00:00 (join) tauntaun
00:10 (quit) tauntaun: Quit: Ex-Chat
00:12 (join) littlebobby
00:12 (join) realitygrill
00:17 (join) lewis1711
00:19 (quit) realitygrill: Ping timeout: 264 seconds
00:25 (join) yoklov1
00:27 (join) realitygrill
00:30 (quit) neilv: Ping timeout: 255 seconds
00:36 lewis1711: https://gist.github.com/1064251 can anybody help with this type error? I have no idea why a polymorphic function can't be applies here
00:41 jonrafkind: vectorof a?
00:41 jonrafkind: meaning the operations should be a vector?
00:45 lewis1711: the arguments should be
00:46 lewis1711: actually I can't even do a simple function :/
00:47 lewis1711: (: add (Number -> Number -> Number)) (define (add a b) (+ a b))
00:47 lewis1711: is an error
00:50 lewis1711: oh, I am using ML or whatever notation
00:50 lewis1711: should be (Number Number -> Number)
00:51 lewis1711: herp derp working now
00:58 jonrafkind: ah thast what I thought at first
00:59 jonrafkind: but i havent used typed/racket enough to be sure
01:04 (quit) ehine1: 
01:05 (join) saint_cypher
01:15 lewis1711: just submitted a bug report on drrackets handling of colours. maybe I'm the only one that likes a white on black colour scheme with visible line numbers lol
01:16 jonrafkind: i fixed that a long time ago
01:16 jonrafkind: what version of drracket are you using?
01:17 lewis1711: 5.1
01:17 lewis1711: is there a new version?
01:17 lewis1711: will grab it
01:17 jonrafkind: i thought it was suppsoed to work in 5.1..
01:17 (quit) yoklov1: Quit: Leaving.
01:17 jonrafkind: ill try in the latest
01:18 lewis1711: there's nothing in the 5.1.1 release notes
01:18 jonrafkind: works in 5.1.1
01:18 lewis1711: alright, i'll try it out
01:19 (quit) saint_cypher: Ping timeout: 246 seconds
01:20 jonrafkind: http://bugs.racket-lang.org/query/?cmd=view%20audit-trail&database=default&pr=11655&return_url=http%3A%2F%2Fbugs.racket-lang.org%2Fquery%2F%3Fdatabase%3Ddefault%3Bdebug%3D%3BState%3Dany%3BSynopsis%3Dline%2520numbers%3Bmultitext%3D%3Bcolumns%3DState%3Bcolumns%3DSynopsis%3Bcolumns%3DCategory%3Bcolumns%3DLast-Modified%3Bcolumns%3DRelease%3Bcmd%3Dsubmit%2520query%3Bsortby%3DNumber
01:20 rudybot: http://tinyurl.com/3brcfdx
01:20 lewis1711: 5.1 also is very hard to read with a black on white gtk colour scheme. a small but sizeable portion of programs fail this test, so I am reluctant to submit a bug report about that.
01:20 lewis1711: as I am not sure if it's the themes fault or dr. rackets
01:23 jonrafkind: you're building 5.1.1 from source?
01:23 lewis1711: no, downloading it over a painfully throttled connection
01:23 lewis1711: using the shell script installer for ubuntu
01:23 jonrafkind: ok
01:23 jonrafkind: actually i should try that too
01:24 jonrafkind: oh 5.1 might have come out just before my fix
01:27 lewis1711: sweet. download should be done by christmas :)
01:27 jonrafkind: whats your download speed?
01:28 lewis1711: 7.3 kB/s
01:29 jonrafkind: omfg
01:34 jonrafkind: yea it doenst work in 5.1 but it should work in 5.1.1
01:36 jonrafkind: ill jsut reply to your bug saying it should be fixed and to confirm
01:36 jonrafkind: oh robby already did
01:53 (join) neilv
01:59 (quit) realitygrill: Quit: realitygrill
02:37 (quit) jonrafkind: Ping timeout: 252 seconds
02:38 ohwow: Can somebody please explain to me what does '(#%# xexpr ...) renders an X-expression forest.' mean?
02:38 ohwow: xexpr forest?
02:38 eli: neilv: I have an example where speed is going to be a very big problem.
02:39 eli: ohwow: forest = many trees.
02:39 ohwow: I still don't get it :(
02:40 eli: ohwow: XML is made of text file with a tree structure of nested nodes, right?
02:42 ohwow: Hm, so it's just for generating several XML-expressions?
02:43 eli: Yes, because what I said is wrong -- it's not a single tree.
02:43 ohwow: Why can't I just write (formlet (,(=> input-string x) ,(=> input-string y))) ?
02:43 eli: That, I have no idea.
02:43 mk: a forest is a set of trees, with more than one tree in it
02:43 mk: a tree is a graph with no cycles. If you have a bunch of disconnected trees, you get a "forest" (haha)
02:44 mk: xml, iirc, is a tree, not a forest
02:45 mk: the root node is the document itself... it's somewhat weird to call it a tree in the first place
02:46 eli: mk: If the root node is the document itself, then the *contents* of an xml document is a forest.
02:46 eli: In any case, "forest" is not a joke, it's the way it's used elsewhere in CS.
02:46 mk: they have sibling relations between them
02:47 lewis1711: that's new. a sequence of disconnected trees is a forest. I like it
02:48 lewis1711: also, is it possible to include information about the length when defining a type in typed racket?
02:48 (quit) neilv: Ping timeout: 255 seconds
02:49 lewis1711: I have (define-type NVector (Vectorof Number)), but I'd like the vectors length to be part of the type information, so that an NVector of length 3 is a different type to one of length 2
02:49 ohwow: Do you guys know any other languages which uses contracts?
02:49 ohwow: I think contracts is a nice feature
02:50 mk: it's an official name, but it's also a "joke" in the sense that it's amusing, in the same way that physicists amuse themselves by calling subatomic particles truth and beauty etc.
02:50 lewis1711: ohwow: D uses it.
02:52 mk: lewis1711: to be pedantic, a forest is a set, not a sequence ;)
02:52 lewis1711: a set is a kind of sequence though. type hierarchy!
02:53 mk: a sequence is a kind of set
02:54 ohwow: Hey guys, I don't know if it's a bug, but most of the examples on http://docs.racket-lang.org/web-server/formlets.html dont work for me. I think the problem is in `text-input' procedure. However, if you replace `text-input' with `input-string' it works fine.
02:54 ohwow: I have Racket 5.1.1
02:54 ohwow: should I file a bug or smth?
02:55 lewis1711: mk: double checked that. my mistake.
02:56 lewis1711: it appears sequence means what i hitherto called an "ordered sequence". I wonder what the super generic term is
02:56 mk: what's an unordered sequence?
02:56 ohwow: a set?
02:56 ohwow: oh no
02:57 lewis1711: something that is neither a set or an ordered sequence
02:58 lewis1711: so a collection of objects without order, but also without the set properties of only having one instance of each object
02:58 lewis1711: or something.
02:58 mk: I think that's a bag
02:59 (join) realitygrill
02:59 mk: yep. aka multiset
02:59 lewis1711: ha. the terms in computer science and maths are probably different as well
03:02 eli: mk: In that sense, "forest" is a joke too, as well as 90% of CS terms.
03:03 mk: I suppose, but forest is a particularly good one
03:03 lewis1711: I think I might be able to get vector legnths with this All syntax
03:04 ohwow: Can someone please check that for me?  < ohwow> Hey guys, I don't know if it's a bug, but most of the examples on
03:04 ohwow: http://docs.racket-lang.org/web-server/formlets.html dont work for me. I think the problem is in `text-input'
03:04 ohwow: procedure. However, if you replace `text-input' with `input-string' it works fine.
03:04 lewis1711: (All (A) (Listof A) -> Natural) can anybody tell me what sort of type that describes?
03:09 (join) neilv
03:27 (quit) realitygrill: Quit: realitygrill
03:41 (join) masm
04:24 (join) Skola
04:29 (quit) roblally: Quit: roblally
05:19 (join) roblally
05:20 (quit) roblally: Client Quit
05:21 (join) roblally
05:40 (quit) neilv: Ping timeout: 255 seconds
06:10 (quit) Penten: Remote host closed the connection
06:19 (quit) JuanDaugherty: Read error: Connection reset by peer
06:30 (part) lewis1711
07:35 (join) Lajla
07:52 (join) yoklov1
08:00 (quit) yoklov1: Quit: Leaving.
08:13 (quit) masm: Quit: Leaving.
08:13 (join) yoklov1
08:14 (join) yoklov2
08:14 (quit) yoklov1: Read error: Connection reset by peer
08:37 (join) racket
08:37 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/kTMTd1
08:37 racket: [racket/master] another fix for `vector-{ref,set!}' on chaperone/impersonator - Matthew Flatt
08:37 (part) racket
08:37 (quit) yoklov2: Quit: Leaving.
08:43 (join) yoklov1
08:45 (quit) yoklov1: Client Quit
08:58 (join) mithos28
09:04 (join) racket
09:04 racket: [racket] plt pushed 1 new commit to master: http://bit.ly/isSJhR
09:04 racket: [racket/master] scribble: clarify intent of `#:use-sources' - Matthew Flatt
09:04 (part) racket
09:05 ohwow: Hello, I have problems with WebSockets library for racket
09:05 ohwow: It just doesnt work
09:06 ohwow: http://pastebin.com/sM9v5MnC
09:06 ohwow: this code just wont start up a server
09:07 (join) realitygrill
09:08 (quit) realitygrill: Client Quit
09:15 (join) lucian
09:18 (join) yoklov1
09:24 (join) realitygrill
09:30 (quit) realitygrill: Quit: realitygrill
09:43 (join) dnolen
09:47 (quit) dnolen: Client Quit
10:10 (quit) mithos28: Quit: mithos28
10:13 (quit) mk: Remote host closed the connection
10:17 (nick) elliottcable|zzz -> elliottcable
10:20 (join) tauntaun
10:38 (join) dherman
10:40 (quit) martinhex: Read error: Connection reset by peer
10:47 (join) martinhex
10:52 (join) dnolen
11:00 (join) neilv
11:00 (quit) neilv: Changing host
11:00 (join) neilv
11:01 (nick) samth_away -> samth
11:08 (nick) danking_ -> danking
11:09 (quit) neilv: Quit: Leaving
11:20 (quit) yoklov1: Quit: Leaving.
11:25 (join) anRch
11:41 (quit) dherman: Remote host closed the connection
11:41 (quit) roblally: Quit: roblally
12:13 (quit) anRch: Quit: anRch
12:15 (join) dherman
12:19 (join) anRch
12:34 (join) jonrafkind
12:37 (join) shofetim
12:37 (join) MayDaniel
12:40 (join) yoklov1
12:43 (quit) anRch: Quit: anRch
12:44 (join) masm
12:49 (quit) MayDaniel: 
12:51 (quit) rgee: Ping timeout: 258 seconds
12:52 (join) rgee
12:57 (join) MayDaniel
13:02 (quit) MayDaniel: Read error: Connection reset by peer
13:21 jonrafkind: how can I evaluate an expression in phase 1 if the current phase is 0?
13:30 (join) mithos28
13:30 stamourv: jonrafkind: begin-for-syntax?
13:31 jonrafkind: I want to run (syntax-local-value x) from phase 0
13:31 jonrafkind: but that won't work because it needs to be run at phase 1
13:31 jonrafkind: so how do I send x from phase 0 to phase 1
13:31 stamourv: You can't.
13:31 (quit) rgee: Ping timeout: 244 seconds
13:32 jonrafkind: well.. maybe theres some trick
13:32 stamourv: You can serialize it, or something, but you can't just cross phases like that.
13:33 jonrafkind: something with eval I guess
13:38 (quit) yoklov1: Quit: Leaving.
13:38 (quit) mithos28: Quit: mithos28
13:47 (quit) masm: Read error: Connection reset by peer
13:49 (join) masm
13:56 (quit) masm: Quit: Leaving.
13:57 samth: jonrafkind: (let-syntax ([x e]) ...)
13:57 samth: e is at phase 1
13:58 (join) jao
13:59 (join) masm
14:03 jonrafkind: but I can't make e deal with any phase 0 values
14:04 jonrafkind: well maybe i can just do the whole test at phase 1
14:04 (quit) masm: Read error: Connection reset by peer
14:07 (join) carleastlund
14:14 (join) anRch
14:19 (join) masm
14:21 (join) rgee
14:34 (join) rgrinberg_
14:36 (quit) rgee: Ping timeout: 276 seconds
14:38 (quit) masm: Quit: Leaving.
14:39 tauntaun: samth: a couple of days ago I asked about Slideshow and you pointed me to some of the talks on your website.  Any chance I could get the Racket source code to a couple of the flashier ones, say, "Growing Software: From Scripts to Programs" and "Logical Types for Untyped Languages"?
14:39 (quit) rgrinberg_: Quit: Leaving
14:40 tauntaun hopes the request isn't inappropriate!
14:40 (join) rgee
14:41 (join) masm
14:42 (quit) masm: Client Quit
14:47 (join) mithos28
14:48 stamourv: mithos28: Have you seen Matthew's reply?
14:49 (join) superjudge
14:52 (join) masm
14:53 jonrafkind: so how can I get a syntax object at phase1 imbued with lexical information from phase 0? read-syntax seems to produce an unenriched ssyntax object
14:53 jonrafkind: basically I just want to emulate #lang
14:55 carleastlund: jonrafkind, #lang does not add lexical information to a syntax object
14:55 jonrafkind: wonderful
14:55 (quit) dnolen: Ping timeout: 252 seconds
14:56 jonrafkind: ah I know ill just make a regular macro since that what module-begin does
14:58 carleastlund: jonrafkind, what kind of context are you trying to add here?  Perhaps there is an appropriate mechanism?
14:59 jonrafkind: lexical context
14:59 jonrafkind: I want #'foo to come from the foo.rkt module
14:59 carleastlund: Right, but for what purpose?  The context of a module's base language?
15:00 jonrafkind: uhm yea I guess
15:00 carleastlund: Or perhaps you just want some core form to refer to some specific #'foo, I take it.  Then yes, it sounds like a macro is appropriate.
15:01 carleastlund: For comparison, take the reader macro that turns 'x into (quote x).  It doesn't determine what the binding for quote is -- that's up to the program.  You get a different quote depending whether you use racket or bsl or whatever.
15:05 (join) dnolen
15:10 (join) racket
15:10 racket: [racket] plt pushed 9 new commits to master: http://bit.ly/mlkQ5X
15:10 racket: [racket/master] Added srclocs to errors raised by match. - Eric Dobson
15:10 racket: [racket/master] Add types for kernel struct constructors. - Eric Dobson
15:10 racket: [racket/master] Fix match to not use apply and work with TR. - Eric Dobson
15:10 (part) racket
15:14 (quit) mithos28: Quit: mithos28
15:25 (quit) superjudge: Read error: Connection reset by peer
15:26 (join) superjudge
15:35 (quit) superjudge: Quit: superjudge
15:40 (join) zx80
16:06 (quit) zx80: Ping timeout: 255 seconds
16:21 (quit) anRch: Quit: anRch
16:23 (quit) tauntaun: Quit: Ex-Chat
16:29 (join) racket
16:29 racket: [racket] plt pushed 2 new commits to master: http://bit.ly/rdSOVh
16:29 racket: [racket/master] Fix syntax test harness to work with compiled tests. - Vincent St-Amour
16:29 racket: [racket/master] Fix props. - Vincent St-Amour
16:29 (part) racket
16:34 (quit) dherman: Quit: dherman
16:34 (join) tauntaun
16:36 (join) mithos28
16:39 (join) dherman
16:44 (join) yoklov1
16:55 (join) roblally
16:59 (quit) dherman: Quit: dherman
17:18 samth: tauntaun: "Growing Software" was written in keynote
17:21 samth: tauntaun: "Logical Types" is now at: http://www.ccs.neu.edu/home/samth/tmp/logical.tgz
17:25 tauntaun: samth: Many thanks...and sorry to put you to the trouble.
17:26 samth: tauntaun: no, it was easy
17:26 tauntaun: (I've downloaded "Logical Types")
17:26 samth: tauntaun: there's currently a compile error w/ the latest git
17:26 samth: b/c there's now a binding for `identity'
17:27 tauntaun: samth: I'm not sure that'll be a big deal for me, as I wanted the source mostly to correlated it with the snazzy effects; I should be able to iron out bugs by checking the docs.
17:27 tauntaun: s/correlated/correlate
17:37 Skola: how's racket for webapps :[]
17:37 Skola: is it used often for webapps?
17:38 stamourv: The Racket web server is quite nice, and it's fairly widely used.
17:39 ohwow: It's pretty neat
17:39 ohwow: I havent been using it much, but I've been skimming through the docs and playing with it for the past 3 days
17:41 Skola: I'm wondering how fast development is/can be
17:42 Skola: not for a LISP god, but on average
17:43 stamourv: The documentation is pretty good, so that probably flattens the learning curve a bit.
17:43 jonrafkind: well theres no debugger, so keep that in mind
17:45 Skola: ok
17:49 (join) bugQ
17:51 danking: jonrafkind: The REPL doesn't count?
17:51 jonrafkind: not even close
17:52 danking: jonrafkind: What does it lack?
17:53 jonrafkind: analyzing runtime data - variables, stack trace
17:53 jonrafkind: stepping through code
17:53 shofetim: Skola: after the learning curve it feels to be about as fast as PHP + a framework of your choice. You get more value as the complexity of the application increases.
17:56 danking: hmm.
17:56 danking: It would be nice to have scheme48's repl in Racket.
17:57 jonrafkind: does that have the abilities I alluded to?
17:58 danking: You can dump the current stack
17:58 danking: And I'm pretty sure you can analyze the current environment frame as well.
17:59 jonrafkind: oh yea that would be magnificent
17:59 danking: The interface to do this isn't entirely obvious to me but I've definitely looked at stack traces and replaced bogus values mid-execution after an exception was thrown.
17:59 danking: I think that last part is especially sexy.
18:13 (quit) dnolen: Quit: Page closed
18:17 (quit) yoklov1: Quit: Leaving.
18:20 (join) yoklov1
18:21 (quit) yoklov1: Client Quit
18:26 _p4bl0: danking: I think MIT Scheme has this kind of debugging features too, maybe even Guile, but not quite sure about the latter
18:44 (join) realitygrill
18:47 (quit) masm: Quit: Leaving.
19:06 (quit) lucian: Remote host closed the connection
19:16 (quit) Skola: Ping timeout: 240 seconds
19:19 jao: _p4bl0, yes, guile 2.x has that
19:21 (quit) bugQ: Ping timeout: 240 seconds
19:27 (join) bugQ
19:35 _p4bl0: jao: okay, I have the debian package and the git version on my laptop so that's why I wasn't sure I guess ^^
19:47 (part) shofetim: "ERC Version 5.3 (IRC client for Emacs)"
20:03 (quit) tauntaun: Quit: Ex-Chat
20:03 (join) tauntaun
20:09 (quit) jonrafkind: Ping timeout: 246 seconds
20:18 (quit) fmu: Ping timeout: 276 seconds
20:22 (join) fmu
20:29 (join) dsrguru
20:30 dsrguru: hi
20:30 dsrguru: I have three servlets that link to each other
20:30 dsrguru: and each one provides interface-version, manager, and start
20:31 dsrguru: and I have an entry.rkt servlet
20:31 dsrguru: that calls another provided function in one of them
20:31 dsrguru: to render the starting page of the web app
20:31 dsrguru: since they all provide start
20:31 dsrguru: the interpreter is telling me there's a naming conflict
20:32 dsrguru: is there a way around having a servlet provide a function called "start" ?
20:32 dsrguru: (the error message was entry.rkt:12:9: module: identifier is already imported)
20:34 (join) Demosthenes
20:40 (join) dherman
20:41 (join) yoklov1
20:48 (quit) mithos28: Quit: mithos28
20:53 (quit) roblally: Quit: roblally
20:53 (join) mithos28
21:10 eli: dsrguru: You can require it with a different name, or with a prefix or whatever.  But more generally, I don't see why you need to require all of them.
21:10 eli: If they're servlets, then the web server requires them for you.
21:11 dsrguru: when I didn't require them
21:11 dsrguru: and then clicked on a link
21:11 dsrguru: that pointed to those servlets
21:11 dsrguru: I would get a not found page
21:11 dsrguru: saying that that servlet didn't provide a start function
21:12 eli: That sounds like something else is wrong.
21:12 dsrguru: hmm
21:12 (join) Penten
21:12 eli: You probably didn't *provide* a `start' function -- that's needed.
21:12 dsrguru: oh
21:12 dsrguru: like literally
21:12 dsrguru: have a start function
21:12 eli: But there's no need to require the file from some other file.
21:12 dsrguru: oh
21:12 dsrguru: wait
21:12 dsrguru: should I export the start function
21:13 dsrguru: using the provide keyword
21:13 eli: Yes.
21:13 dsrguru: ok
21:13 dsrguru: just don't require that servlet's file?
21:13 eli: Yes, there shouldn't be any need to require it that I see so far.
21:13 (join) dnolen
21:14 dsrguru: oh I see
21:14 dsrguru: so in entry.rkt
21:14 dsrguru: I've defined
21:14 dsrguru: start
21:14 dsrguru: to equal a function
21:14 dsrguru: in another module
21:14 dsrguru: so I'm requiring that module
21:15 dsrguru: but now
21:15 dsrguru: running `racket entry.rkt'
21:15 dsrguru: doesn't create a web server
21:15 dsrguru: but instead
21:16 dsrguru: entry.rkt:12:9: module: identifier is already imported at: start in: (define-values (start) (new-lambda (request) (render-module2-page request)))
21:17 eli: dsrguru: You went beyond what I can understand from an irc chat -- better to send the question to the list.
21:17 dsrguru: kk
21:17 dsrguru: eli: thanks for your help anyway
21:17 eli: No problem -- and please do send it there.
21:18 (quit) Demosthenes: Ping timeout: 240 seconds
21:18 dsrguru: I will if I don't resolve it in the next few minutes
21:18 (quit) dherman: Quit: dherman
21:22 (part) mithos28
21:22 (join) mithos28
21:26 (join) neilv
21:26 (join) Demosthenes
21:46 (quit) carleastlund: Quit: carleastlund
21:46 (join) lewis1711
21:48 lewis1711: ok, non-trolling question: how "complete" is typed racket considered? is it still experimental at this stage? because I find it very difficult to use (and yes that could just be down to me:P)
21:50 eli: It's complete enough that it is no longer "experimental".
21:53 neilv: lewis1711: it is being used for some real work
21:54 neilv: the biggest thing i think it's missing is keyword arguments, which i'm very fond of
21:56 lewis1711: I find it hard to use because the error messages can be rather long and cryptic. for reference I can generally parse what ghc tells me
21:56 eli: neilv: So I have this convenient utility which packs some files into a single one in a readable format, possibly running a command on the result.
21:56 eli: For example, `multifile -c emacs /tmp/X **/*.rkt' will pack all the *.rkt files in /tmp/X, run emacs on the result, then unpack it back.
21:57 eli: It's really useful for doing global changes.
21:57 eli: Using your thing would be a challenge though...
21:57 neilv: lewis1711: that might be a good thing to mention on the email list
21:57 neilv: eli: yes
21:59 neilv: eli: it has a watchdog timer to make sure it doesn't take too long parsing. if it does, it sets a 'hopeless' flag on the buffer and very little works
22:00 neilv: (if meow-hopeless-blv (error (format "Sorry, Meow already gave up on this buffer (%s). Try restarting meow-mode, or try scheme-mode."
22:01 neilv: eli: if your thing puts a ton of files into one buffer
22:03 neilv: eli: you generally do that for search&replace across multiple files?
22:04 lewis1711: neliv, alright, will do
22:06 (quit) Demosthenes: Ping timeout: 240 seconds
22:07 dsrguru: eli: I figured it out
22:08 dsrguru: eli: in entry.rkt I called serve/servlet with a function called start
22:08 eli: dsrguru: Good.
22:08 eli: neilv: Well, if you do parse it incrementally, then the right way to do things would be to (1) parse enough to show the current screen and a bit more, (2) have an idle timer thing to do more stuff "in the background", and (3) switch back to parsing more as you move around.
22:08 dsrguru: that function conflicted with the start that I imported in the module that contained the rendering function
22:09 dsrguru: so I just had to rename the start function in entry.rkt
22:09 lewis1711: "Type Checker: untyped identifier sequence-fold imported from module <typed/racket> in: sequence-fol" does this mean that typed racket can't do any type checking with that?
22:09 eli: neilv: The thing is that this is hard, and if instead you'd bend `font-lock' to do what you want, then you'd get it for free,
22:09 eli: s/,/./
22:09 lewis1711: I would use vector-fold but no such function exists.
22:10 eli: lewis1711: Best to ask on the list, or wait until samth or stamourv are around.
22:10 neilv: eli: i know it's hard. font-lock gets in the way of doing it correctly, however
22:11 eli: neilv: If you do plan to do the proper thing, then you have a much larger problem that you're solving...
22:11 eli: (There were some parsing frameworks for emacs around, I think, maybe they do those things?)
22:12 neilv: if you mean "semantic" and the emacs ide stuff, i looked at that after i wrote my scanner, but it was not up to the task
22:12 eli: neilv: And yes, I do these for all kinds of difficult search/replace jobs, where the search/replace are more difficult.  (Eg, refactored code that has a slightly different interface.)
22:13 (join) dherman
22:13 neilv: eli: i would like to do that kind of cross-file search&replace using sexp-based pattern-matching and manipulation anyway. and there's no reason you have to kludge it as all files in one buffer
22:14 eli: Doing some sexpr-based thing is more like resorting to some sed script -- it reduces the advantage you get from an interactive work.
22:15 neilv: eli: there's no reason it can't be interactive
22:16 neilv: and people have done various cross-file commands for emacs over the years
22:16 (quit) mithos28: Quit: mithos28
22:16 eli: Yes, it can, but I said "reduced" -- you're beyond the edit-a-buffer situation, which is not as interactive as it can be.  (For example, what happens when you want to do some other kind of editing in two edge cases?)
22:17 eli: It's the same reason why this is generally more convenient than using grep-mode, for example.
22:17 neilv: i know what you're talking about
22:18 neilv: we could also do a special mode that is sensitive to file boundaries but lets you keep your single buffer
22:18 (join) mithos28
22:18 (join) Demosthenes
22:19 eli: BTW, something that is slightly related -- I was talking about Emacs vs Drr, and one thing that came up is how Emacs keeps `find-file' inside itself rather than use some gtk dialog -- you get an immediate advantage of scripting openning files via keyboard macros that you can't do otherwise.
22:19 neilv: but this is a tangent. 99% of users will not be loading all of the racket core source files into a single emacs buffer
22:20 eli: Re file boundaries -- right, that would be a solution to that -- something like clone buffer or indirect buffer, but then that goes back to the single-file interface on top of the multifile one...
22:20 eli: Yes, I'm fully aware that this is not a typical use case...
22:20 neilv: i have a prioritization, and 1GB buffers of racket source is pretty low priority
22:21 eli: Sure.
22:25 neilv: i like the idea of refactoring tools and sexp-pattern-based changes across multiple files (which could be semi-interactive)
22:26 neilv: i'd also be open to the idea of adding support for multiple-files-per-buffer. it wouldn't be TOO hard to add
22:27 neilv: we'd get around the expensive scanning part by scanning a file at a time, as needed
22:27 neilv: that would be the easy way to do it
22:28 eli: You mean doing it without my tool?  That would be nice, though I was shooting for something generic because (a) it's much easier, (b) I can use other tools on it.
22:28 neilv: meow would need to recognize the file separation markers
22:30 neilv: i would not start that until i have everything else implemented, because i don't want the 1GB people to be a barrier to functionality for the 99% people
22:32 (join) rgrinberg_
22:35 (quit) rgee: Ping timeout: 255 seconds
22:44 (quit) mithos28: Quit: mithos28
22:47 (quit) littlebobby: Quit: Ex-Chat
22:59 (quit) yoklov1: Quit: Leaving.
23:10 (quit) neilv: Ping timeout: 255 seconds
23:26 (quit) bugQ: Ping timeout: 252 seconds
23:27 (join) yoklov1
23:28 (join) yoklov2
23:28 (quit) yoklov1: Read error: Connection reset by peer
23:32 (join) bugQ
23:48 (join) jonrafkind
23:58 lewis1711: jonrafkind: I think I may have found an actual bug now ;)
23:58 lewis1711: afternoon :)
23:58 jonrafkind: in the line numbers?
23:58 lewis1711: no