00:04 (join) dnolen
00:16 (quit) ohwow: Ping timeout: 260 seconds
00:16 (join) ohwow
00:18 (quit) dnolen: Quit: dnolen
00:26 (join) jeapostrophe
00:27 (quit) jeapostrophe: Client Quit
00:56 gf3: found a minor typo
00:56 gf3: http://docs.racket-lang.org/reference/places.html#(tech._place)
00:56 gf3: a few paragraphs down
00:56 gf3: "For example, the following expression lanches two places…"
00:56 gf3: lanches → launches
00:58 (join) hussaibi_
00:58 (quit) 13WAALCEC: Read error: Connection reset by peer
01:09 asumu: gf3: Thanks, there'll be a fix in git soon.
01:09 gf3: I guess I could have submitted a pull req
01:09 gf3: next time
01:19 (quit) ohwow: Ping timeout: 258 seconds
01:26 (join) ohwow
01:28 (quit) Nanakhiel: Ping timeout: 252 seconds
01:29 (join) RacketCommitBot
01:29 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://bit.ly/ncsOUL
01:29 RacketCommitBot: [racket/master] Fix typo. - Asumu Takikawa
01:29 (part) RacketCommitBot
01:33 (quit) ohwow: Ping timeout: 260 seconds
01:38 ozzloy: dang, that was quick
01:46 gf3: :)
01:48 (join) hakkum
01:52 (quit) kennyd: Ping timeout: 250 seconds
01:55 (join) kennyd
01:58 (quit) jonrafkind: Ping timeout: 255 seconds
02:10 (join) ohwow
02:16 (quit) ohwow: Ping timeout: 264 seconds
02:17 (join) ohwow
02:20 (join) realitygrill
02:30 (quit) ohwow: Remote host closed the connection
02:35 (join) ohwow
02:40 (quit) kennyd: Ping timeout: 246 seconds
02:44 (join) kennyd
02:53 (quit) kennyd: Ping timeout: 276 seconds
02:57 (join) kennyd
03:07 (join) hkBst
03:50 (join) RacketCommitBot
03:50 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://bit.ly/qWpx8c
03:50 RacketCommitBot: [racket/master] New Racket version 5.1.2.4. - Eli Barzilay
03:50 (part) RacketCommitBot
04:48 (quit) 31NAAUN3H: Ping timeout: 240 seconds
04:49 (quit) hussaibi_: Ping timeout: 260 seconds
05:45 (join) masm
05:47 (quit) hkBst: Quit: No Ping reply in 180 seconds.
05:47 (join) hkBst
05:47 (quit) rekahsoft: Ping timeout: 255 seconds
05:48 (join) rekahsoft
06:22 (join) rekahsof1
06:24 (join) MayDaniel
06:24 (quit) MayDaniel: Changing host
06:24 (join) MayDaniel
06:27 (quit) rekahsoft: Ping timeout: 255 seconds
06:27 (quit) DT``: Ping timeout: 255 seconds
06:34 (join) Lajla
06:34 (join) DT``
06:37 (join) tfb
06:39 (quit) realitygrill: Quit: realitygrill
06:40 (quit) hkBst: Read error: Connection reset by peer
06:48 (join) hkBst
07:21 (quit) DGASAU: Ping timeout: 250 seconds
07:25 (join) DGASAU
07:58 (join) dnolen
08:41 (join) jeapostrophe
08:46 (join) RacketCommitBot
08:46 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://bit.ly/oCdshc
08:46 RacketCommitBot: [racket/master] use the supplied security guard in one more place where compiler/cm needs write permission - Robby Findler
08:46 (part) RacketCommitBot
09:57 (quit) dnolen: Quit: dnolen
10:25 (quit) jeapostrophe: Quit: jeapostrophe
10:26 (join) ChibaPet
10:28 (join) RacketCommitBot
10:28 RacketCommitBot: [racket] plt pushed 3 new commits to master: http://bit.ly/r34sh6
10:28 RacketCommitBot: [racket/master] download LNCS and JFP class files on demand - Matthew Flatt
10:28 RacketCommitBot: [racket/master] fix doc typos - Matthew Flatt
10:28 RacketCommitBot: [racket/master] ftp: fix data download for some servers - Matthew Flatt
10:28 (part) RacketCommitBot
10:28 (quit) MayDaniel: Read error: Connection reset by peer
10:39 offby1: No brain, no problem!
10:39 offby1: We lend to anyone!
10:44 ohwow_: Hm
10:44 ohwow_: this package actually looks very interesting :D
10:44 ohwow_: Now I want to develop some sort of static analysis tool with Racket + Racketlog
10:45 ohwow_: *Racklog
10:45 (quit) hkBst: Remote host closed the connection
11:03 (join) jonrafkind
11:12 (join) dnolen
11:21 (join) samth
11:28 gf3: coooooool
11:29 gf3: kinda lame that it requires DrScheme
11:38 (join) jeapostrophe
11:39 (quit) jeapostrophe: Client Quit
11:53 (join) rlb3`
11:55 jonrafkind: can keywords be used with `define-syntax-rule' ?
12:18 (quit) tfb: Quit: sleeping
12:20 clklein: jonrafkind: seems to work
12:21 (join) anRch
12:23 (join) dherman
12:25 jonrafkind: can you paste the code
12:25 jonrafkind: racket needs a database of examples..
12:26 (quit) anRch: Ping timeout: 255 seconds
12:29 clklein: sure, one second
12:32 (join) anRch
12:41 clklein: jonrafkind: http://paste.lisp.org/display/123941
12:44 jonrafkind: ah interesting
12:50 jonrafkind: clklein, are you sure the keyword works and #:keyword is not just acting as a pattern variable?
12:50 jonrafkind: i mean a literal
12:50 jonrafkind: try to do (uses-keyword)
12:50 jonrafkind: and define the keyword as #:keyword [x 5]
12:51 clklein: All I tried are those two examples. It doesn't act as a pattern variable in the second one.
12:54 jonrafkind: right but it looks like a literal
12:58 (quit) anRch: Ping timeout: 250 seconds
12:59 clklein: Oh, I see what you want.
12:59 clklein: No, define-syntax-rules doesn't do that.
13:01 clklein: If you want something like keyword arguments, you need to use syntax-parse. Check out the examples section of the syntax-parse docs.
13:03 (join) RacketCommitBot
13:03 RacketCommitBot: [racket] plt pushed 3 new commits to master: http://bit.ly/oraBOi
13:03 RacketCommitBot: [racket/master] start making planet thread safe - Robby Findler
13:03 RacketCommitBot: [racket/master] fix the initial frame position code for the multiple monitors case - Robby Findler
13:03 RacketCommitBot: [racket/master] change some prints to fprintfs to make it easier to remove debugging printfs from this file when checking in - Robby Findler
13:03 (part) RacketCommitBot
13:03 (join) anRch
13:08 (quit) anRch: Quit: anRch
13:10 jonrafkind: clklein, well yes.. but i want to use define-syntax-rule :p
13:11 clklein: I don't blame you.
13:12 clklein: syntax-parse is a big step up from handlign keywords by hand, but it's still a big step down from `define'
13:12 (nick) depy -> Guest9666
13:13 jonrafkind: its interesting that keywords are the new way of doing symbolic identifiers in syntax-parse
13:13 jonrafkind: because right now if you want to make a literal you have to bind it
13:13 jonrafkind: but in syntax-case you can just use identifiers as literals whenever you want
13:14 (join) RacketCommitBot
13:14 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://bit.ly/oc7lKj
13:14 RacketCommitBot: [racket/master] cocoa: fix sign on `get-display-left-top-inset' x result - Matthew Flatt
13:14 (part) RacketCommitBot
13:14 (join) depy
13:15 (join) anRch
13:16 clklein: It's generally considered bad practice for a module not to provide bindigns for its literals.
13:17 clklein: I'm guessing that syntax-parse is trying to encourage good behavior.
13:18 jonrafkind: right
13:19 jonrafkind: but its interesting that if you want the old behavior you can still get it (through keywords)
13:19 clklein: You mean the #:-thingies, not the identifiers that might be considered a macro's keyword (e.g., `else' in `cond')?
13:20 jonrafkind: yea
13:21 clklein: It's a little different because keywords don't have bindings, but I see what you mean.
13:24 clklein: Actually, wait. How is this any different than just comparing identifiers symbolically?
13:24 clklein: If that's bad, then why aren't keywords as literals bad?
13:25 clklein: I take it this is what you meant. Took me a while to catch up :)
13:25 (quit) masm: Quit: Leaving.
13:26 (join) RacketCommitBot
13:26 RacketCommitBot: [racket] plt pushed 12 new commits to master: http://bit.ly/oDUKve
13:26 RacketCommitBot: [racket/master] Fixes define-judgment-form at the top-level with errortrace enabled - Casey Klein
13:26 RacketCommitBot: [racket/master] Changes syntax for define-judgment-form mode and contract specs - Casey Klein
13:26 RacketCommitBot: [racket/master] Adds an alternative syntax for define-judgment-form rules - Casey Klein
13:26 (part) RacketCommitBot
13:30 (quit) dherman: Read error: Connection reset by peer
13:31 (join) dherman
13:37 (join) RacketCommitBot
13:37 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://bit.ly/qTTyWE
13:37 RacketCommitBot: [racket/master] Revert "start making planet thread safe" -- didn't mean to push that one. Whoops, sorry. - Robby Findler
13:37 (part) RacketCommitBot
13:48 samth: clklein: the difference is that keywords (like numbers) *can't* be bindings
13:50 (join) jeapostrophe
13:54 dnolen: samth: question about the Racket pattern matcher if you have a moment.
14:08 jonrafkind: is it a crazy idea if .rkt files were just assumed to have #lang racket on top of no #lang is found?
14:10 clklein: samth: so?
14:10 samth: jonrafkind: yes
14:11 samth: clklein: because they can't be bound, confusing things relating to binding can't happen
14:11 jonrafkind: samth, care to explain further?
14:11 samth: and you know that keywords are data, whereas you don't know that arbitrary identifiers are data
14:11 samth: dnolen: ask away
14:12 samth: jonrafkind: one of the central parts of racket (IMO) is that the language is specified linguistically and not environmentally
14:13 samth: jeapostrophe: drdr almost to the front page!
14:13 jonrafkind: seems sort of redudnant to say 'foo.rkt' and then #lang racket
14:13 jonrafkind: might as well just say 'foo'
14:13 jeapostrophe: samth: yes :)
14:13 samth: but it isn't!
14:13 jeapostrophe: very close
14:14 dnolen: samth: so I saw that you used some bit's from the Maranget paper (compiling pattern matching to good decision trees). However when I read your paper on Racket's implementation I saw that you went with a backtracking implementation anyway, any particular reason?
14:14 samth: what about #lang typed/racket or #lang lazy or #lang scribble/manual or #lang setup/infotab or ...
14:14 jonrafkind: so then write those
14:14 samth: dnolen: which paper?
14:14 jonrafkind: but if nothing is given then #lang racket seems like a nice default
14:14 clklein: samth: What confusing thing can happen if you compare identifiers symbolically?
14:15 dnolen: samth: Extensible Pattern Matching
14:15 samth: dnolen: ok
14:15 samth: i see the confusion
14:15 samth: maranget has several papers
14:15 dnolen: samth: Compiling Pattern Matching to Good Decision Trees
14:15 samth: the racket impl basically follows his 2001 paper, but takes some optimizations from the decision trees paper
14:16 samth: the decision tree implementation has exponential code size in the worst case
14:16 samth: and i found the backtracking implementation simpler to understand
14:16 samth: jonrafkind: the whole point is that #lang racket doesn't have special status
14:18 jonrafkind: its on the same level as public static void main
14:18 samth: dnolen: i also took some stuff from marangets paper on warnings (although there are no warnings issues by `match')
14:18 dnolen: samth: yeah the code-size I understand, though I thought they avoided that with maximal sharing via hash-consing.
14:18 dnolen: samth: yeah I've looked at the warning paper as well.
14:18 jonrafkind: maybe it should be a raco command instead, 'raco run' will try to figure out what #Lang you wanted
14:18 samth: dnolen: you can't do that if you're macro expanding
14:19 dnolen: samth: in Racket because you're not expanding a data structure?
14:19 samth: dnolen: the expansion of `match
14:19 samth: is just more syntax
14:19 samth: so there's no way to create the sharing internally in the implementation
14:20 samth: whereas the ocaml impl is in the compiler, so they can construct whatever sort of ast they want
14:20 dnolen: samth: what I mean is, if you had say CL  macros, no obstacles tho right?
14:20 samth: racket could be extended to support hash-consed syntax
14:20 samth: dnolen: i don't think so
14:20 samth: how do CL macros help here?
14:26 dnolen: samth: you take the source build the ast, then emit new source
14:27 dnolen: samth: perhaps you can already do this in Racket as well, I've never delved much into Racket macros capabilities
14:33 samth: dnolen: the problem is the new source is exponential in size
14:33 samth: you can only avoid that if your macro output (sexpressions in CL, syntax objects in Racket) support explicit sharing
14:35 dnolen: samth: yeah in the decision tree paper, in the first example they lift the shared code into a local fn
14:36 (join) realitygrill
14:41 (quit) anRch: Quit: anRch
14:46 (part) rlb3`: "ERC Version 5.3 (IRC client for Emacs)"
14:52 (join) jao
15:01 dnolen: samth: what it mean if Racket supported hash-consed syntax? How those branches actually be shared in the final output?
15:01 dnolen: what would it mean
15:23 samth: dnolen: first, sharing those functions doesn't avoid the exponential blowup -- you can have exponentially many calls to those shared fns
15:28 samth: dnolen: hash-consed syntax would be compiled only once (although scoping, macro expansion, etc get very tricky)
15:33 dnolen: samth: hmm I guess I can't quite "see" the exponential blowup from calls to the shared fns. I did find it interesting that for the patterns the Maranget experimented with, they never saw code blowup more than 1.5-2X over maximal sharing.
15:34 samth: i think the exponential construction is in maranget 2001
15:34 samth: rather, i think it's given there; it was known earlier
15:43 (quit) jeapostrophe: Quit: jeapostrophe
15:46 (join) jeapostrophe
15:46 (quit) jeapostrophe: Client Quit
15:46 (join) RacketCommitBot
15:46 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://bit.ly/nxmwUZ
15:46 RacketCommitBot: [racket/master] More TR test failures more easily detectable. - Vincent St-Amour
15:46 (part) RacketCommitBot
15:47 (quit) kennyd: Ping timeout: 250 seconds
15:58 dnolen: samth: thx much this was helpful.
16:06 (quit) ohwow_: Quit: leaving
16:09 (join) tfb
16:20 (join) RacketCommitBot
16:20 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://bit.ly/nD39nw
16:20 RacketCommitBot: [racket/master] Fixes Redex doc's links to examples - Casey Klein
16:20 (part) RacketCommitBot
16:30 (join) RacketCommitBot
16:30 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://bit.ly/oHLNkn
16:30 RacketCommitBot: [racket/master] small fix to plural - Danny Yoo
16:30 RacketCommitBot: [racket/master] adding example to the prose about using (for-syntax (for-syntax ...)) - Danny Yoo
16:30 (part) RacketCommitBot
16:31 jonrafkind: probably (for-syntax (for-syntax ...)) is bad style
16:31 jonrafkind: and (for-meta 2 ...) is more clear
16:36 stamourv: jonrafkind: The paragraph just above explain that they are equivalent.
16:36 stamourv: So I don't think there's going to be much confusion there.
16:40 (join) jeapostrophe
16:42 (join) anRch
16:44 (quit) jeapostrophe: Client Quit
17:14 (quit) ChibaPet: Quit: Leaving.
17:15 jasoegaard: Guessing contest: How many pages in the Racket documentation contains all of the words "foo", "bar", and, "baz" ?
17:18 ozzloy: https://gist.github.com/1138256 how do i best determine whether a file was run as a script?  right now i'm using the name of the file, but i'd like something better.  is it possible to determine the name of the current source code file at run time?
17:18 stamourv: jasoegaard: Too many?
17:18 jasoegaard: ozzloy: It is in the vicinity of "this-expression-source-directory" in the manual.
17:19 jasoegaard: ozzloy: Also: lookuo runtime-path
17:19 ozzloy: jasoegaard, thanks!
17:19 jasoegaard: stamourv: actually no - which suprised me
17:19 samth: ozzloy: see `define-runtime-path'
17:20 stamourv: jasoegaard: Ok, so what's the number?
17:20 jasoegaard: 8
17:20 jasoegaard: But - mostly due to the inclusion of srfis and r*rs
17:21 jasoegaard: It is mentioned only two (2)! times in the "proper" documentation.
17:22 jasoegaard: Oh not two. More like 4 or 5.
17:22 jasoegaard: Is is probably due to: "When writing examples, refrain from using nonsense  words like “foo” and “bar.” For example, when documenting member, resist the  temptation to write"
17:23 stamourv: Ah, so it seems we've mostly been following that advice. Good.
17:25 (quit) anRch: Quit: anRch
17:30 jasoegaard: The word "foo" occurs in 51 documents thogh.
17:31 stamourv: Oops.
17:31 stamourv: But how many of these are srfis, or R*Rs
17:31 stamourv: s/R*Rs/R*RSs/1
17:33 jonrafkind: if foo is outlawed then so should non-word variable names
17:34 jasoegaard: 9 of them are from srfis
17:34 jasoegaard: And roughly the same numbers are from r5rs and r6rs docs.
17:54 samth: jonrafkind: "lst" is much more helpful than "foo"
17:59 jasoegaard: I prefer both "a-list" and "xs" to "lst". I like "xs" the most.
17:59 jonrafkind: the hipocracy of such a statement is astounding
18:02 stamourv: Hippocracy = A political system where horses are in power.
18:02 stamourv: This is probably not what you meant ;).
18:03 samth: jonrafkind: the hypocrisy of what statement?
18:03 jonrafkind: well that is an sat word, and i did badly on the english part
18:04 jonrafkind: of both "I prefer both "a-list" and "xs" to "lst". I like "xs" the most." and ""lst" is much more helpful than "foo"", they suggest that because 'foo' isn't used consistently that it is not a good word, ignoring the fact that 'lst' is just as nonsensical as 'foo'
18:06 jonrafkind: as you can tell, purposely misspelling things does not amuse me
18:07 samth: no, "lst" is not a non-sensical as "foo" -- "lst" conveys information where "foo" does not
18:08 jonrafkind: i declare 'foo' to be a list. happy now?
18:08 samth: jonrafkind: you aren't humpty-dumpty
18:09 jonrafkind: lst conveys information because you've been using it so long that its obvious to you what it means,it is not obvious to everyone
18:09 jamessan: lst only conveys information if you understand what it means. granted, lst is a bit more obvious than some common variable names
18:09 jamessan: elt was particularly confusing the first time I saw it
18:10 samth: right, but the same is true of things like "expr" or even full words (like "mutator")
18:10 jonrafkind: its true that some full words aren't immediately obvious, but its at least a starting point. what does 'expr' expand to? who knows
18:11 samth: jonrafkind: what claim are you making?
18:11 jonrafkind: misspelling variables is indefensible
18:15 jamessan: is there an automated test suite that can be run to help verify a build?
18:22 jonrafkind: collects/tests/racket
18:22 jonrafkind: racket -f quiet.ss or something
18:23 stamourv: jamessan: That's for the racket core. Most libraries have automated tests too.
18:23 stamourv: jamessan: Check the docs for DrDr house calls.
18:24 stamourv: DrDr is our continuous testing tool, and house calls make it possible to run it on your own machine.
18:24 jamessan: quiet.ss looks good.  I'll poke around with DrDr stuff too.  might be nice to have a test target in the Makefile :)
18:24 stamourv: It may take a while, though. We have a lot of tests.
18:25 jamessan: I'm basically looking for a way to run some extra sanity tests as part of building the Debian packages
18:25 stamourv: Then the core Racket tests should do.
18:40 (join) jeapostrophe
18:42 (quit) dnolen: Ping timeout: 252 seconds
18:48 jasoegaard: I am having problems with moving a directory containg serialized structs.
18:48 jasoegaard: One more time.
18:48 jasoegaard: I want to move a directory with a data file.
18:49 jasoegaard: The data file contains serialized data including some structs.
18:49 jasoegaard: The serialization format includes the full path to the source file wherin the struct was defined.
18:50 jasoegaard: When I move the root folder containg both source and data file, I can no longer load the old data file.
18:50 jasoegaard: Any obvious solutions?
18:51 samth: jasoegaard: not really
18:51 samth: sed?
18:54 jasoegaard: sed were an option, if it wasn't for the fact that want others to dowload a zip file with both source and data file.
18:56 jasoegaard: Hmm. The format has changed from (lib "serializa.scm:) to racket/serialize. Now the full path is no longer included.
18:56 (quit) jeapostrophe: Quit: jeapostrophe
19:17 (join) jeapostrophe
19:25 (nick) samth -> samth_away
19:34 (join) RacketCommitBot
19:34 RacketCommitBot: [racket] plt pushed 2 new commits to master: http://bit.ly/pCS4ep
19:34 RacketCommitBot: [racket/master] [honu] set associativty for operators - Jon Rafkind
19:34 RacketCommitBot: [racket/master] [honu] add dot as an operator. parse parenthesized expressions - Jon Rafkind
19:34 (part) RacketCommitBot
19:38 jasoegaard: Anyone interested in testing "Little Helper" which provides full text searching on docs.racket-lang.org ?
19:39 jasoegaard: 1. Go to https://github.com/soegaard/little-helper
19:39 jasoegaard: 2. Click the download button. Download and unpack.
19:39 jasoegaard: 3. Open and run   web/run-me.rkt in DrRacket.
19:40 jasoegaard: I am especially interested in whether the user interface makes sense to others.
19:41 jonrafkind: ok
19:42 jonrafkind: does it have to run in drracket?
19:42 jonrafkind: or can i run from the command line
19:42 jasoegaard: The command line ought to be fine.
19:42 jasoegaard: run-me.rkt just starts a servlet
19:42 jonrafkind: oops i just did a git pull on the plt tree.. have to rebuild first
19:45 (join) dnolen
19:53 jasoegaard: Just tested from the command line. It works if you delete the contents of two compiled/ directories.
19:54 jasoegaard: It seems DrRacket ignores them, so I didn't notice this before.
19:57 (quit) Lajla: Quit: Phallic Illuminatus
19:59 jonrafkind: i see compiled directories under planet/
19:59 jonrafkind: shold i erase those?
19:59 jasoegaard: yes just delete them
20:05 (join) Lajla
20:13 (join) RacketCommitBot
20:13 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://bit.ly/n4YbZ1
20:13 RacketCommitBot: [racket/master] trace the language-position require(s) specially - Robby Findler
20:13 (part) RacketCommitBot
20:18 (quit) dherman: Quit: dherman
20:20 jonrafkind: jasoegaard, how do I make the index?
20:20 jasoegaard: The index is included.
20:20 jonrafkind: also I did a similar thing using jetty and lucene a few months ago
20:20 jonrafkind: with-input-from-file: cannot open input file: "/tmp/docs.racket-lang.org/r6rs-lib-std/r6rs-lib-Z-H-19_html.txt" (No such file or directory; errno=2)
20:20 jonrafkind: i guess it should point at data
20:21 jasoegaard: open config.scm
20:21 jasoegaard: then change the reposioty-root to ...
20:21 jasoegaard: (build-path (current-data-directory) "docs.racket-lang.org")
20:22 jonrafkind: that should work.. why does it end up at /tmp?
20:22 jonrafkind: the only /tmp is in original-repostiory-root
20:22 jasoegaard: I download the docs.racket-lang.org to /tmp and made the index there.
20:23 jasoegaard: During the indexing all html files are converted to txt files.
20:23 jonrafkind: but shouldn't (current-data-directory) point at ./data ?
20:23 jonrafkind: where does the /tmp come from?
20:24 jonrafkind: i changed the-repository-path to be  (build-path "data" "docs.racket-lang.org") but its still pointing at /tmp
20:24 jasoegaard: It comes from the seriazlization. I thought I had eliminated it, but I must have overlooked something.
20:24 jonrafkind: ok ill just symlink it to /tmp
20:26 (quit) realitygrill: Quit: realitygrill
20:26 jonrafkind: seems to work I guess
20:26 jasoegaard: That ought to work.
20:26 jonrafkind: takes a long time, i searched for 'string'
20:26 jasoegaard: The first search load the index. Subsequent searches ought to be faster.
20:26 jonrafkind: oh yea now its faster
20:26 jonrafkind: how does your search compare with lucene?
20:27 jasoegaard: But string has a lot of hits and I show them all - maybe I should just show the first 50?
20:27 jasoegaard: I haven't tried lucene.
20:27 jasoegaard: I had fun implementing the algorithms from "Managing Gigabytes:.
20:28 jonrafkind: anyway i gave up on my thing becuase theres a git search too
20:28 jonrafkind: http://git.racket-lang.org/
20:32 jasoegaard: I could probably solve my problem simply by copying the repository into the data folder. Do the indexing, and then delete everything but hte txt-files.
20:33 jasoegaard: Do the regexp search button make sense?
20:35 (join) ohwow_
20:37 jonrafkind: what does a keyword consist of
20:37 jonrafkind: i mean i typed 'str.*', so what will that match?
20:38 jasoegaard: Each document is split into keywords.
20:38 jasoegaard: The syntax of a keyword is a Scheme identifier.
20:39 jasoegaard: str.* (or simply str) first finds all keywords containing str then finds all documents that contain those keywords
20:39 (join) RacketCommitBot
20:39 RacketCommitBot: [racket] plt pushed 1 new commit to master: http://bit.ly/nWiwEc
20:39 RacketCommitBot: [racket/master] Added a match expander for objects in unstable. - Asumu Takikawa
20:39 (part) RacketCommitBot
20:39 jasoegaard: It is useful when you know a function is named list- something.
20:40 jonrafkind: list- just brings up srfi-1
20:40 (join) 45PAAF7NN
20:40 (join) hussaibi
20:40 jasoegaard: Hmm. With regexp search? That's a bug.
20:41 jonrafkind: oh oops i might have just done regular search
20:41 jonrafkind: ok now it finds a lot of stuff
20:42 jasoegaard: ok
20:43 jasoegaard: But it reveals another bug. Some snippets are missing when searching for list- .
20:45 jasoegaard: Try searching for   easter   Robby (I think) has included an example that shows how to make easter eggs.
20:46 (quit) dnolen: Quit: dnolen
20:47 jonrafkind: yea
20:55 jasoegaard: For some reason GitHub thinks the language is JavaScript.
20:55 jasoegaard: How on earth do I change that?
20:55 jonrafkind: hm.. eli added racket as a language to github
20:56 jonrafkind: maybe change .scm to .rkt?
20:59 jasoegaard: That might work. I just don't understand where the javaScript came from. I don't think there are any js-files at all.
21:00 jonrafkind: probably github's lexer got confused
21:01 jasoegaard: It says 86% is JavaScript, 12% is Scheme and 2% is Racket.
21:01 jonrafkind: you mean for the whole project?
21:02 jasoegaard: yes
21:10 jasoegaard: The extenstions ss and scm are not recognized as Racket: https://github.com/github/linguist/blob/master/lib/linguist/languages.yml
21:11 (join) dnolen
21:15 jasoegaard: Ah! the file plt-index.js is huge!
21:16 jasoegaard: 3.4 Mb !
21:22 (quit) jonrafkind: Read error: Operation timed out
21:36 asumu: jasoegaard: I got an exception when I searched.
21:36 asumu: "with-input-from-file: cannot open input file: "/tmp/docs.racket-lang.org/web-server/dispatch_html.txt" (No such file or directory; errno=2)"
21:39 jasoegaard: asumu: I need to change the serialization from full paths to relative paths.
21:40 jasoegaard: In the mean time: Make a symbolic link:   Change to the little-helper folder.
21:40 jasoegaard: Then  ln -s data /tmp   (I think)
21:42 jasoegaard: ln -s data/docs.racket-lang.org /tmp/docs.racket-lang.org
21:44 asumu: I just copied over the data files into tmp instead, which seemed to work.
21:44 jasoegaard: ok
21:49 asumu: This is pretty neat. "inspector" got me exactly what I was looking for. "match expander" was a bit off.
21:50 jasoegaard: The ranking is based on how many times the word is mention (and the page size is factored in too).
21:55 ozzloy: jasoegaard, thanks again.  here's what your suggestion led me to: https://gist.github.com/1138256
21:55 ozzloy: jasoegaard, also i think you have your ln backwards.  ln -s file-that-already-exists name-of-symlink
21:56 ozzloy: at least, that's how i think of it.  like cp file-that-exists name-of-copy
21:57 jasoegaard: ozzloy: Yep. In this case the data/docs.racket-lang.org exists and /tmp/docs.racket-lang.org must be created.
21:57 ozzloy: oic
21:57 jasoegaard: But I need to find way around that, so "it just works"
21:58 jasoegaard: Kinda funny to see (simplyfy-path (cleanse-path ...))  .
21:58 jasoegaard: I need to lookup cleanse-path to see what it does.
21:58 ozzloy: yeah, i thought so too.
21:59 ozzloy: on unix it does // -> /
21:59 ozzloy: on ms idk, that page was really long and i'm not on ms anyways
21:59 ozzloy: i was about to ask if you could look at it and make suggestions. :)  thanks
22:00 offby1: it can be dirty & complex, like my car's engine ...
22:00 offby1: dirty & simple like my toothbrush
22:00 ozzloy: as long as it runs?
22:00 offby1: clean & complex like my espresso maker
22:00 offby1: clean & simple like my code :)
22:00 ozzloy: hah, nice
22:00 (quit) jao: Ping timeout: 252 seconds
22:00 ozzloy: brb
22:07 ozzloy: thanks so much for looking at the code
22:07 ozzloy: this is my first real lisp btw
22:12 ozzloy: i did (simplify-path (cleanse-path ...)) because that's in the description for (current-directory).  also without simplify the comparison failed if i ran it with ./foo.rkt
22:35 (quit) jeapostrophe: Quit: jeapostrophe
23:40 (join) dherman