00:02 (quit) bugQ: Ping timeout: 240 seconds
00:18 (join) dnolen
00:20 (quit) dnolen: Client Quit
00:46 (join) mithos28
02:03 (nick) elliottcable -> ec|detached
02:14 (quit) realitygrill: Quit: realitygrill
02:25 (join) sir_lewk
02:26 sir_lewk: is there a proper way to do basically a 'popen'? ("popen, pclose - pipe stream to or from a process")
02:31 (join) littlebobby
02:40 (quit) mithos28: Quit: mithos28
02:57 (quit) littlebobby: Quit: Ex-Chat
03:31 (join) igor_
03:45 (part) igor_
03:55 (join) masm
06:53 eli: sir_lewk: ?
07:03 (join) samth_
07:26 (join) reynir
08:41 (join) racket
08:41 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/lF4wOf
08:41 racket: [racket/master] The format change script should not be run by drdr. - Vincent St-Amour
08:41 (part) racket
08:52 (join) lucian
08:56 (join) MayDaniel
09:12 (join) realitygrill
09:23 (join) racket
09:23 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/mPwwCf
09:23 racket: [racket/master] PPC repairs - Matthew Flatt
09:23 (part) racket
09:33 (join) mithos28
09:49 (quit) mithos28: Quit: mithos28
10:02 (quit) samth_: Ping timeout: 255 seconds
10:05 (quit) MayDaniel: Read error: Connection reset by peer
10:19 (join) mithos28
10:52 (join) anRch
11:28 (join) bugQ
11:35 (quit) bugQ: Ping timeout: 240 seconds
11:39 (join) littlebobby
11:56 (quit) gienah: Quit: leaving
12:02 (join) dnolen
12:03 (quit) dnolen: Client Quit
12:11 sir_lewk: eli: basically I'm wondering about how IPC is generally done
12:14 (quit) mithos28: Quit: mithos28
12:17 offby1: rudybot: doc subprocess
12:17 rudybot: *offby1: http://docs.plt-scheme.org/reference/subprocess.html#(def._((quote._~23~25kernel)._subprocess))
12:17 offby1: sir_lewk: see that ^^
12:25 (part) reynir
12:39 (quit) anRch: Quit: anRch
12:53 sir_lewk: excellent, thanks
12:57 (join) lucian_
12:58 (quit) lucian: Read error: Operation timed out
13:02 (join) gwdean
13:06 (quit) gwdean: Ping timeout: 252 seconds
13:09 (join) mithos28
13:18 (quit) littlebobby: Ping timeout: 240 seconds
13:27 (join) kenjin2201
13:27 (join) jamiltron
13:29 (join) lucian
13:31 (quit) lucian_: Ping timeout: 258 seconds
13:32 (join) kenjin2202
13:36 (part) shofetim: "ERC Version 5.3 (IRC client for Emacs)"
13:36 jamiltron: Hello, I was wondering if someone could help me with an error I am getting in a program that was working a week ago (and seems to still work if I use gambit scheme) but is giving an unusual error today.
13:37 offby1: if it's super-tiny, sure
13:38 jamiltron: The error I am getting is "compile: unbound identifier (and no #%top syntax transformer is bound)"
13:38 offby1: *shrug*
13:38 (join) bugQ
13:38 offby1: is the whole thing small enough to paste to (e.g.) gist.github.com ?
13:38 (quit) mithos28: Quit: mithos28
13:38 jamiltron: Not really
13:39 jamiltron: I have a github repo of it, but its a little long
13:39 offby1: well, I don't know what that error means, off the top of my head.
13:39 bugQ: bremner: no but one would expect type-case to work like cond
13:40 offby1: I can certainly guess what "unbound identifier" means, but then so can you.
13:40 bugQ: (18 hours late)
13:40 jamiltron: I mean its one 330 line program that probably could be cleaned up: https://github.com/jamiltron/forsight/blob/master/forsight.ss
13:40 offby1: jamiltron: why don't you shrink the program as much as possible, while still reproducing the error?  That will make it easier for lazy strangers such as myself to help.
13:40 jamiltron: It works in gambit scheme, which makes me think it may be something weird with racket - like how I installed the sicp module.
13:40 jamiltron: Ok let me give it a try
13:41 offby1: jamiltron: I just tried it, and I don't get any error; instead I get what looks like a prompt.
13:41 jamiltron: type in something simple like 2 3 +
13:41 offby1: I saved it to a file named 'forsight', then typed 'racket forsight'
13:41 offby1: ah
13:41 jamiltron: and hit return
13:41 jamiltron: sorry
13:42 offby1: gotcha
13:43 jamiltron: The main error is throwing down in "interpret-f" in the else-clause of ((symbol-f fun-f) ...
13:44 jamiltron: Which is weird, if I comment out the repl stuff and go in and do the fun-f lookup for say "+" it retrieves the symbol for the add function.
13:44 jamiltron: Then if I call the eval on it like is being done in the next line it resolves appropriately
13:45 offby1: my hunch is that it has something to do with your calling "eval"
13:45 offby1: it's not obvious to me that you should need to do that
13:45 jamiltron: Hmmm, you might be right, let me try something.
13:46 offby1: yeah, you're calling (eval 'add-f)
13:46 jamiltron: That gets the procedure add-f, doesn't it?
13:46 offby1: I can't remember the circumstances in which "eval" is actually the right thing, but they're rare.
13:47 offby1: jamiltron: obviously, (eval 'add-f) gets an error :)
13:47 offby1: as you can see
13:47 offby1: now, I can't explain why it worked before.
13:47 jamiltron: If I run (eval 'add-f) in the interpreter (not in the repl) it doesn't error.
13:47 offby1: that doesn't surprise me.
13:47 offby1: even though I'm not certain what you mean by "interpreter" and "repl"
13:48 jamiltron: Oh yeah sorry name overloading :P
13:48 offby1: I predict that if you quit using "eval", your problem will go away.
13:49 jamiltron: No that doesn't seem to do it
13:49 jamiltron: because then it tries to resolve ('add-f ...)
13:49 jamiltron: and add-f is a symbol
13:49 offby1: well, like I say: reduce the problem to a (say) five-line program that reproduces the error; it'll be much easier to chew on
13:49 offby1: I didn't mean to simply delete the token 'eval' from your program.
13:49 offby1: I meant that you should write a replacement.
13:49 jamiltron: And it was working with eval just a week ago
13:50 offby1: That shouldn't be hard, since all you're using it for (at first glance) is consants and your own procedures, the latter of which you're already storing in a dict, indexed by symbol
13:50 offby1: *constatnt
13:50 offby1: GAAH
13:50 offby1: *constants
13:56 (quit) masm: Ping timeout: 240 seconds
13:58 (join) mithos28
14:01 jamiltron: Hmmm, if I keep the whole forth-repl thing and the tokenizing and how functions are defined I can pair it down to about 55 lines.
14:02 jamiltron: https://gist.github.com/1033340
14:03 bugQ: *pare
14:03 bugQ: >_>
14:03 jamiltron: Sure
14:03 jamiltron: that too
14:05 bremner: bugQ: I guess so. I was less familiar with scheme when I started PLAI, so the issue never occured to me.
14:08 bugQ: well it could be worse
14:09 bugQ: it could have required a monad instead :P
14:11 offby1: jamiltron: you might want to read http://docs.racket-lang.org/guide/eval.html#(part._namespaces)
14:12 jamiltron: Right, I've gone through that before but I'll give it another look.
14:13 jamiltron: As far as I know it should still work (in fact, it does when ran outside of my repl-f loop)
14:13 jamiltron: Because the dict-get function is retrieving the quoted form of add-f
14:13 jamiltron: so when I evaluate it I get the procedure it represents.
14:16 offby1: jamiltron: try these tweaks: http://ix.io/1Kf
14:19 jamiltron: Hmmm: forsight.ss:272:48: compile: unbound identifier in module in: ns
14:21 jamiltron: Racket doesn't seem to be using eval how it says it does any longer.
14:22 jamiltron: Using the example from the section on namespaces in the docs throws the same error I am getting
14:22 jamiltron: Or maybe I am incorrect
14:22 jamiltron: Yeah nevermind on that
14:22 jamiltron: I screwed something up
14:24 offby1: you must have failed to include the first two lines of my patch -- the second of them defines "/me nods sagely"
14:24 offby1: heh
14:24 offby1: "ns"
14:25 jamiltron: Yeah sorry I copied in the wrong line there
14:26 offby1: anyway, when I ran my fixed version, it no longer raised that error; instead it printed <ok>.  Dunno if that's OK or not :)  I'd have sorta expected it to print 5
14:26 (join) samth_
14:26 jamiltron: It will is you type in .s
14:26 jamiltron: or .
14:26 jamiltron: which it does
14:27 offby1: ah
14:27 jamiltron: Forth is stack-based so when you call add it pops off the two top-most operators on pushes the result back onto the data-stack
14:27 jamiltron: I really appreciate the help.
14:27 jamiltron: But I guess I don't really understand the neccessity of this change, or why it was working just a week ago.
14:28 offby1: ok, I pronounce it healthy
14:28 jamiltron: And why it works in another scheme implementation.
14:28 jamiltron: Thanks
14:28 offby1: I can't explain why it used to work in racket.
14:28 offby1: But it's easy to explain why it works in another scheme: that other scheme does things differently :)
14:28 jamiltron: Right
14:29 jamiltron: It was weird because I originally did it in gambit, which was the scheme I originally learned. Then I got tired of doing stuff like rolling my own dictionary and string functions, so I decided to start updating it to Racket
14:29 jamiltron: And it worked fine for a little while. I went on vaction, came back, tried running it before I futher updated it and I started getting that error.
14:30 offby1: clearly, while you were on vacation, brigands broke into your house and redefined racket
14:30 jamiltron: That's usually what happens to all my code that doesn't work
14:30 jamiltron: Or that's what I tell my boss.
14:32 jamiltron: Do you know why it needs that namespace anchor? Why shouldn't ((eval 'add-f ) '(2 3)) work?
14:33 jamiltron: I mean, I can read the docs more, but I guess I still don't see the issue. Granted I'm a little dense but I still thought I had the whole symbol vs. procedure thing down.
14:35 bugQ: wouldn't apply be nicer ?
14:37 jamiltron: I'm not applying a function to a list, I'm taking a symbol for a procedure and calling it on a list. The procedure then determines what elements from the list it needs to operate on, although its usually the first one, or the first two.
14:39 bugQ: ok, but that's still just (apply fun-f (list d-stack))
14:40 bugQ: I realize that may not be aesthetically appealing
14:41 jamiltron: fun-f is a symbol, so wouldn't it need to be eval-d to get the procedure it represents?
14:41 bugQ: wait ignore me
14:41 bugQ: yeah
14:43 jamiltron: Thanks, I'm just really confused by the need for a namespace anchor.
14:44 offby1: jamiltron: I'd put the procedure fun-f into a dictionary, with 'fun-f as the key.
14:45 jamiltron: So like cons fun-f onto the data stack and then evaluate it?
14:45 bugQ: danger will robinson
14:51 bugQ: I think what offby1 means is to put the procedures themselves in the dictionary instead of their names
14:51 bugQ: but keyed by their names
14:51 offby1 nods sagely
14:52 bugQ: thus eliminating eval
14:53 jamiltron: Hmm, so the basic idea I am going for is that the dictionary holds a key value pair for what user input relates to what function (so in this case - "+" relates to add-f), and then the value of that is the name of the procedure to be executed.
14:54 bugQ: not quite
14:54 bugQ: functions are values too
14:54 bugQ: just put add-f itself in there
14:54 jamiltron: Hmm, isn't there a better way to enscapulate that, or am I missing something?
14:55 jamiltron: So you're saying in my dictionary correspond "+" to be the whole procedure of add-f (the lambda-on-down part of add-f's definition)
14:56 bugQ: when you type add-f it _refers_ to the procedure
14:57 bugQ: so if you have an entry that looks like "+" -> add_f, it will point to the procedure, neither its name nor a simile of it
14:57 bugQ: so whatever name you bind it to later will work just like add_f
14:57 bugQ: sorry add-f
14:58 bugQ: e.g. (fun-f d-stack)
14:58 jamiltron: I do have any entry that is ("+" add-f)
14:59 bugQ: then that's the problem, you shouldn't even be evaluating it
14:59 offby1: jamiltron: this should give you the idea of what I'm talking about: http://ix.io/1Kg
14:59 bugQ: it's just like how functions work in C
14:59 jamiltron: Right, but when I get the value from the dictionary using the key it gives it back to me as a symbol.
15:00 bugQ: oh right, because you constructed as a quoted list
15:02 jamiltron: Wait... so if I use list instead of quoting it should bind the procedure instead of the name - right?
15:03 jamiltron: the list keyword in place of the quote, that is
15:03 bugQ: indeed
15:03 jamiltron: And reorganize my code so that the forth procedures come first :P
15:04 ohwow: Hm, I want to make something like a GUI designer
15:04 ohwow: like Glade or smth
15:04 ohwow: what's the good place to start?
15:04 ohwow: I know some racket, but I am clueless about this specific task
15:10 (join) littlebobby
15:12 (join) ShereKahn
15:13 (quit) ShereKahn: Client Quit
15:13 (quit) mithos28: Quit: mithos28
15:20 (quit) kenjin2202: Read error: Connection reset by peer
15:20 (quit) kenjin2201: Read error: Connection reset by peer
15:25 (quit) jamiltron: Quit: jamiltron
15:28 bremner: maybe http://docs.racket-lang.org/guide/More_Libraries.html?q=gui#(part._graphics)
15:33 (join) masm
15:42 bugQ: what are the appropriate uses of assignment transformers ?
15:42 bugQ: and are they necessary ?
15:46 (join) mithos28
15:47 (quit) masm: Ping timeout: 276 seconds
15:48 bugQ is just a little annoyed that set! can't return anything
15:48 (join) masm
15:49 bremner: write a function?
15:51 (quit) mithos28: Client Quit
15:52 rapacity: could always write a macro set!* maybe that returns a value
15:54 rapacity: rudybot: (define-syntax-rule (set!* id val) (let ([tmp val]) (set! id tmp) tmp))
15:54 rudybot: rapacity: your sandbox is ready
15:54 rudybot: rapacity: Done.
15:54 rapacity: rudybot: (define x 10) (set!* x 90)
15:54 rudybot: rapacity:   (define (add-complex z1 z2)
15:55 (join) mithos28
15:56 bugQ: that one of course assumes that accessing the var gives the same result without side effects
15:58 bugQ: it's just sometimes nice to know I won't launch the missle because I set the sprinkler timer for Tuesday
16:03 (quit) littlebobby: Ping timeout: 255 seconds
16:07 (quit) masm: Ping timeout: 276 seconds
16:09 (join) masm
16:13 (quit) masm: Ping timeout: 255 seconds
16:14 (quit) mithos28: Quit: mithos28
16:25 (join) littlebobby
16:34 (quit) littlebobby: Ping timeout: 264 seconds
16:36 (quit) bugQ: Ping timeout: 244 seconds
16:38 (join) littlebobby
16:40 (join) mithos28
16:44 (quit) samth_: Ping timeout: 255 seconds
16:47 (join) masm
17:04 (quit) masm: Ping timeout: 252 seconds
17:09 (join) masm
17:27 (join) realitygrill_
17:29 (quit) realitygrill: Ping timeout: 240 seconds
17:29 (nick) realitygrill_ -> realitygrill
17:35 (nick) samth_away -> samth
17:41 (quit) mithos28: Quit: mithos28
17:43 samth: ohwow: there's a gui designer on planet
17:51 offby1: rapacity: I think you need to wrap those two forms in a (begin), otherwise rudybot doesn't realize it's scheme code
17:52 offby1: rudybot: (begin (define x 10) (set!* x 90))
17:52 rudybot: *offby1: error: reference to an identifier before its definition: set!* in module: 'program
17:52 offby1: well, you know.
17:52 offby1: rudybot: (define-syntax-rule (set!* id val) (let ([tmp val]) (set! id tmp) tmp))
17:52 rudybot: *offby1: Done.
17:52 offby1: rudybot: (begin (define x 10) (set!* x 90))
17:52 rudybot: *offby1: ; Value: 90
18:02 offby1: What problem do impersonators solve?
18:02 offby1: rudybot: x
18:02 rudybot: *offby1: ; Value: 90
18:07 (join) mithos28
18:09 (quit) mithos28: Client Quit
18:13 (join) mceier
18:25 samth: offby1: do you mean, as opposed to chaperones?
18:25 samth: or do you mean, what problem do chaperones and impersonators solve?
18:28 offby1: the latter.
18:28 offby1: both are new to me.
18:29 offby1: My number 1 complaint about documentation is always: it doesn't tell me the problem which the documented technology solves.  Lacking that, I can never make any sense out of the documentation, no matter how thorough it is.
18:29 offby1: So I read the docs without really understanding, and I say "that sounds vaguely interesting.  I cannot imagine how it could possibly be useful", and promptly forget it.
18:34 (quit) lucian: Remote host closed the connection
19:09 samth: offby1: the answer is "how do you implement contracts"
19:09 samth: or more specifically, "how do you implement contracts for mutable data"
19:09 offby1: hm.
19:10 offby1: I confess that that doesn't enlighten me, but that's probably because I haven't thought about it
19:10 offby1: I do wish more documentation had _motivating_ examples.
19:10 samth: mostly we write those in papers
19:11 offby1: that's not useful for the guy who's just reading the docs.
19:11 samth: which isn't a good answer, but it's true
19:11 samth: and the chaperones paper hasn't appeared yet
19:11 offby1: it might suffice to simply include a link to the paper (once it appears, of course)
19:11 samth: that happens in some places in the docs
19:11 samth: but not enough
19:12 samth: if you give me a couple minutes, i'll explain the rationale at more length
19:12 samth: but the short piece is -- think about creating a contract on a mutable vector.  how should it behave?
19:12 (join) mithos28
19:12 offby1: why not just write the rationale down and include it in the docs?  More effective than explaining it to me directly
19:13 offby1: given that I've already decided that it's beer-thirty here
19:13 samth: i wouldn't write it well enough for that
19:13 offby1: don't let the perfect be the enemy of the good!
19:13 offby1: for example: I just uploaded a rather awful PLaneT package.
19:13 offby1: It's surely better than nothing, but just as surely not very good.
19:14 samth: the quality of the documentation is more important to preserve, i think
19:14 offby1: but it can always be improved, yes?
19:15 offby1: so upload something less-than-ideal-but-better-than-nothing now, and improve it later.
19:15 offby1: *shrug*
19:16 samth: that's true, but currently the documentation is written in a uniform style, with uniform high quality
19:16 samth: whereas planet, no so much :)
19:16 offby1: heh
19:16 offby1: maybe there needs to be a canonical place for "not really polished but perhaps helpful" documentation.
19:17 samth: i'd like to use a wiki for that
19:17 samth: which is why i put that wiki page up
19:17 samth: on which you should add a link to your planet package, btw
19:17 offby1: oho!
19:17 offby1: gimme the link again
19:17 offby1: it was on github, yes?
19:19 samth: yeah, just the github wiki
19:20 samth: https://github.com/plt/racket/wiki/HTTP-over-SSL
19:22 offby1: ok, just added a ref to it
19:25 samth: cool
19:50 (quit) mithos28: Quit: mithos28
20:04 (join) mithos28
20:05 (quit) masm: Ping timeout: 240 seconds
20:18 eli: sir_lewk: If you use `process' etc, they read from the usual `current-input-port', and write to `current-output-port' and `current-error-port', so the plumbing work is done by the library.
20:19 eli: samth: Using that wiki is a bad idea.
20:20 samth: eli: provide me a better one, and i'll move all my content
20:20 eli: samth: I knew you'd flame about it, so I'd leave it at that.  If it becomes a problem I'll make it read-only or something.
20:40 (join) gienah
20:41 (quit) mithos28: Quit: mithos28
21:02 (quit) mceier: Quit: leaving
21:02 (nick) samth -> samth_away
21:22 (join) neilv
21:40 neilv: does everyone like the way drracket does paren-matching with the gray background?
21:42 littlebobby: It can be a little bit distracting, but I haven't been using drracket for too long. maybe that's the reason. just started reading littleschemer (after my initial interest in sicp broke off a little)
21:49 (join) samth_
22:07 (join) mithos28
22:14 (join) jonrafkind
22:22 (quit) mithos28: Quit: mithos28
22:45 (join) ckrailo
22:45 (quit) ckrailo: Client Quit
23:03 (join) realitygrill_
23:04 (quit) realitygrill: Ping timeout: 240 seconds
23:04 (nick) realitygrill_ -> realitygrill
23:07 (quit) neilv: Ping timeout: 250 seconds
23:35 (quit) samth_: Ping timeout: 255 seconds