00:41 (quit) danbrown: Remote host closed the connection
00:45 (join) bahrieinn
00:45 bahrieinn: hey can anyone here help with racket install on linux?
00:51 (quit) bahrieinn: Ping timeout: 265 seconds
01:12 (join) jonrafkind
03:00 (quit) jonrafkind: Ping timeout: 240 seconds
03:41 (quit) emma: Ping timeout: 252 seconds
04:29 (quit) Fare: Quit: Leaving
05:33 (join) masm
06:20 (join) hanDerPeder
06:39 (quit) _p4bl0: Read error: Operation timed out
07:29 (nick) Lajla -> binrapt
07:29 (nick) binrapt -> b-ratius
07:30 (nick) b-ratius -> b-ratjus
07:30 (nick) b-ratjus -> bjnrapt
07:48 (join) b-man_
08:00 (quit) b-man_: Ping timeout: 252 seconds
08:09 (join) b-man_
08:16 (quit) b-man_: Ping timeout: 240 seconds
08:51 (quit) jao: Ping timeout: 240 seconds
09:10 (join) MayDaniel
09:14 (quit) MayDaniel: Client Quit
09:56 (join) danbrown
09:59 (join) pentra
10:12 (quit) jeapostrophe: *.net *.split
10:18 (quit) danbrown: Remote host closed the connection
10:33 (join) jeapostrophe
10:39 (join) jonrafkind
10:45 (quit) pentra: Quit: Page closed
11:29 clklein: samth: What's the bug?
11:30 samth: i have this pattern: github.com/offby1/rudybot
11:30 samth: sorry, wrong paste
11:30 samth: (error '(side-condition variable_1 (regexp-match #rx"^[a-zA-Z!?-]*$" (symbol->string (term variable_1)))))
11:30 samth: that pattern
11:30 samth: and it generated (error 'F2)
11:31 samth: I haven't been able to get this to happen in a simpler example
11:31 samth: the `redex-check' expression is: (redex-check chaprack
11:31 samth: (side-condition e (null? (term (free-locations () e))))
11:31 samth: (same-result? (term e))
11:31 samth: #:attempts 1000)
11:32 samth: clklein, see above
11:38 clklein: samth: OK, I'm looking into it. Thanks for the report!
11:40 (join) carleastlund
11:41 bremner: does Mathew Flatt ever hang out on IRC?
11:42 samth: bremner, no
11:42 bremner: thanks
11:42 samth: nobody here but us chickens
11:42 bremner: wrong scheme variant :)
11:43 samth: :)
11:43 samth: what is this scheme you speak of?
11:43 bremner: oh, just some racket to trick people into using parens
11:44 (join) sstrickl
11:45 (join) danbrown
11:46 clklein: samth: I can't get it to happen with just the pattern. Can you send the whole program?
11:48 samth: clklein, see pltgit:sstrickl/papers/chaperones , model.rkt & test.rkt
11:50 clklein: thanks
12:07 (join) fjl
12:08 samth: another thing, clklein
12:08 samth: it can be very confusing when you've failed to provide a metafunction, and (term (mf x)) just produces '(mf x)
12:09 clklein: I know. I hate that.
12:09 samth: so, i wish something somewhere were different so that was better
12:09 clklein: heh
12:09 samth: how's that for actionable feedback
12:10 clklein: I've considered making you declare all of your keywords.
12:10 clklein: Robby thinks that would be too painful. He might be right.
12:10 samth: yeah, it might be
12:10 samth: but it took me about 15 minute to track down that bug
12:11 carleastlund: I am always wary of systems where identifiers can be references to binders, but do something else silently if there's no binder.  Syntax pattern bindings and Redex metafunction bindings both have this unfortunate property.
12:12 clklein: Just yesterday, I was helping someone get started Redex-ifying a part of their dissertation, and we ran into one of these bugs.
12:12 clklein: It took several minutes to track down.
12:13 clklein: It's very embarrassing.
12:13 clklein: That this class of bug exists; not that it took me minutes to track down.
12:13 sstrickl piles on.
12:14 clklein: carleastlund, sstrickl: Do you agree that declaring your keywords would be too much of a burden?
12:14 carleastlund: Clarify for me what this would look like.
12:16 (join) MayDaniel
12:21 clklein: Actually, I'm not sure how to make it work with `term'. I was thinking that there'd be some kind of define-keywords form that declares names like `if', `begin', `:', etc., and every symbol would need to be one of those, a meta-function name, or a reference to a bound pattern variable.
12:21 clklein: But `term' needs to be able to construct expressions that contain symbols representing variables.
12:23 carleastlund: This starts to sound like a job for unquote-style escaping to me when one wants to jump from literal terms to metafunction calls, or have some identifier that signifies metafunction calls so instead of (mf x) you have ($ mf x) or something.  But I may be more willing to make these things explicit than other people.
12:25 clklein: That still leaves the problem of binding e in a rule's LHS and using e-oops in the right, though that's probably much less common.
12:26 carleastlund: I think that one is not *quite* so evil to debug, but I could be wrong.
12:26 (quit) bjnrapt: Read error: Connection reset by peer
12:26 (join) Lajla
12:27 clklein: I think it just happen less frequently. I can't remember making that mistake personally, for instance.
12:29 carleastlund: I've definitely done it more than once.
12:29 clklein: samth: BTW, I did reproduce the bug; in fact, I got an even weirder output: (error (quote (loc x1)))
12:32 samth: that is very weird
12:33 samth: clklein, if you declared the list of metafunctions up-front, then variables with those names could be rejected
12:33 samth: part of the problem is that (term x) is valid in the absence of a language
12:34 clklein: Ah, that might be a good compromise.
12:35 samth: and if you didn't declare anything, it would keep the current behavior
12:35 clklein: Matthias already wants the term generate to avoid choosing variables whose names coincide with the names of non-terminals.
12:35 samth: clklein, but that's so great for finding bugs :)
12:36 clklein: that's why I'm reluctant
12:37 clklein: But I can see how the printed form of such a counterexample can easily be mistaken for a non-ground term, where the odd variable name looks like a meta-variable representing any term from that non-terminal
12:42 (quit) jonrafkind: Ping timeout: 255 seconds
12:44 clklein: samth: I think it's actually your reduction relation that's injecting this bogus `error' expression.
12:45 samth: hmm
12:45 samth: well, that would explain why i couldn't find smaller counterexamples
12:45 clklein: I put a contract on the reduction relation, and it blames the `let' rule
12:45 samth: oh, i know what's going on
12:46 samth: substitution is descending inside (quote x)
12:46 samth: and substituting x with a fresh variable
12:46 samth: which may have a new name that has a number in it
12:46 sstrickl: yay for contracts
12:47 clklein: This is both a win and a loss for Redex.
12:47 clklein: It found a bug in something that, arguably, you shouldn't have had to write.
12:51 (join) _ryanc_
12:52 clklein: (or should have been easier to write)
12:54 (quit) danbrown: Remote host closed the connection
12:54 sstrickl: clklein: Are you talking about the substitution metafunction?
12:54 clklein: sstrickl: yea
12:59 (join) jonrafkind
13:01 jonrafkind: clklein, is there a way to change the default size of the boxes that show up in `traces' ?
13:01 jonrafkind: i want them wider than they are by default
13:07 samth: jonrafkind, your line-number stuff causes a scribble warning
13:08 clklein: jonrafkind: I think you might be able to control by supplying a pretty printer that produces wider lines. I'll check.
13:08 samth: raco setup: WARNING: undefined tag in scribblings/framework/framework.scrbl:
13:08 samth: raco setup:  ((lib "framework/main.rkt") text:line-numbers<%>)
13:11 clklein: jonrafkind: Yes, that works. For example, #:pp (λ (x) (format "~s" x)) produces boxes that are all single lines
13:12 samth: _ryanc_, ping
13:16 lisppaste: clklein pasted "`traces' box width" at http://paste.lisp.org/display/116201
13:16 clklein: jonrafkind: ^^^
13:21 (quit) asumu: Read error: Connection reset by peer
13:29 (quit) hanDerPeder: Quit: hanDerPeder
13:44 (join) hanDerPeder
13:51 (quit) hanDerPeder: Read error: No route to host
13:56 (join) danbrown
14:19 (join) hanDerPeder
14:35 (quit) chturne: Ping timeout: 255 seconds
14:36 (join) asumu
14:38 samth: _ryanc_, ping
14:39 (join) jao
14:58 (join) mceier
15:02 (quit) Lajla: Quit: Phallic Illuminatus
15:03 (quit) sstrickl: Quit: sstrickl
15:10 (join) anRch
15:41 _ryanc_: samth, pong
15:42 samth: _ryanc_, rename-transformers confuse the macro stepper
15:43 _ryanc_: how so?
15:43 samth: also, the idea of the 'not-free-identifier=? property is very broken
15:43 samth: you can look at a module with a rename transformer in it and not see any expansion steps
15:45 _ryanc_: hmm, I have an idea of what might be going on, but can you send me an example with what you expect to see, just in case/
15:45 samth: working on it
15:46 samth: also, the fact that you can write programs that just behave incorrectly without that syntax property is very bad
15:47 samth: darn, just hung the macro stepper
15:49 _ryanc_: In addition to your bug about text updates, I've still got some work to do to make the macro stepper more responsive and interruptible.
15:50 samth: this program repeatably hangs the macro stepper:
15:51 samth: #lang racket/load
15:51 samth: (module m1 racket
15:51 samth: (define f 1)
15:51 samth: (define-syntax f* (make-rename-transformer #'f))
15:51 samth: (provide (rename-out [f* f])))
15:51 samth: (module m2 racket
15:51 samth: (require 'm1)
15:51 samth: f)
15:51 samth: (require 'm2)
15:51 samth: similar programs should show the bug w/ rename transformers
15:51 samth: the bug is that you expand m2, and you don't see any expansions of `f'
15:52 samth: and the syntax properties say that it's just `f', not `f*'
15:52 jonrafkind: oh good catch samth about line-numebrs<%>, i meant to do it but it slipped my mind
15:53 samth: _ryanc_, you should make sure not to do that in a drracket you care about
15:53 samth: as it will be gone for good
15:54 (quit) hanDerPeder: Quit: hanDerPeder
15:57 samth: and now _ryanc_ is gone for good :)
15:57 _ryanc_: well, not entirely
15:58 _ryanc_: my drracket instance, though...
16:08 _ryanc_: samth, anyway, I think I know why it's not showing the step and why it's confusing the binding information
16:08 _ryanc_: not sure why that example hangs drracket, though; I've tried the non-tool version of the macro stepper, and it seems to handle it fine
16:21 (quit) anRch: Quit: anRch
16:31 (quit) MayDaniel: 
17:01 (join) Fare
17:17 (quit) jonrafkind: Read error: Connection reset by peer
17:38 (join) jonrafkind
17:40 (nick) samth -> samth_away
17:48 (join) seth
17:51 seth: when my script encounters an error, i get a backtrace === context === etc.  is there a way to know the actual line number or sexp?
17:51 jonrafkind: you can try to use errortrace
17:51 jonrafkind: but its not always completely accurate
17:52 jonrafkind: racket -l errortrace -t blah.rkt
17:52 seth: thanks
17:52 jonrafkind: make sure you erase the compiled/ directory if you have one
17:56 seth: if the script has #!/usr/bin/racket on the first line and #lang racket/load on the 2nd, how would I do this?
17:56 seth: i can run racket -l errortrace -t blah.scm, but the script has arguments
17:56 seth: "racket -l errortrace -t blah.scm --something" doesn't seem to work
17:57 seth: nor does "racket -l errortrace -t blah.scm -- --something"
17:59 seth: ah (require errortrace) did it.  thanks again
18:11 (join) Lajla
18:27 (quit) danbrown: Remote host closed the connection
19:08 (quit) asumu: Ping timeout: 265 seconds
19:16 (quit) _ryanc_: Quit: Leaving
19:40 (quit) mceier: Quit: leaving
20:06 (quit) jonrafkind: Ping timeout: 264 seconds
20:12 (join) danbrown
20:31 (quit) carleastlund: Quit: carleastlund
20:37 (join) asumu
21:09 (quit) pygospa: Ping timeout: 255 seconds
21:11 (join) pygospa
22:23 (join) emma
22:50 (quit) danbrown: Remote host closed the connection
22:50 (quit) emma: Ping timeout: 276 seconds
23:04 (quit) rudybot: Remote host closed the connection
23:06 (join) rudybot
23:07 (join) danbrown
23:08 (quit) mario-goulart: Ping timeout: 245 seconds
23:23 (join) emma
23:30 (quit) lisppaste: Ping timeout: 255 seconds
23:40 (quit) emma: Read error: Connection reset by peer
23:40 (join) emma_
23:40 (quit) masm: Ping timeout: 265 seconds
23:41 (nick) emma_ -> emma
23:42 (join) masm
23:42 (quit) masm: Client Quit