Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2008-04-16

Perl 6 | Reference Documentation | Rakudo

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:01 mncharity re STD.pm, are :: and ::: getting parsed?  I'm not immediately seeing it.
00:02 TimToady haven't had much chance to work on it, what with $dayjob being ferocious at the moment...
00:03 mncharity ah :/
00:06 TimToady I don't think it does parse :: and ::: at the moment, sb a regex_metachar, but it's not there
00:06 mncharity ok, thanks
00:07 Auzon joined #perl6
00:28 wtgee joined #perl6
01:12 yewenbin joined #perl6
01:14 pugs_svnbot r20293 | putter++ | elf_e_src/EmitRegexYare.pm: Sketching what needs to be done.
01:14 pugs_svnbot r20293 | putter++ | Backtracking api changes from a Filter::Simple to a s/// method.
01:14 pugs_svnbot r20293 | putter++ | Engine core becomes a prelude.
01:14 pugs_svnbot r20293 | putter++ | Emitters become p6, emitting p5 source.
01:14 pugs_svnbot r20293 | putter++ | IR analysis becomes p6.
01:14 pugs_svnbot r20293 | putter++ | The regexp and regex parsers are temporarily retained, as some of the action logic will need to end up in IRx1_FromAST or elsewhere.
01:14 pugs_svnbot r20293 | putter++ | And the unicode rules will need to be metaprogrammed in p6.
01:14 pugs_svnbot r20293 | putter++ | Discarded:
01:14 pugs_svnbot r20293 | putter++ |  Match - no need for painful 'use overload' games.
01:14 pugs_svnbot r20293 | putter++ |  Functional regex representations - we're bootstrapping on STD.
01:14 pugs_svnbot r20293 | putter++ |  Api and Filter packages - emitter has understanding and control of the code.
01:14 pugs_svnbot r20293 | putter++ |  Interactive shell and command line - though may again need something like it for testing.
01:14 pugs_svnbot diff: http://dev.pugscode.org/changeset/20293
01:14 lambdabot Title: Changeset 20293 - Pugs - Trac
01:14 Auzon ... wow.
01:17 bradb joined #perl6
01:18 mncharity yeah.  some commit messages are just not meant for irc. ;)
01:19 mncharity good night all &
01:27 meppl joined #perl6
01:35 silug joined #perl6
01:36 silug joined #perl6
02:00 alanhaggai joined #perl6
02:22 alanhaggai joined #perl6
02:29 rindolf joined #perl6
02:31 cognominal_ joined #perl6
02:42 TimToady ooh, we'll have to allow small comments with ⋕ and even tinier ones with ⌗
02:43 rindolf TimToady: heh.
02:43 rindolf TimToady: hi.
02:43 rindolf Hi all.
02:43 bradb left #perl6
02:43 rindolf I'm going crazy without my home email.
02:44 rindolf I hope I won't get unsubscribed from a lot of mailing lists.
02:45 TimToady I think I might go sane without my home email...
02:48 rindolf TimToady: heh.
02:48 rindolf "Can't live with it. Can't live without it."
02:53 rindolf TimToady: what's up?
02:54 rindolf TimToady: BTW, can you tell me how do you make your living now?
02:59 TimToady I'm working for a company called Netlogic Microsystems
03:00 rindolf TimToady: ah.
03:00 TimToady they do pattern matching in hardware, among other things
03:00 rindolf TimToady: hmm... interesting.
03:00 TimToady and somehow they thought I knew something about pattern matching...
03:01 rindolf TimToady: imagine that!
03:02 Auzon So... like regexes in hardware?
03:02 elmex_ joined #perl6
03:02 rindolf TimToady: many modern Greek speakers can speak ancient Greek, right?
03:02 TimToady they can read it, at least...
03:02 rindolf Someone I talked with once wanted to have a C compiler in hardware.
03:02 rindolf He used Gentoo.
03:03 Auzon Makes sense now.
03:03 TimToady Auzon: yeah, like that
03:03 rindolf TimToady: how much memory can they afford to use?
03:04 TimToady they try to avoid using memory, which is slow...
03:04 rindolf Because compiling a regex into a state machine can take a lot of memory.
03:04 rindolf TimToady: ah.
03:04 rindolf TimToady: not computer memory - on-chip memory.
03:05 TimToady yeah, well, now you're getting into NDA territory :)
03:05 rindolf TimToady: oh sorry.
03:05 rindolf TimToady: NDA or NFA?
03:05 * TimToady snickers
03:05 rindolf J/K.
03:06 rindolf Non-finite Automata.
03:06 TimToady flipflopping on the question, are you?
03:06 rindolf TimToady: heh.
03:06 rindolf TimToady: J/K flip-flop, right.
03:06 rindolf TimToady: can't remember any others.
03:06 rindolf RS Flip-flop.
03:07 rindolf I think.
03:07 TimToady Radio Shack?
03:07 rindolf TimToady: :-)
03:07 rindolf TimToady: I studied EE for 6 years. Well, one year was composed of two vacation semesters.
03:07 rindolf TimToady: I only have a B.Sc. in EE.
03:07 rindolf But it would be a stretch to call me an EEer.
03:08 TimToady well, my family is calling me, so I'd better go...
03:08 TimToady not sure I want to know what they're calling me...
03:08 Auzon See you.
03:08 TimToady bbl &
03:08 rindolf TimToady: bye.
03:14 rindolf We still need to compose the words for "I'm the real TimToady".
03:14 rindolf To be sang by the Randal Schwartz Condition.
03:43 wtgee joined #perl6
03:44 PerlJam blah, too late for TimToady
03:45 rindolf PerlJam: hi.
03:45 PerlJam hello rindolf
03:45 rindolf PerlJam: are you on San-fran.pm?
03:46 PerlJam nope.
03:46 rindolf PerlJam: oh, so you missed the funny thread.
03:46 rindolf Too bad.
03:46 PerlJam do you happen to know how to use the msg bot? It's one of those things I've done maybe once, so it didn't stick in my long term memory.
03:46 rindolf PerlJam: /msg MemoServ HELP
03:46 rindolf PerlJam: or do you mean lambdabot?
03:47 rindolf lambdabot: help
03:47 Auzon I think lambdabot is just @tell whoever whatever
03:47 PerlJam whichever
03:47 rindolf PerlJam: I can grep my logs.
03:47 PerlJam heh, I've stopped keeping logs of IRC.  I rarely ever went back and searched them for anything.
03:49 PerlJam @tell TimToady I think the definition of statement or stdstopper is wrong in STD.pm.  stdstopper includes statement_mod_loop and statement_mod_cond.  Coupled with the definition of statement, that lets us have 2 statement_mod_cond in a row and I thought that's now allowed.
03:49 lambdabot Consider it noted.
03:50 PerlJam Grr.  s/(.*)now/${1}not/
03:51 PerlJam trying to implement more of STD.pm for rakudo is ... trying  :)
03:56 meppl joined #perl6
04:03 alanhaggai joined #perl6
04:17 wtgee joined #perl6
04:32 silug joined #perl6
04:32 eternaleye joined #perl6
04:54 Psyche^ joined #perl6
06:13 BinGOs joined #perl6
06:30 syle joined #perl6
06:30 syle I notice memory is never freed on a variable when you use concatenation on it in perl5, is that normal?
06:35 Alias_ joined #perl6
06:39 silug joined #perl6
07:03 iblechbot joined #perl6
07:08 spinclad syle: this is a perl6 channel.  for perl5 help, try #perl here on freenode, or #perl-help on irc.perl.org or #perlhelp on EFnet (or so purl tells me).
07:55 wknight8111 joined #perl6
09:11 IllvilJa joined #perl6
09:20 ruoso joined #perl6
09:29 hanekomu joined #perl6
09:33 elmex joined #perl6
09:49 agentzh left #perl6
09:49 agentzh joined #perl6
09:58 Psyche^ joined #perl6
10:15 chris2 joined #perl6
11:13 meppl joined #perl6
11:23 Alias__ joined #perl6
11:25 riffraff joined #perl6
11:35 xinming joined #perl6
11:35 masak joined #perl6
12:11 iblechbot joined #perl6
12:22 moritz_ ... and a log crumbles through the gate
12:33 pmurias joined #perl6
12:37 BinGOs joined #perl6
12:44 BinGOs joined #perl6
12:44 dalek joined #perl6
12:48 Juerd_ joined #perl6
12:53 Zygo joined #perl6
12:56 mtkatwork joined #perl6
12:59 mtkatwork left #perl6
13:45 BinGOs joined #perl6
13:52 rdice joined #perl6
13:54 TJCRI joined #perl6
14:00 alester joined #perl6
14:04 Auzo1 joined #perl6
14:40 REPLeffect joined #perl6
14:41 jhorwitz joined #perl6
15:01 kcwu joined #perl6
15:12 peeps[work] joined #perl6
15:16 moritz_ any kp6 developer around?
15:16 moritz_ I noticed a possible problem with gather { ... take }
15:16 moritz_ take() just calls Coro::cede
15:17 moritz_ which has no control over which coro is being executed
15:17 moritz_ so if you have more than one coro, you end up triggering the wrong one
15:17 moritz_ one possible solution could be Coro::Channel
15:22 lisppaste3 joined #perl6
15:27 awwaiid if you know exactly which coro you want to run isntead, you could do transfer()
15:27 brbpack joined #perl6
15:28 moritz_ awwaiid: aye, but then you have the problem that take() has to redispatch to the previous Coro::State object
15:28 moritz_ awwaiid: I tried that with Perl6::GatherTake (now on CPAN), but I didn't succeed
15:28 awwaiid ah
15:29 moritz_ but it could be that kp6's abstraction layer helps here, and makes ->transfer() possible
15:30 moritz_ for the p5 implementation I used a tied, quasi-lazy array
15:30 moritz_ which is kinda ugly
15:30 awwaiid I'm completely unfamiliar with the kp6's architecture, I was just making a completely uneducated suggestion :)
15:31 moritz_ it was certainly educated ;-)
15:32 moritz_ I'n not familiar enough to fix it, I just noticed that kp6 might have a problem
15:38 awwaiid moritz_, see http://thelackthereof.org/projects/perl/Continuity-Widget/lib/Continuity/Coro/Continuation.pm (though a glance at it suggests that the scope of @yieldstack needs to be more limited)
15:38 lambdabot http://tinyurl.com/6a8dbg
15:38 awwaiid maybe you can use that to effectively transfer only back to the caller
15:39 awwaiid Coro used to have Coro::Cont that did this better
15:40 moritz_ awwaiid: I think I'd need a hash that maps the coro reference to the array I'm filling
15:40 moritz_ awwaiid: but something like this might work, thanks for the hint
15:42 awwaiid np
15:44 japhb joined #perl6
15:48 pmurias moritz_: hi
15:48 moritz_ hi pmurias
15:49 pmurias i think kp6 is frozen/dead right now so i don't suppose the bug will be fixed
15:50 moritz_ what is it blocking on? smop?
15:50 moritz_ $no_free_time?
15:52 avar fglock disappeared and development pretty much dried up after that
15:53 moritz_ too bad
15:53 pmurias and the default backend proved too slow to be usefull
15:57 pmurias on the positive side nested gather works ;)
15:58 moritz_ that surprises me a bit
15:59 moritz_ but it's great ;-)
16:00 pasteling "pmurias" at 81.168.228.98 pasted "nested gather test case" (14 lines, 269B) at http://sial.org/pbot/30835
16:04 pmurias moritz_: i'm afraid i was over enthusiastic :(
16:05 pmurias the inner gather is called more times then it should
16:08 justatheory joined #perl6
16:08 Daveman joined #perl6
16:18 awwaiid moritz_, I looked at that code I sent you and now re-understand why using a global yieldstack is ok. the yieldstack holds the caller, so the only reason it is a stack at all is for nested continuations
16:18 awwaiid (s/continuations/generators/ if you like)
16:19 moritz_ awwaiid: ok, so as long as I don't do *really* evil things like threads it should all be fine
16:19 moritz_ but I think that Coro and threads conflict anyway
16:19 awwaiid Coro, afaik, is thread safe as long as it is wholly contained in a single thread
16:20 moritz_ which is not my definition of "thread safety" ;-)
16:20 awwaiid I think you can use my code directly, and add your hash cache as wrapping code
16:21 awwaiid er, array
16:22 moritz_ thanks
16:22 awwaiid Actually I should just bundle this properly as Coro::Generator
16:24 wtgee joined #perl6
16:24 * awwaiid adds it to the TODO list
16:31 pugs_svnbot r20294 | moritz++ | [irclog] filter [off] comments (the logging bot should have don it, but
16:31 pugs_svnbot r20294 | moritz++ | it seems not to work reliably)
16:31 pugs_svnbot diff: http://dev.pugscode.org/changeset/20294
16:31 lambdabot Title: Changeset 20294 - Pugs - Trac
16:43 rindolf joined #perl6
17:03 Auzon joined #perl6
17:30 tobeya joined #perl6
17:37 barney joined #perl6
17:44 eternaleye joined #perl6
17:51 mj41 joined #perl6
17:55 overlast joined #perl6
18:10 rff joined #perl6
18:32 Tene joined #perl6
18:34 Tene joined #perl6
18:44 [particle] joined #perl6
19:01 lichtkind joined #perl6
19:05 lichtkind TimToady: is it right that basically in perl 6 are only 4 namespaces: $%@ and that for (sub)routines ?
19:06 moritz_ there's the :: sigil for namespaces, isn't
19:08 TimToady arguably there's only packages and lexical scopes, and the sigils are just part of the name
19:08 lambdabot TimToady: You have 1 new message. '/msg lambdabot @messages' to read it.
19:12 moritz_ TimToady: is there a 'real' @@ sigil?
19:13 TimToady depends on what you mean by 'real'
19:13 moritz_ TimToady: STD.pm considers it to be a sigil, but it's not listed in the synopsis as a sigil, iirc
19:13 moritz_ TimToady: having @@a being a different variable from @a
19:14 TimToady it parses as a separate sigil, but @@foo and @foo are the same array as it is currently specced
19:14 moritz_ ok
19:14 TimToady all it really controls is the list vs slice context if you bind to it
19:15 PerlJam moritz: http://dev.perl.org/perl6/doc/design/syn/S02.html#Names_and_Variables
19:15 lambdabot Title: Synopsis 2: Bits and Pieces - perl6, http://tinyurl.com/y648lg
19:16 PerlJam The :: sigil seems to have an aweful lot of conceptual weight put upon it as I look at that list.
19:16 moritz_ PerlJam: slice view of @, that's what I missed
19:17 PerlJam though maybe that's just because I don't quite think of all of those items as specializations of the same thing yet.
19:18 PerlJam package/module/class/grammar all go together in my head, but the others don't quite
19:18 lichtkind TimToady: le me put it this way: filehandle, format name, and dirhandle namespaces are gone and captures and pairs and junctions are "simple" skalars?
19:19 TimToady PerlJam: the statement token prevents you from having two conditional modifiers in a row; the stdstoppers are a superset of what the parser will accept at that point (though there are other issues with transient terminators that need fixing)
19:20 TimToady lichtkind: that's a fair statement of the situation
19:20 lichtkind TimToady: thanks
19:21 PerlJam TimToady: I'll have to look at it again, but it seemed like a statement could be parsed as <expr> <statement_mod_cond> <stdstopper>  and stdstopper could be parsed as <statement_mod_cond> thus you'd get two in a row.
19:21 b_jonas joined #perl6
19:21 PerlJam anyway, it was late for me when I looked.  Maybe my eyes were blurry and my brain fuzzy  :)
19:21 TimToady PerlJam: I'd say :: is really type space, and some types happen to map to what we used to think of as packages
19:22 TimToady stdstopper is only doing lookahead
19:22 PerlJam ah, I completely missed that then.
19:23 PerlJam I may implement that in rakudo next if I can figure out how. :)
19:23 TimToady note that all calls to stdstopper are of the form <?stdstopper> or <!stdstopper>
19:23 lichtkind TimToady: what do you know about kurila
19:23 TimToady neither of those consume chars
19:23 * allbery_b sees and understands (cf. mailing list yesterday)
19:24 TimToady as I've mentioned, stoppers need some reworking in order to be properly dynamic
19:27 PerlJam TimToady: aye.  Maybe I won't attempt it just yet.  I only have a limited amount of time to hack these evenings and meager brain power.
19:28 TimToady lichtkind: I ignore kurila, for the most part.  I'm not very interested in half-baked fixes to Perl.
19:28 lichtkind TimToady: that was a straight statement of the situation :)
19:28 bockmabe_ joined #perl6
19:30 Coke_ TimToady: ... that's quite a straight line there... =-)
19:32 buchetc joined #perl6
19:40 lidden_ joined #perl6
19:56 justatheory joined #perl6
19:56 Juerd_ joined #perl6
19:56 agentzh joined #perl6
19:56 jan_ joined #perl6
19:56 xdg joined #perl6
19:58 armagad joined #perl6
20:20 mj41 joined #perl6
20:29 justatheory joined #perl6
20:37 elmex joined #perl6
20:37 larsen_ joined #perl6
20:41 jan_ joined #perl6
20:43 ruoso joined #perl6
20:45 eternaleye joined #perl6
20:54 jferrero joined #perl6
21:08 pugs_svnbot r20295 | moritz++ | [t/spec] s/use v6-alpha;/use v6;/
21:08 pugs_svnbot diff: http://dev.pugscode.org/changeset/20295
21:08 lambdabot Title: Changeset 20295 - Pugs - Trac
21:09 Coke_ ... I mean in parrot.
21:10 moritz_ Coke_: but in p6 it's not necessary either, is it?
21:11 moritz_ Module.pod say "Before the full specification of Perl 6.0.0 is released, you can use C<alpha>
21:11 moritz_ to denote a program using syntax that is still subject
21:11 moritz_ to change"
21:11 moritz_ but since the tests should always reflect the current status of the language I see no need for it
21:12 moritz_ if anybody objects I can reverse that commit
21:14 Coke_ I just didn't want to suddenly have all of pugs tests failing, is all.
21:16 moritz_ since pugs doesn't run in current versions of the repo that's no a good point
21:16 moritz_ and compatibilty to the implementations should be achieved with fudge
21:16 moritz_ but I can understand your concern
21:16 Coke_ Just trying to be a good neighbor.
21:16 moritz_ aye
21:17 moritz_ we do appreciate that ;)
21:20 rdice joined #perl6
21:42 wtgee left #perl6
22:06 kane_ joined #perl6
22:15 wknight8111 joined #perl6
22:20 eternaleye joined #perl6
22:30 r0bby joined #perl6
22:37 meppl good night
23:55 alester_ joined #perl6

| Channels | #perl6 index | Today | | Search | Google Search | Plain-Text | summary

Perl 6 | Reference Documentation | Rakudo