Awesome
01:22 mithos28: and you always have call/cc if you need the continuation, you don't need to cps
01:22 Sgeo: All I know is that Oleg dislikes call/cc.
01:22 Sgeo: How does call/cc interact with threads?
01:22 Sgeo: I should port my mamb operator to Racket
01:22 mithos28: vyz: I use some internal languages which have no equal in terms of unmaintainablity
01:23 Sgeo: Is better monads the preferred monad library?
01:23 mithos28: Sgeo: It works how you would expect
01:23 mithos28: Sgeo: I don't know of a standard monad library
01:24 vyz: thats the problem, its just not feasible to integrate into a system where there's more than 3 or 4 people working on a system, it gets too  fluid
01:24 vyz: typed racket is a welcome component to confront that
01:24 Sgeo: Are exceptions resumable?
01:25 Sgeo: Because I looked at the guide and see no indication that they are.
01:25 Sgeo: Maybe I skimmed too fast?
01:25 mithos28: Sgeo: Sorta
01:25 mye: lately I've been running in a weird drracket bug where it keeps darkening some paren highlights. Anyone has encountered this?
01:25 vyz: i havent
01:25 mithos28: rudybot: doc raise
01:25 rudybot: mithos28: http://docs.racket-lang.org/reference/exns.html#(def._((quote._~23~25kernel)._raise))
01:25 mithos28: Sgeo: Look there more info
01:26 mithos28: The primitive one's are not, but you can make custom ones resumable
01:26 (quit) jonrafkind: Ping timeout: 264 seconds
01:27 mithos28: vyz: I don't see racket any more hard to work with than python for large groups of developers
01:27 vyz: that contradicts the fact that large groups are using python in most cases
01:27 Sgeo: Hmm, does Racket's web framework mandate a continuation-based web site?
01:27 vyz: when was the last time you saw racket code on hackernews or somewhere?
01:27 Sgeo: I think continuations make web sites ... ugly
01:28 vyz: I had an idea to recreate jekyll in racket
01:28 vyz: just generate html with markup
01:28 Sgeo: easier to program, but very ... stateful
01:28 vyz: simple, clean
01:28 mithos28: vyz: Do you have evidence that the reason that people are not using racket because it doesn't scale with large groups, versus some other reason
01:28 vyz: mithos, no and when you put it like that, i may have been misleading in saying that, but I believe it to be true nevertheless
01:29 mithos28: Sgeo: No, you can do standard RESTful stuff, but it emphasizes the continuation part because thats what is special that it brings to the table
01:30 Sgeo: Hygienic macros present a sort of barrier to writing macros, and that barrier might make code easier to understand as fewer macros are made, I think
01:30 Sgeo: (I understand unhygienic macros, hygienic macros ... confuse me)
01:30 mithos28: vyz: What about the Racket dev team?
01:30 mithos28: Sgeo: What is confusing about them, you don't need to do gensym and things work how you expect.
01:31 vyz: open source development is fundamentally dissimilar from closed source corporate development
01:31 Sgeo: mithos28, I expect to be able to run arbitrary code at macroexpand time to transform the input code as I want
01:31 mithos28: Sgeo: You can
01:31 Sgeo is reading some tutorial on Racket's macros now
01:31 mithos28: I think you must be looking at syntax-rules
01:31 vyz: corporate development environments are full of delegation and talking and emails, while open source devs just want the code to be fast and furious
01:32 Sgeo: mithos28, but that mechanism is more ... difficult to understand, dealing with syntax objects, than defmacro
01:32 mithos28: syntax objects are a separate concern from hygiene
01:32 mithos28: they could exist in an unhygenic system to add annotations such as line numbers
01:33 mye: but they are the objects carrying scope info right?
01:34 mithos28: mye: Racket's do, but in theory you could have syntax-objects without hygiene
01:34 mithos28: and certain implementations of hygiene don't use syntax objects
01:35 mye: for me the confusing part was and is with-syntax :) It got better when I read it's just a nested syntax-case
01:37 Sgeo: I guess saying that I like Tcl is not the right thing to say here? >.>
01:37 Sgeo keeps reading about Racket
01:38 mithos28: I dislike Tcl because gitk is written in it, and they somehow cannot figure out how to get it to come to the front when I launch it
01:39 mye: I wish DrRacket had an entry in the context menu when clicking on a macro that would show the expansion inline. That would be awesome.
01:44 mye: sigh, what am I doing wrong when instead of a response? set-servlet-handler! gets a request… This is the most useless blame in a contract violation ever, showing only internal webserver functions
01:45 vyz: i wish drracket didnt abuse my ram
01:46 (quit) mithos28: Quit: mithos28
01:46 Sgeo: I do like how ... easy it sounds to redefine function application and put an entire module inside a form
01:47 Sgeo: I take it that if Racket didn't have first-class continuations, it would be possible to make a language that provided them
01:47 Sgeo: WIthout needing to explicitely wrap everything in some form
01:48 (join) mithos28
01:49 vyz: im just here for the paren
01:51 mithos28: Sgeo: No, because CPSing is a whole program transformation
01:51 mithos28: How would you use 'map' with a function that needed the continuation
01:51 mithos28: without rewriting map
01:59 Sgeo: Does let over defines not work? I can only assume that they don't, otherwise what is splicing-let for?
01:59 mithos28: What do you want to do?
02:00 mithos28: Splicing let allows the defines to be visible outside the let
02:01 Sgeo: Ah
02:01 Sgeo: So defines in a let don't leak out?
02:01 Sgeo: Ok
02:16 dyoo: Sgeo: back at 7pm, when you're asking about that paragraph on Custodians, they mean something like this:
02:16 dyoo: https://gist.github.com/4411396
02:17 dyoo: Where one tries to call things that allocate resources, when the current custodian has already shut down
02:17 Sgeo: Ah, ok
02:17 dyoo: but yeah, that paragraph's confusing.  I'll propose a patch to the docs with this example just to disambiguate.
02:17 Sgeo: Cool
02:18 Sgeo: I'm glad that I indirectly helped out Racket's documentation
02:18 dyoo: :)
02:28 dyoo: Sgeo: ok, I've asked on dev  (http://lists.racket-lang.org/dev/archive/2012-December/011373.html)  So unless I get completely ignored, we'll get this fixed.
02:29 Sgeo: dyoo, thank you
02:30 dyoo: Sgeo: with regards to hooking your own values into custodians, you will probably want to look at: http://docs.racket-lang.org/foreign/Custodian_Shutdown_Registration.html
02:31 dyoo: but that starts touching foreign-function-interface stuff, which I have admittedly not played with enough to be competent.
02:31 Sgeo: Why is that in ffi stuff?
02:33 (join) Nisstyre-laptop
02:33 dyoo: Because the other built-in resources in Racket (files, ports, etc.) already register with the current custodian.  It's when we need to connect to foreign resources
02:33 dyoo: that we need to customize how they shut down under custodian shutdown
02:36 Sgeo: :/ not sure why that should prevent me from defining my own kind of resource that needs management. What about a connection to an IRC server, and when the custodian tells everything to shut down, I want to tell the server good-bye rather than it being abrupt?
02:36 Sgeo: As an example
02:38 dyoo: Sgeo: yes, that makes sense.  But I think the custodian system is sort of the catch-all for managing low-level resources
02:39 dyoo: For the situation that you're considering, that sounds more like something that can be initiated at a higher level, as a shut-down procecdure, or with an exception handler.  That being said, maybe the custodian concept can be expanded to do high-level shutdown.   Couldn't hurt to ask on the mailing list for feedback
02:40 Sgeo: Hmm, mailboxes aren't themselves first-class?
02:40 Sgeo: I know the same situation is in Erlang, but Factor has first-class mailboxes
02:40 dyoo: Sgeo: mailboxes?
02:40 Sgeo: Rather than "one mailbox per thread"
02:41 Sgeo: Oh, there are async channels
02:41 dyoo: Unfortunately, I'm not too familiar with mailboxes from Factor.  Are the similar to "Channels"?  http://docs.racket-lang.org/reference/channel.html?q=channel#(tech._channel)
02:42 Sgeo: More similar to http://docs.racket-lang.org/reference/async-channel.html?q=channel
02:42 Sgeo: Async channels
02:42 dyoo: I see.  Well then, yes.  :)
02:42 dyoo: Ack, gotta go.  Good luck to you.
02:42 (quit) dyoo: Quit: dyoo
02:42 mithos28: Sgeo: Async channels are built upon the same thing as threads, but thread channels have less synchronization overhead on reading
02:43 mithos28: as there is only one possible reader
02:43 mithos28: I believe you are misunderstanding what custodians do, if you want them to handle graceful shutdown
02:44 mithos28: In the situation they are useful, there are 3 parties. The library writer, the library user, and the host.
02:45 mithos28: The host is running code for the library user in its environment. This code may be malicious/buggy
02:46 mithos28: When the library user's code misbehaves, the host wants to get rid of it immediately.
02:47 mithos28: Custodians allow the library writer to tell the host how to clean up all the resources that the user created
02:52 Sgeo: wrap-evt is basically fmap on events?
02:52 mithos28: Sgeo: Yep
02:52 Sgeo: Monad libraries should really consider events to be a monad imo
02:53 Sgeo: Or, well, event..ness
02:54 mithos28: Isn't that FRP?
02:54 Sgeo: Not... sure, but what's wrong with using synchronizable events in that manner?
02:55 mithos28: Nothing, I was just saying that it had been done
02:58 asumu: Sgeo: re: understanding macros, I recommend greghendershott's macro tutorial.
02:58 asumu: http://www.greghendershott.com/fear-of-macros/
03:00 Sgeo: Looked through that
03:02 (quit) mithos28: Quit: mithos28
03:03 Sgeo: How safe is the sandboxed evaluation stuff?
03:04 Sgeo: I don't understand this snippet
03:04 Sgeo: (let ([e (make-evaluator 'racket/base)])
03:04 Sgeo: (e (,e 1)))
03:04 Sgeo: How is , being used outside a ` ?
03:05 asumu: Usually pretty safe. It doesn't allow filesystem access, has timeouts, etc. Sometimes there are bugs.
03:05 Sgeo: Someone should make a try-racket site
03:06 asumu: That example probably has a typo...
03:07 asumu: Yeah, I think it's supposed to have a quasiquote around the (,e 1)
03:08 asumu will fix that doc bug later
03:09 asumu: Sgeo: thanks for finding it
03:10 Sgeo: You're welcome
03:10 Sgeo: http://25.media.tumblr.com/tumblr_m3c7wfyTr91r8lqovo1_400.png
03:10 Sgeo: </homestuck-addiction>
03:16 Sgeo: The contact's -> is going to confuse me because I spent a lot of time playing with Clojure
03:16 Sgeo: *contract
03:21 (join) bitonic
03:24 Sgeo: Minor typo:
03:24 Sgeo: "a name for the function or method associated with the contract and either the phrase ?contract violation? or ?violated it?s contract? depending on whether the contract was violated by the server or the client;"
03:24 Sgeo: it's instead its
03:24 Sgeo: Don't know if that's in docs or in Racket
03:25 Sgeo: (As in, whether docs are accurately reporting on grammar error in Racket, or incorrect)
03:42 (quit) bitonic: Remote host closed the connection
03:46 Sgeo: What happens when I want to mix the laziness of Racket's lazy language with the static typing of Typed Racket?
03:55 Sgeo: rudybot, help
03:55 rudybot: Sgeo: bug <more> ..., help [<what>], version, quote, source, url, seen <nick>, uptime, t8 <from> <to> <text> ..., init [<lang>], eval <expr> ..., give <to> <expr> ..., apropos <str> ..., desc <id>, doc <id>
03:56 (quit) sw2wolf: Quit: ERC Version 5.3 (IRC client for Emacs)
03:56 Sgeo: eval 'test
03:56 Sgeo: ,'test
03:56 Sgeo: rudybot, eval 'test
03:56 rudybot: Sgeo: your sandbox is ready
03:56 rudybot: Sgeo: ; Value: test
03:56 Sgeo: rudybot, eval #:test
03:56 rudybot: Sgeo: error: #<undefined>:1:0: #%datum: keyword used as an expression in: #:test
03:56 Sgeo: rudybot, eval #'#:test
03:56 rudybot: Sgeo: ; Value: #<syntax:1:2 #:test>
03:56 Sgeo: rudybot, eval (syntax->datum #'#:test)
03:56 rudybot: Sgeo: ; Value: #:test
03:56 Sgeo: ?
03:57 Sgeo: Oh, is that the same as '#:test
03:57 Sgeo: ?
03:58 (quit) francisl: Quit: francisl
04:06 Sgeo: Why does frtime need a whole new language?
04:12 Sgeo: Ah, so it's that one could use a behavior as though it were an ordinary value?
04:21 Sgeo: Hmm, or not
04:47 (quit) Nisstyre-laptop: Ping timeout: 250 seconds
04:55 (join) ThePawnBreak
05:06 (join) bitonic
05:11 Sgeo: asdf can't copy from the text editor that slideshow's tutorial opens
05:13 Sgeo: http://pastie.org/5596719
05:13 Sgeo: The two disks drawn are not the same, the lower one is filled in while the upper one is not
05:13 Sgeo: On my system, I mean
05:48 (join) Okasu
05:53 Okasu: Hello, i'm trying to eval region with Geiser but it fails with "#lang not enabled in the current context"? How can i fix it?
06:56 (join) RacketCommitBot
06:56 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/LA4nEw
06:56 RacketCommitBot: racket/master 4041e3c Asumu Takikawa: Fix sandbox doc example
06:56 (part) RacketCommitBot
07:12 (join) MayDaniel
07:20 Okasu: For example i have basic hello-world form "Continue" article, how can i run it without drracket? Plain "racket -f hello-world.rkt" does nothing.
07:21 (join) hash_table
07:42 (join) RacketCommitBot
07:42 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/wS4wdA
07:42 RacketCommitBot: racket/master 05ce8bc Matthew Flatt: win32: fix MinGW vs. MSVC build problem
07:42 (part) RacketCommitBot
07:45 (quit) kofno: Read error: No route to host
07:45 (join) kofno
07:56 (join) add^_
08:07 (join) soegaard
08:49 Sgeo: Would it be a terribad idea to use Datalog as a database?
08:49 Sgeo: Occasionally serializing to disk and reading from disk?
08:51 (join) masm
09:19 (join) masm_
09:19 (quit) masm: Quit: Leaving.
09:20 (join) dzhus
09:22 (nick) masm_ -> masm
09:23 (join) mizu_no_oto
09:27 (join) anRch
09:28 (quit) mizu_no_oto: Ping timeout: 264 seconds
09:40 (quit) kofno: Remote host closed the connection
09:51 (join) masm
09:54 (join) kofno
09:56 (join) mizu_no_oto
09:57 (join) mithos28
10:01 (quit) dzhus: Remote host closed the connection
10:17 (quit) add^_: Remote host closed the connection
10:20 (join) wwj
10:20 (quit) dca: Quit: ZNC - http://znc.in
10:24 Sgeo: I swear I'm going to slap the author of better-monads
10:25 (part) wwj
10:28 Cryovat: What's up, Sgeo?
10:28 (quit) soegaard: Quit: soegaard
10:29 Sgeo: Why is everyone and their mother under the delusion that if they write lift, the user is going to need to pass in the number of arguments the function will take
10:30 Sgeo: There's a handy dandy function that both Clojure and Racket have called "apply". Haskell doesn't have it. Clojure and Racket functions can do things that Haskell functions cannot.
10:30 Sgeo: Including writing a generic lift
10:32 Okasu: Guys?
10:32 Sgeo: Actually, looking at better-monads more closely, it may be so fundamentally broken that it's flat out impossible to even write functions that use bind or return outside of the lexical scope of a with-monad
10:32 (join) jeapostrophe
10:32 (quit) jeapostrophe: Changing host
10:32 (join) jeapostrophe
10:32 Sgeo: I'm not sure, I'm not good with Racket macros
10:33 Cryovat: If you can't do that, it seems like it defeats the point
10:33 Okasu: How can i run example from Continue without drracket?
10:34 Okasu: Only with plain old racket.
10:34 Okasu: One option is to make module out of that example but i wonder if it possible to run it without mods?
10:34 (quit) kofno: Remote host closed the connection
10:35 Sgeo: Cryovat, look at the definition of with-monad in http://planet.racket-lang.org/package-source/toups/functional.plt/1/1/better-monads.rkt
10:35 rudybot: http://tinyurl.com/byclhgl
10:35 Sgeo: It looks like it's just shoving the body into a let form
10:36 Sgeo: Okasu, I looked before, I'm not sure I see a hello world in Continue, unless the web related one?
10:37 Okasu: Sgeo: http://docs.racket-lang.org/continue/#(part._.Getting_.Started) This one.
10:37 Okasu: Sorry for inaccuracy.
10:39 Cryovat: My insufficient macro-foo is confusing me a bit
10:39 Cryovat: But I can't see how that would achieve anything close to Haskell
10:40 Sgeo: Well, the big problem of doing monads in a dynamically-typed language (or a statically-typed language where the types don't change the running program) is the lack of return-type polymorphism
10:40 Sgeo: So to do return, there needs to be another way to pass in the type information
10:41 Sgeo: Here, it's being done with a with-monad macro
10:41 Sgeo: You give with-monad a value representing the monad, and within its scope bind and return are defined
10:41 Sgeo: But... it looks like it's lexical scope. With no way to get inside it outside of with-monad
10:41 Sgeo: So, I don't see a general way to, say, define sequence
10:42 Sgeo: Okasu, try without -f
10:42 Sgeo: racket hello.rkt worked for me
10:43 Okasu: Sgeo: Heh, thanks, works fine for me to. :)
10:43 Cryovat: That seems inconvenient
10:43 Okasu experiencing serious brain lags.
10:44 Cryovat: I don't really see the big need for monads in Scheme
10:44 Okasu: http://dfns.dyalog.com/max_scode.htm http://dfns.dyalog.com/max_examp.htm Interesting.
10:44 Okasu: Cryovat: Agreed.
10:45 (quit) anRch: Quit: anRch
10:46 (nick) Raynes -> baranosky
10:46 Sgeo: better-monads defines a monad called the-syntax-monad
10:47 (join) kofno
10:47 Sgeo: That sounds like it could be useful, if it wasn't trapped in a piece of junk library
10:51 (join) soegaard
10:52 Cryovat: Probably better to steal the idea and start over then ;)
10:52 Sgeo: I was hoping to steal better-monads and implement it in Clojure
10:53 Sgeo: Because the Clojure monad implementations aren't all that great
10:53 Sgeo: But.... no way
10:53 Sgeo: Not now
11:00 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
11:03 (join) jao
11:03 (quit) jao: Changing host
11:03 (join) jao
11:09 (quit) kofno: Remote host closed the connection
11:21 (quit) soegaard: Quit: soegaard
11:29 (join) soegaard
11:54 mye: I've a question about send/suspend and continuations (kinda new to this). In the docs for it it is said send/suspend "returns" the request. These quotes are there literally. send/suspend has a request? return value. So am I right that it returns the value as usual? Or must I do something special to capture the continuation?
11:58 (join) netrino
12:00 (join) cipher
12:04 (quit) mithos28: Quit: mithos28
12:06 (join) mithos28
12:09 (join) mizu_no_oto
12:32 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
12:32 (quit) netrino: Read error: Connection reset by peer
12:40 (join) kofno
12:40 mithos28: asumu: ping
12:43 (quit) kofno: Remote host closed the connection
12:50 (quit) jeapostrophe: Ping timeout: 250 seconds
12:58 (quit) cipher: Read error: Operation timed out
13:14 (join) jonrafkind
13:16 (join) francisl
13:24 (join) dnolen
13:32 (quit) dnolen: Ping timeout: 245 seconds
13:33 (quit) mithos28: Quit: mithos28
13:33 (join) mithos28
13:39 (join) add^_
14:13 (join) blubberdiblub
14:30 (quit) mithos28: Quit: mithos28
14:48 (part) firefux
14:53 (join) RacketCommitBot
14:53 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/Tl_7cA
14:53 RacketCommitBot: racket/master ede0e3c Matthias Felleisen: at JPR's request: modified memq to return boolean and added memq? to teaching languages...
14:53 (part) RacketCommitBot
15:19 (quit) karswell: *.net *.split
15:19 (quit) em: *.net *.split
15:19 (join) em
15:20 (quit) jao: Ping timeout: 260 seconds
15:24 (join) karswell
15:27 (join) nejucomo
15:29 (quit) soegaard: Quit: soegaard
15:39 (quit) francisl: Quit: francisl
15:49 (join) bro_grammer
15:52 (join) soegaard
15:52 (quit) hash_table: Ping timeout: 255 seconds
15:59 (quit) nejucomo: Ping timeout: 276 seconds
16:03 (quit) blubberdiblub: Ping timeout: 252 seconds
16:04 (join) blubberdiblub
16:05 (join) nejucomo
16:18 (quit) Okasu: Quit: Lost terminal
16:23 (join) mye
16:27 (quit) vyz: Read error: Connection reset by peer
16:39 (join) univyrse
16:39 (join) offby1
16:40 offby1: Is there some way to get "raco test" to exit with a non-zero status if some tests failed?
16:48 (join) Nisstyre-laptop
16:50 (nick) univyrse -> vys
16:50 (part) vys: "Leaving"
17:04 (part) nejucomo
17:25 (join) Fare
17:30 mye: when using stateless or stateful servlets, send/formlet seems to use the send/suspend from the stateful implementation, so is that a problem when used in stateless servlets?
17:32 (join) neilv
17:35 (join) anRch
17:38 (quit) masm: Ping timeout: 260 seconds
17:38 asumu: mye: nothing special you have to do. The continuation plumbing is not something you usually have to think about.
17:38 asumu: The URL argument is how you use the continuation
17:42 asumu: offby1: I don't know that it can do that in general, because test submodules can have tests defined in any way (manually, rackunit, etc.)
17:42 asumu: I suppose if you use rackunit, you could call `run-tests` and call `exit` if there are unsuccessful tests.
17:43 (join) masm
17:46 offby1: asumu: yeah, but I assume that if I call "exit" in one file, it'll prevent the other files from running.
17:46 asumu: Ah, yeah, that's true. Hmm.
17:48 offby1: I always get the feeling that I'm not doing tests the "right way", but I've never figured out what the right way is.
17:50 neilv: i released a new version of overeasy the other day that gets closer to right. but i threw out the old stuff for multi-file and haven't replaced that part yet
17:51 neilv: http://www.neilvandyke.org/overeasy/
17:57 Sgeo: Hi neilv. I started working through SICP a while ago (put it down a while ago too, but intend to resume), and am using your SICP language
18:00 (join) francisl
18:03 mye: I don't understand why this code loops the formlet page https://gist.github.com/4415889
18:04 mye: Someone here does?
18:04 neilv: hi Sgeo. i hope that sicp language was of some help. i am looking for someone to take it over and improve it
18:15 Sgeo: I didn't really get that far in SICP, I should start again
18:20 offby1: neilv: cool.  ``Note that a given instance test-section syntax appear inside procedures and loops'' doesn't parse as English though
18:28 (join) mithos28
18:30 (quit) bro_grammer: Ping timeout: 245 seconds
18:30 neilv: offby1: thanks
18:31 Sgeo: Hmm. What happens if I want to document examples via Scribble but I don't want to actually run the examples because they do IO stuff?
18:32 neilv: Sgeo: you use forms like "racketblock" and "racketresult"
18:32 neilv: and "racketinput"
18:39 (quit) karswell: Remote host closed the connection
18:41 (join) jeapostrophe
18:43 (join) kofno
18:43 (quit) Fare: Ping timeout: 260 seconds
18:45 (quit) add^_: Quit: The Garbage Collector got me...
18:49 (join) karswell
18:51 (quit) anRch: Quit: anRch
18:55 asumu: mye: works for me if you take out the #:stateless thing
18:57 mye: asumu: ah, thank you for trying, I'm very irritated right now about the send/suspend stuff… :(
19:00 asumu: mye: well, I don't think send/formlet is supposed to be used with stateless things. It uses send/suspend from the stateful library, after all.
19:00 asumu: Though I could be wrong, I don't use the webserver stuff enough to know.
19:01 mye: yeah, I thought so too, I'll just rewrite it with the other send/suspend.
19:03 (quit) kofno: Remote host closed the connection
19:06 mye: Here is the error now that plagues me since yesterday https://gist.github.com/4416224
19:07 mye: No idea why it gets a request and where
19:07 mye: all I changed was add the send/formlet
19:07 mye: and reenable stateless
19:07 (join) nejucomo
19:08 mye: I must be doing something stupid but I don't see it!
19:29 mye: Indeed! The module must use #lang web-server instead of #lang racket
19:29 mye: weird but… so it is. I was initially using this as a submodule, the servlet that is
19:29 mye: and you can't use a #lang for a submodule
19:29 mye: right?
19:30 mye: it's just like require with submodules
19:31 (join) Fare
19:33 (quit) mithos28: Quit: mithos28
19:36 (quit) MayDaniel: Read error: Connection reset by peer
19:40 (quit) soegaard: Quit: soegaard
19:44 (quit) ThePawnBreak: Quit: Leaving
19:49 (join) kofno
19:54 (quit) kofno: Remote host closed the connection
19:55 (join) mithos28
19:58 (quit) mithos28: Client Quit
19:59 (join) kofno
20:01 (quit) kofno: Remote host closed the connection
20:02 (quit) nejucomo: Ping timeout: 276 seconds
20:02 Sgeo: In Racket, if you want to radically change the syntax, it gets its own language, but in Factor, you can have different syntax within a program using mostly normal syntax
20:02 Sgeo: :/
20:03 (join) kofno
20:04 asumu: mye: no, submodules can have a #lang... in a sense.
20:05 asumu: A (module foo bar ...) names a module `foo` using "language" `bar`
20:05 asumu: OTOH, module* and module+ use the surrounding module as the base, so you can't specify a language.
20:05 mye: asumu: but giving it web-server gave me an error
20:05 asumu: What error?
20:06 mye: will check swiftly
20:06 asumu: Sgeo: you can use #reader, but I'm not sure that's really a desirable feature in general...
20:07 mye: module: no #%module-begin binding in the module's language in: (module srv web-server (...(start req) (send/suspend ...
20:07 Sgeo: asumu, Factor uses it to embed EBNF syntax into code, for example
20:07 Sgeo: It also has XML literals due to that feature but I think XML literals suck
20:08 Sgeo: Also used for defining lexical scope when that's needed
20:08 (join) mye_
20:10 mye_: seems to work with /lang. Five letters to pull my hair out *grml*
20:11 asumu: Yeah, that's odd. Submodules came far after the web-server so there's probably some unanticipated weirdness there.
20:11 (quit) mye: Ping timeout: 245 seconds
20:11 (nick) mye_ -> mye
20:16 Sgeo: Can a loaded module be renamed such that re-declaring the module under the old name results in the module being in memory twice, once under the new name and the newer version under the old name?
20:18 jonrafkind: you could write a macro that defines the module with two different names
20:18 jonrafkind: but I guess the proper answer is to use units and instantiate them twice
20:19 Sgeo: I was hoping for the module to not necessarily be co-operating
20:20 jonrafkind: the only "normal" way is to instantiate the module in another phase, you get a unique instance for each phase
20:20 jonrafkind: what do you need it for?
20:21 Sgeo: jonrafkind, thinking of the possibility of module A needing version 1 of module X, and module B needing version 2 of module X, and module C needing both A and B
20:21 Sgeo: How can this be resolved/
20:21 Sgeo: ?
20:22 jonrafkind: but version 1 and 2 of X will be different files right
20:22 Cryovat: http://www.4clojure.com/problems
20:22 Cryovat: This was linked on Hacker News
20:22 jonrafkind: (require "X-1.rkt") (require "X-2.rkt")
20:22 Cryovat: That's pretty cool
20:23 Sgeo: jonrafkind, but what if A calls it X and B calls it X, would C have to change A and B?
20:23 Sgeo: And they're typically on the hard drive as X, but the user of C is willing to call it X-1 and X-2
20:23 Sgeo: Or does Planet do this sort of thing?
20:23 Sgeo: I'd like to see a Planet without the reliance on a central server
20:23 jonrafkind: so you want in A (require x) and in B (require x) ?
20:23 Sgeo: yes
20:24 jonrafkind: well someone is going to have to put the different versions of x in different files
20:24 jonrafkind: in planet you can do (require (planet 1 2 x)) or whatever to specify a version
20:24 Cryovat: Would be cool if there was some file where essentially
20:24 (quit) neilv: Ping timeout: 260 seconds
20:25 Cryovat: (define planet (make-repository "http://planet.racket-lang.org/index.something"))
20:25 Sgeo: I can almost guarantee that try-racket and 4racket would be more accurate to Racket than try-clojure and 4clojure are to Clojure
20:25 jonrafkind: I think in planet2 you have more control over the server
20:25 jonrafkind: i mean the repository
20:26 Sgeo: tryclj and 4clojure use Clojail for sandboxing. Clojail uses a not completely accurate recursive macroexpander.
20:27 Cryovat: Having a try-racket would be very cool
20:27 Sgeo: Yes
20:28 Cryovat: I've mentioned it before, but I don't have the knowhow to make it :(
20:32 (quit) bitonic: Ping timeout: 252 seconds
20:33 Sgeo: Probably cobble together code from tryhaskell with code from rudybot
20:35 (quit) cdidd: Ping timeout: 252 seconds
20:35 Sgeo: Actually, we'd want users to be able to try writing modules, right?
20:36 rudybot: Careful, don't poke. Ow.
20:36 Cryovat: Would that be necessary?
20:36 Cryovat: It'd be a text box on a web page
20:36 asumu: You can write modules at the REPL anyway.
20:37 asumu: rudybot: (module foo racket (define x 3) (provide x))
20:37 rudybot: asumu: Done.
20:37 asumu: rudybot: (require 'foo)
20:37 rudybot: asumu: Done.
20:37 asumu: rudybot: x
20:37 rudybot: asumu: ; Value: 3
20:37 asumu: Though yeah, a try-racket with #lang would be nicer in some ways.
20:37 Sgeo: Can't use #lang languages at a REPL
20:37 Cryovat: Where does that semicolon come from? :o
20:39 jonrafkind: you can use #lang's at the repl
20:39 jonrafkind: its some argument to racket
20:39 jonrafkind: racket -I blah, I think
20:40 (join) mizu_no_oto
20:40 Sgeo: jonrafkind, o.O
20:40 asumu: Random: is it just me or is there a "Could not connect to remote server" message in the background of http://blog.racket-lang.org/ ?
20:40 asumu: Well, you cna start the REPL in a language, but you can't change it.
20:40 asumu: *can
20:40 (quit) masm: Remote host closed the connection
20:41 asumu: I guess try-racket could just have a choice widget or something.
20:41 asumu: And work like rudybot
20:41 Cryovat: Where do you see it asumu?
20:41 Sgeo: Someone should make a language with freeform language extension
20:41 asumu: rudybot: init typed-racket
20:41 rudybot: asumu: error: default-load-handler: cannot open input file: "/mnt/racket-5.2.900.1/collects/typed-racket/main.rkt" (No such file or directory; errno=2)
20:41 asumu: rudybot: init typed/racket
20:41 rudybot: asumu: your typed/racket sandbox is ready
20:41 Sgeo: So, just use one language in the middle of antother
20:41 Sgeo: Reader and all
20:41 asumu: Cryovat: near the definition of `b1` in the top post
20:42 Cryovat: Can't see anything like that
20:42 Cryovat: Blogger having a bad moment?
20:42 asumu: Hmm, okay my browser must be weird. I don't see it in firefox either.
20:42 asumu: Oh, and it went away.
20:45 (join) mithos28
20:45 (quit) mithos28: Client Quit
20:48 (quit) spiderweb: Read error: Connection reset by peer
20:54 Cryovat: Heh
20:54 Cryovat: For some reason, I ended up on the Intro Projects wiki page, which sent me on to Rosetta Code
20:54 Cryovat: Racket is currently missing an entry on the Quine page
20:55 Cryovat: There is one for Scheme though
20:59 (quit) Nisstyre-laptop: Quit: Leaving
21:03 Sgeo: What happens if I want to redefine function application just within the body of one function that should see other definitions in the surrounding module and should be available to the surrounding module?
21:03 Sgeo: A macro that says "ok, function application is different here")
21:05 jonrafkind: hm, im not sure you can do that..
21:05 Sgeo: I'd like to be able to though
21:06 Cryovat: Would said function be called bind perchance? ;)
21:06 Sgeo: Would it be possible to make a macro that rebinds #%app within the body given to it?
21:09 Sgeo: Hmm, I think it might be possible
21:11 (quit) kofno: Remote host closed the connection
21:16 jonrafkind: oh right actually I think that is reasonable
21:17 jonrafkind: maybe
21:18 Sgeo: I don't entirely understand the label phase level
21:19 (join) sw2wolf
21:19 Sgeo: If it's used to bring stuff in for documentation, won't phase 1+ stuff still need to run to find all relevant definitions that would be available to a phase 0 program?
21:21 (join) kofno
21:22 (quit) kofno: Read error: Connection reset by peer
21:22 (join) nejucomo
21:26 Cryovat: Stupid question time: Whhy does both apply and #%app exist?
21:27 (join) mithos28
21:30 jonrafkind: #%app is a primitive form that is really no different from normal function application
21:30 jonrafkind: so (f 1 '(2 3)) is different from (apply f 1 '(2 3))
21:30 jonrafkind: the latter is the same as (f 1 2 3)
21:31 (join) RacketCommitBot
21:31 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://git.io/baVy6A
21:31 RacketCommitBot: racket/master 8b7379a Matthew Flatt: docs: clarification for `immutable?'
21:31 RacketCommitBot: racket/master 7cb7bea Matthew Flatt: fix non-JIT build...
21:31 (part) RacketCommitBot
21:31 jonrafkind: #%app is kind of only relevant for the expander. its part of the core AST that is sent to the compiler
21:31 (quit) jschuster: Ping timeout: 252 seconds
21:31 Cryovat: So every normal function application has an #%app in front of it?
21:32 Cryovat: +implicit
21:32 jonrafkind: yea
21:32 jonrafkind: the expander rewrites (f a) as (#%app f a)
21:32 Cryovat: Could apply have been used there, or do they have different semantics?
21:33 jonrafkind: apply is a normal function
21:34 jonrafkind: (apply f a) is rewritten as (#%app apply f a)
21:34 sw2wolf: then why the compiler needs #%app ?
21:34 jonrafkind: its how it recognizes function applications
21:35 Cryovat: Ah, tried entering #%app and apply into the REPL
21:35 jonrafkind: you know when you write a normal parser you parse f(a) into something like (Invocation f (Args a))
21:35 jonrafkind: that (Invocation) AST thing is what racket calls #%app
21:35 Sgeo: Cryovat, it's so that things can change what function application does
21:36 Sgeo: (+ 1 2 3) turns into (#%app + 1 2 3). So if I change the meaning of #%app, I change what (+ 1 2 3) means
21:36 Cryovat: Ah, that makes sense
21:37 Cryovat: Thanks :)
21:37 Sgeo: You're welcome
21:38 sw2wolf: user program donot need to use #%app ?
21:39 jonrafkind: no
21:39 Cryovat: sw2wolf: It seems very much like one of those things you should never touch unless you have a very good reason for it
21:39 (join) jschuster
21:39 Cryovat: And if you do, you're probably doing things too complicated ;)
21:39 jonrafkind: Sgeo, actually I know why rebinding #%app might not work, because the app that you bind in the macro will not be used by the expander
21:39 sw2wolf: Cryovat: i see
21:41 Sgeo: Cryovat, you know the lazy language? It's easy to imagine it changing #%app for its purposes
21:41 Cryovat: That's true
21:42 Cryovat: I have a feint hope that one day, there might be a marriage of Lazy Racket and Typed Racket ;)
21:42 (join) kofno
21:42 Sgeo: That's sort of one of my criticisms of Racket currently, it doesn't seem obvious how to combine two languages
21:43 Sgeo: Cryovat, Haskell fan?
21:43 asumu: Sgeo: label phase import happens on compiled programs, so phase 1+ computations should already have run.
21:43 Cryovat: Kind of
21:43 Cryovat: I've been trying to get into Haskell for a while
21:43 Cryovat: And keep on having "oh, awesome!" moments reading about it
21:43 asumu: AFAIK, anyway. I don't know the implementation details of label-phase.
21:44 Cryovat: But it still makes me feel incredibly stupid :\
21:49 sw2wolf: haskell is a "theory language" which is not VERY practical
21:50 Sgeo: The most impractical things about Haskell to me are difficulty of macros and difficulty of changing running code
21:50 Sgeo: It occurs to me that both do kind of apply to Racket
21:50 Cryovat: I think it can be practical language once you get into it
21:50 Cryovat: It's just that it forces so much theory on you
21:50 sw2wolf: You can learn its concepts but hard to use i real programming
21:51 (quit) mithos28: Quit: mithos28
21:51 Sgeo: Its form of I/O can be really convenient sometimes
21:55 Sgeo: Why is this disallowed? (define (foo) (define bar 5))
21:55 Sgeo: rudybot, eval (define (foo) (define bar 5))
21:55 rudybot: Sgeo: error: #<undefined>:1:0: begin (possibly implicit): no expression after a sequence of internal definitions in: ((define bar 5))
21:56 offby1: rudybot: (define (foo) (define bar 5) bar)
21:56 rudybot: *offby1: your sandbox is ready
21:56 rudybot: *offby1: Done.
21:56 offby1: rudybot: (foo)
21:56 rudybot: *offby1: ; Value: 5
21:57 Sgeo: offby1, I know, but what's so bad about a useless define in a useless definition?
21:57 Sgeo: It might be bad to write by hand, but macros can easily produce "bad" (ugly) code in some circumstances, and should macros have to check for that?
21:58 Cryovat: I'd say yes
21:58 (join) mithos28
21:58 Cryovat: Macros are magical
21:58 Sgeo: I'd say no.
21:58 Cryovat: So they should fail as quickly as possible
21:58 (join) spiderweb
21:59 Sgeo: Why should a macro need to be burdened by checks to make sure it's not producing "useless" code that still actually makes sense?
21:59 asumu: It's not that it's useless. It's just that definitions are not expressions and there cannot return a result.
21:59 asumu: *therefore cannot
22:00 asumu: In an imaginary world where definitions evaluated to void and were expressions, it would be legal code.
22:00 Cryovat: Why don't they?
22:00 Cryovat: Scheme legacy?
22:01 Cryovat: To me, the most obvious thing would be for a definition to evaluate to its value
22:01 asumu: It makes scoping weird. What's the scope of the binding of x in (f (define x 3) 5)?
22:01 Cryovat: Ah, that's true
22:01 asumu: (assuming 'f' is bound to a function and not a macro)
22:04 asumu: I guess in languages with implicitly declared variables, the answer is that there's a default scope.
22:05 asumu 's personal opinion is that implicit variables are more trouble than they're worth
22:06 (join) serhart
22:06 Cryovat: Agreed
22:06 Cryovat: It's a nightmare in Javascript
22:07 mye: for me, programming is a nightmare. Good thing I'm a masochist :-P
22:08 sw2wolf: Can plt-web-server be used to production ?
22:09 sw2wolf: Are there any VPS which supports using plt-web-server ?
22:09 asumu: sw2wolf: sure, and there are people that do. Like http://iwl.me/
22:09 sw2wolf: great !
22:10 Cryovat: Is racket-lang.org "self-hosted"?
22:10 Sgeo wonders if Heroku could be made to use Racket
22:10 sw2wolf: me too
22:10 asumu: (also apparenlty I write like Arthur Clarke)
22:11 asumu: https://github.com/onixie/heroku-buildpack-racket
22:11 Cryovat: Are you a fan? :D
22:11 (join) RacketCommitBot
22:11 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/-pHXIg
22:11 RacketCommitBot: racket/master 9b62726 Robby Findler: tweak the latex rendering of ╔ and company
22:11 (part) RacketCommitBot
22:12 asumu: Yeah, I enjoyed his _Childhood's End_ especially.
22:12 Cryovat: I haven't read that :/
22:13 Cryovat: Hmm
22:13 Cryovat: There is a Kindle version
22:13 asumu also liked the 2001 series, of course. :p
22:13 Cryovat: I haven't read too much classical sci fi
22:14 Cryovat: I started on one of the Foundation books by Asimov, but found it... weird
22:14 Sgeo: Apparently I write like Isaac Asimov
22:14 Sgeo: Based on one of my Tumblr posts
22:14 (quit) mizu_no_oto: Ping timeout: 255 seconds
22:15 (join) RacketCommitBot
22:15 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://git.io/3Bdvjg
22:15 RacketCommitBot: racket/master 0d01db8 Sam Tobin-Hochstadt: Fix type of `log-message`.
22:15 (part) RacketCommitBot
22:15 Sgeo: What happens if someone is using the nick RacketCommitBot when it comes in?
22:17 jonrafkind: it probably just alters its name, appending a _ or 1 to it
22:17 sw2wolf: What is the difference between "(define (out format-string . args) ...)" and "(define (out format-string args) ... )" ?
22:17 mye: sw2wolf: args is a list with the arguments
22:17 (join) mizu_no_oto
22:17 jonrafkind: common lisp doesn't have rest args?
22:18 sw2wolf: CL has &rest
22:18 jonrafkind: oh, so its like that i guess
22:20 sw2wolf: mye: thx
22:21 sw2wolf: jonrafkind: it is simple than CL's &rest
22:21 sw2wolf: at least type little
22:23 (join) Kaylin
22:23 Sgeo: sw2wolf, you may or may not be pleased to know that in Racket, rest parameters and keyword parameters won't get in eachother's way
22:23 sw2wolf: oh
22:24 Sgeo: um
22:24 Sgeo: Looking at the guide, actually, accepting both rest and keyword arguments is ugly
22:24 sw2wolf: i would like to hear
22:25 Sgeo: n/m not ugly
22:25 Cryovat: It sounds a bit icky :P
22:26 Sgeo: rudybot, eval (define (foo #:a a #:b b . rest) (list b a rest)) (foo #:a 5 #:b 6 1 2 3)
22:26 rudybot: Sgeo: ; Value: (6 5 (1 2 3))
22:26 Sgeo: The trick is that #:a is not actually a value, unlike in CL
22:26 Sgeo: Keyword arguments don't work by looking at the rest of the arguments
22:27 Sgeo: rudybot, eval (define (foo #:a a #:b b . rest) (list b a rest)) (foo 1 2 3 #:a 5 #:b 6)
22:27 rudybot: Sgeo: ; Value: (6 5 (1 2 3))
22:28 Sgeo: rudybot, eval #:not-actually-a-value
22:28 rudybot: Sgeo: error: #<undefined>:1:0: #%datum: keyword used as an expression in: #:not-actually-a-value
22:28 Sgeo: rudybot, eval #:not-actually-a-value 5
22:28 rudybot: Sgeo: error: #<undefined>:1:0: #%datum: keyword used as an expression in: #:not-actually-a-value
22:29 sw2wolf: a bit tricky ?
22:29 Sgeo: #%datum is another hook that's pervasive in fully expanded syntax, similar to #%app
22:29 Sgeo: In your own language, you could redefine #%datum to allow keywords to be arguments
22:30 Sgeo: I guess that #%app does some special thing when it sees keywords, but don't know
22:30 asumu: rudybot: (keyword? '#:foo)
22:30 rudybot: asumu: your typed/racket sandbox is ready
22:30 rudybot: asumu: ; Value: #t
22:30 rudybot: asumu: ; stdout: "- : Boolean\n"
22:30 asumu: They have to be quoted to be values.
22:30 Sgeo: rudybot, eval (define (foo #:a a #:b b . rest) (list b a rest)) (foo 1 2 3 #:a 5 '#:b 6)
22:30 rudybot: Sgeo: error: foo: requires an argument with keyword #:b, not supplied; arguments were: 1 2 3 '#:b 6 #:a 5
22:31 asumu: (you can apply with quoted kws via `keyword-apply`... but usually you don't want/need to)
22:32 Sgeo: The need for a separate keyword-apply is the thing that bothers me about the way Racket handles it
22:33 asumu: `apply` also takes keywords, actually
22:33 asumu: rudybot: (apply (lambda (#:x [x 5]) x) #:x 7)
22:33 rudybot: asumu: error: #<undefined>:1:16: lambda: not an identifier at: #:x in: (lambda (#:x (x 5)) x)
22:34 asumu: rudybot: (apply (lambda (a #:x [x 5]) x) #:x 7)
22:34 rudybot: asumu: error: #<undefined>:1:18: lambda: not an identifier at: #:x in: (lambda (a #:x (x 5)) x)
22:34 Sgeo: asumu, but when you're writing a function that's just going to take the same arguments as some other function, do stuff, then call the other function with those arguments, you need to deal with keywords with keyword-apply and that other thing
22:35 Sgeo: A simple wrapper with a rest argument and apply doesn't cut it, the way it would with CL or Clojure
22:35 asumu: rudybot: init racket
22:35 rudybot: asumu: your sandbox is ready
22:35 asumu: rudybot: (apply (lambda (a #:x [x 5]) x) '(5) #:x 7)
22:35 rudybot: asumu: ; Value: 7
22:36 asumu: Sgeo: yes, that's true.
22:37 (nick) sw2wolf -> sw2wolf{away}
22:42 Fare: keyword-apply ?
22:42 Fare: rudybot: keyword-apply
22:42 rudybot: Fare: Hmm.  (keyword-apply p (append k0 k1) (append a0 a1) (append p0 p10)) doesn't work; you need to merge the sorted lists, or KEYWORD-APPLY needs to sort itself.
22:43 Fare: rudybot: doc, keyword-apply
22:43 rudybot: Fare: The need for a separate keyword-apply is the thing that bothers me about the way Racket handles it
22:44 asumu: rudybot: doc keyword-apply
22:44 rudybot: asumu: http://docs.racket-lang.org/reference/procedures.html#(def._((lib._racket%2Fprivate%2Fbase..rkt)._keyword-apply))
22:44 asumu: Fare: ^
22:45 Sgeo: Why is rudybot quoting me?
22:46 Fare: asumu: thanks
22:47 (quit) serhart: *.net *.split
22:47 (quit) lusory: *.net *.split
22:47 (quit) offby1: *.net *.split
22:47 (join) offby1
22:47 Fare: I don't love keyword-apply but I don't love the alternatives either.
22:47 (quit) nejucomo: Ping timeout: 276 seconds
22:47 (quit) offby1: Changing host
22:47 (join) offby1
22:48 Fare: whichever way you do things, when you make function calls fancy, that fanciness will lead to complexity in "reflective" features that explicitly handle it.
22:51 mye: I read in a post by Rob Pike golang doesn't support optional args bc. it leads to bad interface design. Now I have a little devil in my head that makes me feel bad when I use fancy interfaces and I don't even know why  :x
22:51 Fare: I might have done it by reifying argument calls as finite association maps from argument positional index or keyword to value, or something.
22:52 (join) nejucomo
22:53 Fare: with lists and sequences being objects satisfying the "finite association map" interface.
22:55 (quit) jeapostrophe: Ping timeout: 265 seconds
22:56 (join) jeapostrophe
22:56 (quit) jeapostrophe: Changing host
22:56 (join) jeapostrophe
22:58 (quit) mithos28: Quit: mithos28
22:59 (join) mithos28
23:04 (quit) jeapostrophe: Ping timeout: 255 seconds
23:10 (join) serhart
23:11 (quit) nejucomo: Remote host closed the connection
23:12 (join) nejucomo
23:12 Sgeo: There's a lot of really cool stuff that's in the reference but not in the guide
23:19 (join) Shvillr_
23:19 (quit) Shviller: Disconnected by services
23:19 (nick) Shvillr_ -> Shviller
23:21 mithos28: Is there a reason that there is a box-cas! but not the equivalent for vector?
23:21 (quit) Shvillr: Ping timeout: 260 seconds
23:22 (join) Shvillr
23:28 (quit) spiderweb: Quit: ERC Version 5.3 (IRC client for Emacs)
23:28 (quit) francisl: Quit: francisl
23:30 (join) spiderweb
23:46 Sgeo reads dyoo's brainfuck tutorial
23:46 Sgeo: Erm, I guess it's not really about Brainfuck
23:49 Sgeo: Actually, I think dyoo consistently censors the name
23:52 (quit) mizu_no_oto: Quit: Computer has gone to sleep.
23:58 (quit) kofno: Read error: Connection reset by peer