pugscode.org/ | nopaste: sial.org/pbot/perl6 | ?eval [~] <m oo se> | We do Haskell, too | > reverse . show $ foldl1 (*) [1..4]
Set by diakopter on 9 July 2007.
00:00 kanru joined 00:11 b2gills joined 00:12 Blwood joined 00:41 Schwern joined 00:45 Limbic_Region joined 00:51 ashleyb joined 01:03 ofer0 is now known as ofer
pugs_svnbot r16967 | lwall++ | trunk: 01:07
: [Std] use spiffy new infix matchers for regex parsing.
diff: dev.pugscode.org/changeset/16967
lambdabot Title: Changeset 16967 - Pugs - Trac
01:09 miyagawa joined
pasteling "evalbot_r16966" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26194 01:09
01:15 weinig joined
pugs_svnbot r16968 | lwall++ | trunk: 01:15
: [STD] and also parse the new forms, duh...
diff: dev.pugscode.org/changeset/16968
lambdabot Title: Changeset 16968 - Pugs - Trac
diakopter TimToady: I dunno if you got my @message[0] from a few days after yapc (b/c lambdabot was being attacked) - but I thought I'd remind you about the .keys method on an enum array 01:18
which someone asked about during your/damian's talk
TimToady ah, right 01:22
01:26 Alias_ joined 01:34 ChanServ sets mode: +o diakopter, diakopter sets mode: +o Alias_, diakopter sets mode: -o diakopter
devbot6 planet6: Dive Into Perl 6: Operating with operators <diveintoperl6.blogspot.com/2007/07/...s.html> 01:42
lambdabot diveintoperl6.blogspot.com/2007/07/...s.html>
diakopter lambdabot: nice. 01:44
thoughtpolice lambdabot failing? impossible 01:45
Aankhen`` LOL.
thoughtpolice++
thoughtpolice :> 01:46
01:48 willy_nelson is now known as snafu
thoughtpolice i had no idea this channel had a bot for planet6 01:48
you learn something new all the time 01:49
01:49 snafu is now known as willy_nelson, willy_nelson is now known as snafu
Aankhen`` thoughtpolice: Only if you keep an open mind. ;-) 01:49
thoughtpolice: Do you read XKCD?
01:50 snafu is now known as willy_nelson, sunnavy joined
Aankhen`` thoughtpolice: By the way, your HTML encoding seems to have gone wild on the first code snippet. I see this: (1 &amp;amp;amp;amp;amp; 2 & 3 & 4 & 5) 01:50
thoughtpolice Aankhen``: sometimes. i don't browse web comics as much as I'd like to these days (8-bit theatre was my favorite) 01:51
Aankhen``: thanks. so far my experiances with blogger have proven to be notorious in terms of putting things like code in your posts. :/
Aankhen`` :-\ 01:52
thoughtpolice namely, the compose mode seems to be the issue
Aankhen`` See www.xkcd.com/c242.html if you haven't already.
lambdabot Title: xkcd - A webcomic of romance, sarcasm, math, and language - By Randall Munroe
thoughtpolice it randomly likes to do things like take away line breaks
Aankhen`` Ouch.
thoughtpolice yeah. not too much fun. 01:53
so far the problems have proven to be fairly minor though
Aankhen`` That's why I don't have a journal or anything where I post regularly. I don't want to use anything that's out there already, and I don't feel like writing my own. :-)
01:53 nipotaway is now known as nipotan
diakopter I dunno if anyone noticed, but I added a new .xslt for pugs' Subversion repo.. svn.pugscode.org/pugs/ and so on 01:53
Aankhen`` .xslt? 01:54
diakopter stands for...
Aankhen`` Oh, whoa, I didn't realize it was an XML file.
That looks really nice!
diakopter purl: xslt?
buubot: xslt?
Aankhen`` Yes, I know what XSLT is. I just wasn't sure how it applied to the SVN repository. :-)
diakopter buu: xslt?
:)
I figured...
Aankhen`` I like it.
diakopter now to modify it to link back to trac and/or SVN::Web's versions of the same 01:55
Aankhen`` I'm not sure the border-bottom on hover is a good idea, though.
IMO, it would look nicer with alternating background colours and/or a separate background colour on hover. 01:56
thoughtpolice Aankhen``++ # fixed html-dumbness
Aankhen`` ^_^
thoughtpolice diakopter: looks nice. 01:57
Aankhen`` diakopter: Another thing that might be nice is to put `display: block` on the directories/folders. 01:58
That is, put `display: block;` on the `a` elements surrounding them.
(And anything else surrounding the `a` element that might need to be a block element too in order to accommodate it.)
Aankhen`` sighs. 01:59
Checking out Pugs takes a while. :-)
I think it's been about ten minutes so far.
01:59 [particle2 is now known as [particle]
Aankhen`` I'm about to check the Synopses, but in case anyone knows the answer: are `if` and `else` expressions or statements? 02:00
I think this was mentioned recently in here, but I can't remember what the answer was. 02:02
Though I have a sneaking suspicion that it was along the lines of "they're still statements because otherwise it would cause confusion with the statement modifiers".
Hum. I guess they're statements. 02:04
Yay, Pugs still being checked out.
02:08 agentzh joined
daxim I just released my first module. high 5! (or 6) 02:09
Aankhen`` Congrats!
Alias_ Welcome to CPAN daxim 02:10
daxim I used the pause id already longer, for bug reports on RT 02:11
Alias_ Doesn't count :)
daxim ah well :)
Alias_ Submitting to RT just makes you a rabble rouser :) 02:12
Aankhen`` Rouse the rabble! 02:13
Jeez. `svn co` still in progress. 02:15
02:15 fglock joined
Aankhen`` Further updates as we receive information. 02:15
diakopter Aankhen``: hrm 02:17
Aankhen`` Ot 02:20
It's probably something to do with my laptop.
Vista appears to be a major resource hog.
diakopter Aankhen``: are you checking out via svk or svn? and using https or not? 02:22
Aankhen`` SVN, http.
diakopter I just started a fresh checkout; we'll see who wins :D 02:23
Aankhen`` CURSE YOU
[particle] vista ain't half bad with four cpus :)
Aankhen`` [particle]: Laptop with single dual-core CPU and Vista = not so hot. 02:24
Wait, that's incorrect.
It's actually = a little too hot.
diakopter Aankhen``: finished 02:25
Aankhen`` It's interesting to note that probably my favourite part of Vista is the set of new fonts.
diakopter: Still going. :-(
Consolas is a fantastic fixed-width font (being used in the logs, for reference), and Cambria is a beautiful serif font. 02:26
Firefox + Stylish = Wikipedia using Cambria = lovely.
thoughtpolice oi
diakopter Aankhen``: disable indexing for pugs/../ and below... then add an exception for your antivirus for pugs/../ and below... then add an exception for Windows Defender for svn
thoughtpolice looks like I'm going to have to compile erlang from a tarball. :( darn you pkgsrc! 02:27
Aankhen`` diakopter: Aha, it's probably the indexing.
Thanks for reminding me.
diakopter Consolas in iTerm looks great on the mac. if only iTerm's 'use option/alt as "Meta"' actually worked when using UTF-8. 02:28
Aankhen`` diakopter: Would you happen to know how to disable indexing for a particular folder? 02:29
I can't seem to find anything in either the Indexing Options or the folder options.
diakopter folder properties...
check apply to all children 02:30
Aankhen`` Oooh, right, it's in the "Advanced" file attributes.
diakopter I didn't pass 12 MS cert tests for nothing... <orly?> 02:31
Aankhen`` Twelve MS cert tests?!
...why?
[particle] technologies expire. 02:32
diakopter why not!? they're only $125/pop!
Aankhen`` Yes, but why start?
diakopter heh
this was eons ago
er, epochs ago
02:34 ChanServ sets mode: +o diakopter
diakopter TimToady: any ideas for the enumArray.keys? 02:36
Aankhen`` Uh, weird. 02:37
pugs_svnbot: REPORT request failed on '/pugs/!svn/vcc/default'
pugs_svnbot: REPORT of '/pugs/!svn/vcc/default': 200 OK (svn.pugscode.org)
80,69658662A pugs\misc\old_pugs_perl5_backend\Perl6-MetaModel\t\19_CLASS.t
80,69658662svn: REPORT request failed on '/pugs/!svn/vcc/default'
80,69658662svn: REPORT of '/pugs/!svn/vcc/default': 200 OK (svn.pugscode.org)
What the heck?
CRAZY IRC CLIENT
Sorry for the spam.
diakopter cleanses the logs... (j/k) 02:39
Aankhen`` Hehe.
diakopter hrm
I had a similar error to that a few days ago. 02:40
re: *..0 # -infinity to 0 02:42
how does one represent -infinity to 10? *..0..10 ?
02:44 SubStack joined
[particle] -Inf..10 i suspect 02:49
diakopter oh
Aankhen`` I totally knew that.
Like, totally.
(Still running `svn co`, BTW.) 02:50
I think it's been an hour.
Limbic_Region salutations all
Aankhen`` Greetings, o ye of the region limbic.
diakopter I guess x..y..z isn't implemented in pugs. but is it specified? 02:52
[particle] i don't think so. a Range is *..* 02:53
x..y..z is ERETARDO # twisting a phrase from MJD 02:54
Aankhen`` *..* :by(1/Inf) 02:55
diakopter someone's supposed to tell me that x..y..z simplifies to x..z ;)
[particle] only timtoady can tell you that, and that he hasn't put it in the spec yet :)
Aankhen`` No. We're not your nannies!
02:59 mako132_ joined 03:00 torz joined
Limbic_Region calls it a night 03:01
diakopter calls it a spade 03:07
Aankhen`` calls it a wuss. 03:08
Alias_ calls Aankhen`` a wuss. 03:09
Aankhen`` cries.
avar hush now baby baby don't you cry 03:14
spinclad momma gonna sing you a lullaby 03:20
... and if that rocking horse get broke
momma gonna call it a night
03:30 greentea joined
diakopter Perl for Proles 03:45
Aankhen`` stabbity stabs avar.
avar urgh? 03:46
Aankhen`` I was hoping for something more along the lines of "[sound of dying]".
diakopter sounds like 'urgh' when dying 03:47
Aankhen`` If that holds true for avar too, then I retract my previous statement.
avar stabs diakopter to verify
diakopter urghs 03:48
diakopter dies
avar makes a note
diakopter question for anyone who's awake: what CMS, web app platform, wiki (ugh), blogroll, or everything-ish[, or no] software should run on perlsix.com,org,net? and what content should go in it (if any)? 03:51
spinclad doesn't know except let it run in Perl 6 someday 03:53
(helpful, i know)
Patterner TWiki, what else? :) 03:55
diakopter points them to feather for now 03:59
04:07 cernd_ joined 04:10 Eidolos_ joined
thoughtpolice diakopter: what'll the site's purpose serve? analogous to perl.org, or is a perl6 'official site' just kind of a rough sketch right now? 04:29
diakopter thoughtpolice: that's my question 04:30
thoughtpolice diakopter: hm. i dunno, a wiki would be a viable solution (dokuwiki?) 04:32
there's also things like plone if you're into that kind of stuff
diakopter I think the question of "official implementation" is open to debate. I think the question of "official distribution" [incl libraries] of any successful-testsuite-smoking implementation is open to debate.
therefore, "official site" will by definition be a misnomer 04:34
thoughtpolice fair enough. how's about "center of the universe"? :> 04:35
04:39 REPLeffect joined
diakopter It seems to me that [one of ]TimToady's goal[s] for Perl_6_The_Language has been for it to be descriptive over decretive. Not only is there more than one way to do it, there's now *every* [known ]way to do it. 04:44
TimToady There's Not Many Ways Not To Do It! 04:46
diakopter Once a single implementation definitively smokes the test suite, Perl_6_The_Implementation will only exist until someone adds their own libs to the distribution and redistributes it. Then, there will be many Perl_6_An_Implementation/Distribution. 04:47
04:47 bingoxXx joined
TimToady works for linux... 04:48
bingoxXx Hi can anybody help me with this? Can't locate Parse/RecDescent.pm in @INC
I have already installed the required rpms
diakopter Look harder in @INC?
bingoxXx how to do that?
I'm a newbie..
diakopter what program is erroring that? 04:49
TimToady
.oO(why would perl 6 need Parse::RecDescent?)
bingoxXx thats a demo program for parsing 04:50
amnesiac but why on Perl6? 04:51
bingoxXx Its actually perl 5.8.8
04:51 monomorph joined
amnesiac and why are you looking for that in a perl6 channel? 04:51
try out #perl
diakopter that's T-Christmas too early of a version
TimToady ah, this is #perl6, so we were naturally a bit confused... 04:52
bingoxXx ok.. but berfore leaving can somebosy tell me how perl6 is different?
amnesiac TimToady, TheDamian should port that to p6!
agentzh bingoxXx: Parse::RecDescent is cool, but I think you may want to take a look at Pugs::Compiler::Rule :)
TimToady bingoxXx: Perl 6 regexes are rather more powerful than P::R
agentzh bingoxXx: which is a Perl 6 regex engine written for Perl 5 developers. 04:53
diakopter leaves as a conifer after someone else explains
04:53 id_sonic joined
bingoxXx agentzh: does Pugs comes built in? 04:53
agentzh bingoxXx: hmm?
diakopter but wait, there's more! Pugses in the trunk!
bingoxXx or I have to look for the required rpms?
agentzh has no idea what the built-in means in this context. 04:54
TimToady perl 6 is still in development
you're unlikely to find an rpm for it
w
agentzh bingoxXx: See search.cpan.org/perldoc?Pugs::Compiler::Grammar
id_sonic hi, I want to subscribe [email@hidden.address] but I can not.
bingoxXx ok, thanks a lot guys
lambdabot Title: Pugs::Compiler::Grammar - Compiler for Perl 6 Grammars - search.cpan.org
04:54 buetow joined
TimToady we're not very practical here... 04:54
id_sonic I have sent two mails..
diakopter id_sonic: what makes you think you cannot? 04:55
TimToady the last message to that mailing list was June 22
id_sonic diakopter: I don't know.
agentzh bingoxXx: p6 regex may look a bit more verbose than P::RD, but the match object is really handy.
TimToady perl6-language has more traffic, as does perl6-internals 04:56
most of the pugs development happens here though
diakopter id_sonic: did those two emails bounce back to you?
agentzh
.oO( Pugs::Compiler::Grammar is very P::RD-ish ;) )
 04:57
diakopter guesses that the subscribe confirmation emails got caught in id_sonic's spam filter, or vice versa 05:00
id_sonic TimToady: all of them not bounce back me.
agentzh heh, bingoxXx may think 5.8.8 is very "close" to 6.0.0 :) 05:02
id_sonic I'll try again, dinner time, see you, bye.
diakopter it would be, except in our diradic numeropathy, the inner base is uncountable. 05:06
05:08 devogon joined 05:17 _mjk_ joined 05:23 bingoxXx left
diakopter anyone here attending oscon this year? 05:25
oh, and, is there a SotO talk?
me: "what does Perl stand for?" my wife the med student: "pupils equal, round, reactive to light" 05:30
05:37 Aankhen`` joined
dduncan diakopter, I'd luv to, but can't afford it this year 05:40
05:44 cernd joined
Alias_ heh 05:51
diakopter, it's also a singaporean/malaysian instance coffee targetted at women
instant
diakopter pBracketing characters are defined as any Unicode characters with either bidirectional mirrorings...q 05:52
oh, bidirectional. 05:53
got it.
mmm instance coffee 05:54
05:59 Aankhen`` joined 06:07 rho joined 06:14 bsb left 06:19 hakobe is now known as hkb_lab
moritz re 06:20
diakopter reyo 06:22
06:26 elmex joined 06:29 iblechbot joined 06:35 id_sonic joined
diakopter TimToady: what's the argument against making subs implicitly multi? 06:47
moritz diakopter: I think it makes it harder to detect accidential redefinition of a sub 06:48
diakopter well hopefully it couldn't be *too* accidental, I mean, if the signatures are identical, the compiler can squawk...? 06:51
are you imagining accidental redefinition via accidental reuse of a sub's name for another purpose, or accidental reimplementation of the same sub? 06:53
moritz yes ;)
both, actually
Aankhen`` moritz: Hi. 06:54
moritz hi Aankhen`` ;) 06:55
Aankhen`` moritz: Would it be a problem to use HTML::CalendarMonth instead of HTML::Calendar::Simple in index.pl?
diakopter ok, the latter is detectable as I said above (and seems to me, the compiler must die b/c it doesn't know which to run), but the former is just plain not detectable, but seems to me that is extremely rare.
moritz Aankhen``: I think I tried it, and it doesn't produce valid xhtml :(
06:55 Alias_ joined
thoughtpolice hey moritz 06:55
Aankhen`` moritz: Hmm, I didn't realize.
moritz hey thoughtpolice ;) 06:56
Aankhen`` apathetically shakes a fist at people who don't validate.
moritz Aankhen``: wait, I tried HTML::CalendarMonthSimple
Aankhen`` moritz: I checked ::CalendarMonth, it doesn't seem to produce valid HTML either.
moritz Aankhen``: why do you want to get away from the current solution? not enough class="..." tags? *g* 06:58
Aankhen`` Hehe.
06:58 REPLeffect joined
Aankhen`` ::CalendarMonth seemed to have a few useful tricks at first glance, but on closer inspection it isn't all that great. 06:59
moritz thoughtpolice++ # yet another cool diveintoperl6 post
thoughtpolice moritz: ty. :) 07:01
diakopter @tell TimToady in fact, all built-in methods and operators need to be multi, since they are all extensible... so why not make everything implicitly multi? How about a negator for that flag, instead: 'lone' or 'solo' or 'only' or 'single'. Backlog to here for further context/discussion. 07:02
lambdabot Consider it noted.
Aankhen`` HTML-- # it just sucks, okay? 07:03
thoughtpolice agreed
@karma html
lambdabot html has a karma of 0
thoughtpolice appropriate. :)
Aankhen`` HTML-- # needs more negative karma
@karma html
lambdabot html has a karma of 0
Aankhen`` @karma HTML 07:04
lambdabot HTML has a karma of -2
diakopter markup-- escaping--
thoughtpolice internets--
Aankhen`` diakopter: Those aren't what I have a problem with.
I hate the futility of "semantic HTML".
Aankhen`` checks himself in mid-rant and goes off to watch some cartoons.
s/ in//
moritz Aankhen``: at least xhtml2 will have semantic navigation lists 07:05
Aankhen``: and a href= attribute for every tag
diakopter Aankhen``: okay, but if you look for futility, you will endlessly find it to be endless.
Aankhen``: however, I get eye-roll-y at semantic html insisters. 07:09
moritz I use semantic HTML as long as it's pragmatic as well 07:10
but I don't mind using a few <div>s solely for layout purposes
diakopter er, I once did, back when I was making browser apps 07:11
I once [got eye-roll-], I mean
07:12 franck__ joined
thoughtpolice @karma thoughtpolice 07:17
lambdabot You have a karma of 4
thoughtpolice nice
in any case, sleep. night all.
diakopter l8r thoughtpolice
thoughtpolice --status=sleepings
later diakopter
moritz sleep well ;)
diakopter thoughtpolice++
moritz it's 9am here ;) 07:18
thoughtpolice moritz: 2am here
:)
diakopter it's 2:17am here
thoughtpolice me and diakopter live in the same time zone as each other. we should hang out, we have so much in common.
okay, really though. sleeeep.
diakopter heh 07:25
I grew up in TX. assuming I'm actually "grown up". 07:26
moritz can somebody tell me what 'é»„ęœ›ē§‹' means?
diakopter sees three question marks 07:27
moritz diakopter: then you have broken utf-8 support or missing fonts ;)
07:27 zamolxes joined
diakopter yeah... 07:28
moritz probably the latter
diakopter no
moritz you can look into the web irc logs, there it's correct utf-8 in any case 07:29
diakopter i know; thanks :)
moritz anyway, I have no idea what these three graphs mean, I just came across them in my apache log files, somebody searched for this string and found my site ;) 07:32
so now I'm curious
07:33 theorb joined
agentzh moritz: é»„ęœ›ē§‹ seems to a person's name :) 07:34
pugs_svnbot r16969 | moritz++ | trunk:
: [perlhint]
: * renamed P6Explain to perlhint
: * added more data
: * added a small p5 parser
diff: dev.pugscode.org/changeset/16969
lambdabot Title: Changeset 16969 - Pugs - Trac
moritz agentzh: ok, thanks
agentzh moritz: a Chinese person, it seems.
moritz: you're welcome :)
moritz then it was probably mentioned somewhere in the logs, or google just went beserk for no reason ;) 07:35
agentzh interesting
heh, google returns a lot of results while searching é»„ęœ›ē§‹ ;) 07:36
@google é»„ęœ›ē§‹ 07:37
lambdabot news.sina.com.cn/c/216698.html
Title: ōæ½xD4ōæ½xDAōæ½xB1ōæ½xB1ōæ½xBEōæ½xA9ōæ½xB0ōæ½xF3ōæ½xBCōæ½xDCōæ½xD3ōæ½xD7ĶÆōæ½xB5ōæ½xC4ōæ½xD7ļ·øōæ½xBBōæ½xC6ōæ½xCDōæ½xFBōæ½xC7ōæ½xEFŅ»ōæ½xC9ōæ½xF3ōæ½xB1ōæ½xBBōæ½xC5ōæ½xD0ōæ½xB4ōæ½xA6ōæ½xCBōæ½xC0ōæ½xD0ōæ½xCC
agentzh Oh, this page is in GB2312
moritz and it will totally break the logs again ;) 07:38
agentzh *nod*
pasteling "evalbot_r16968" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26198 07:39
agentzh moritz: btw, occasionally audreyt may paste some Big5 chars here ;)
which is the encoding used in .tw
07:40 weinig joined
moritz agentzh: what I could do is to dump some of the database and build testcases with that data... 07:41
agentzh moritz: that's a great idea!
what i need is just the raw data :)
but anyway, have to go back to $work now ;) later &
07:42 weinig joined
moritz @tell agentzh moritz.faui2k3.org/tmp/irclog.sql.bz2 should contain all you need 07:42
lambdabot Consider it noted.
07:43 weinig joined
diakopter Alias_: ping 07:54
07:55 c9s joined 07:56 jisom joined
Alias_ pong 07:58
08:01 alexe joined
diakopter Alias_: what happened to freepan.org? seems to be spammed 08:05
alexe hi,all. How can i use eval to run some code in a constant?
moritz ?eval my $s = "say 'foo'"; eval $s; 08:07
08:07 evalbot_r16959 is now known as evalbot_r16969
evalbot_r16969 OUTPUT[fooā¤] Bool::True 08:07
Alias_ freepan? what's that? 08:09
diakopter dunno, someplace there used to be an /repos/adamk dir
Alias_ there did? :) 08:10
My repository is currently at svn.ali.as/cpan
lambdabot Title: Revision 1925: /
diakopter svn.pugscode.org/pugs/ext/Config-Ti...ig/Tiny.pm
Alias_ P6 Config::Tiny still gives me a chuckle
From like the third day after pugs got module support 08:12
Audrey wanted some example modules that were small and self-contained, cause you could to "use" yet
So I just pasted them into the repository and forgot all about it
diakopter har 08:14
pugs_svnbot r16970 | Darren_Duncan++ | trunk: 08:29
: ext/Muldis-DB/ : updated mainly Core.pod to flesh out nonscalar type descs and add an initial complement of global, boolean, integer, and type-convert operators
diff: dev.pugscode.org/changeset/16970
lambdabot Title: Changeset 16970 - Pugs - Trac
beppu Does pugs support coroutines like the Coro module? search.cpan.org/dist/Coro/ 08:36
lambdabot Title: Marc Lehmann / Coro - search.cpan.org
beppu If so, does it have an asynchronous socket implementation like Coro::Socket? 08:37
08:55 masak joined 09:06 dduncan left
masak moritz: I just want to say I like your idea with the p6 syntax explainer 09:07
marking up Perl-6.0.0-STD.pm sounds like a double win to me 09:08
Aankhen`` ?eval sub foo { yield 1; yield 2; }; [ foo, foo, foo ] 09:24
09:24 evalbot_r16969 is now known as evalbot_r16970
evalbot_r16970 Error: cannot yield() from a routine 09:24
Aankhen`` Hmm.
masak ?eval coro foo { yield 1; yield 2; }; [ foo, foo, foo ]
evalbot_r16970 [1, 2, 1]
Aankhen`` Oh. 09:25
Let's pretend I said that in the first place. :-P
masak :)
09:25 smg left
masak glad to help 09:25
moritz masak: good to know 09:26
masak moritz: further, I doubt showing docs for user-added ops will be very easy to implement 09:27
it would be way cool, but consider it a 'version 2.0' item
moritz masak: aye. I think I'll assemble a bit of test data manually, and take care of of displaying it nicely... 09:29
masak: and then I can start thinking of different data source 09:30
masak moritz++ # sounds good
09:36 theorbtwo joined 09:37 jisom_ joined 09:51 asksh is now known as dwave 09:57 lichtkind joined 09:59 ddwagnz joined 10:00 ddwagnz left
Juerd I think the ** change is great. 10:01
Does, outside regexes, @foo ** '|' now do a @foo.join:'|'? :) 10:03
masak Juerd: maybe it should 10:07
question is what @foo ** '0' does -- both inside and outside of regexes
Juerd I was just joking. It would be very cryptic to have ** be join, with no advantage at all (that I can think of) 10:08
masak Juerd: you're probably right 10:09
Juerd Besides that, ** is typically a numeric operator, so @foo ** 5 ought to do +@foo ** 5
masak yup
Juerd: still, how does @foo ** '1' work in a regex?
Juerd In a regex, @foo ** anything is interesting, actually.
I think that the LHS is not special, so an array would match any of its elements 10:11
@foo ** '3' is probably @foo 3 @foo 3 @foo 3 @foo 3 @foo 3 ...
lambdabot Maybe you meant: faq ft todo yow
wolverian Juerd, btw, did you get my privmsg from a few weeks ago?
Juerd wolverian: If I did, I forgot. 10:12
masak Juerd: um. my question really was, how does 'word' ** '0' work in a regex?
Juerd masak: word [ 0 word ]*
wolverian hm. you're identified, too.
both of us are.. extra weird
Juerd wolverian: Since yesterday
wolverian oh
well, that explains it
Juerd afk # toilet
masak Juerd: IOW, it recognizes '0' as a string, not a number 10:16
Juerd masak: I think so 10:22
10:37 iblechbot joined 10:43 buetow joined 10:44 chris2 joined 10:56 crashmatrix joined 11:15 Blwood joined 11:17 sci3ntist joined 11:20 cmarcelo joined 11:22 Blwood joined 11:26 masak joined
meppl good morning 11:27
masak meppl: good day to you 11:31
sci3ntist Hey guys 11:33
can any1 help me
?
Alias_ purl, ask to ask?
sci3ntist how to share an array in perl, using threads::shared
Alias_ oh wait, wrong channel 11:34
sci3ntist isn't this channel for perl ?
am i wrong ?
Alias_ can u help me ? 11:35
Alias_ Not really, I generally don't use threads
one issue, how big is the array?
sci3ntist not that much 11:36
maybe 200 elements
Alias_ search.cpan.org/~jdhedden/threads-s.../shared.pm
lambdabot Title: threads::shared - Perl extension for sharing data structures between threads - s ...
Alias_ That reads pretty straight forward to me
Share the array by reference
moritz sci3ntist: threads are NYI in perl 6 :(
sci3ntist Hey Guys i've read that , i just got a compilation error when try to use the array 11:37
thats wont work with me
meppl hello masak 11:38
Juerd sci3ntist: This is a Perl 6 channel. Are you using Perl 6 or 5? 11:39
sci3ntist: Perl 6 is not ready for production use yet.
sci3ntist i'm using perl 5.8
Juerd sci3ntist: Then visit #perl instead
We're building the new Perl here :)
sci3ntist i'm in there 11:40
but they didnt answer me
masak ask again
Juerd Then wait and ask again :)
sci3ntist they didnt know how to solveit
Juerd sci3ntist: There'll be different "they"s another time. Try again in a few hours.
sci3ntist: Or try Perl channels on other IRC networks. Or www.perlmonks.org
sci3ntist i'll try ,anyway thanks 11:41
11:54 autark_ joined 11:57 fglock joined 12:13 lumi joined 12:19 mako132_ joined 12:23 ofer joined 12:28 nipotan is now known as nipotaway 12:53 ofer joined
kolibrie Juerd: sorry it took me so long to get my Dvorak testimonial to you - sent now 13:01
Juerd kolibrie: Thanks. You're the second to send it :( 13:03
13:16 sunnavy_ joined 13:27 TJCRI joined 13:54 chylli joined
pasteling "evalbot_r16970" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26212 13:56
pugs_svnbot r16971 | moritz++ | trunk:
: [perlhint]
: * added Lookup.pm
: * added tests, run with prove -Ilib5 t/*.t
: * added more data
: * perlhint5.pl now has very basic functionality
diff: dev.pugscode.org/changeset/16971
lambdabot Title: Changeset 16971 - Pugs - Trac 13:57
[particle] Juerd: sent testimonial. kolibrie: thanks for the reminder! 13:59
kolibrie [particle]: :) 14:04
14:09 dwave joined 14:10 franck___ joined 14:17 stevan__ joined 14:19 sunnavy_ is now known as sunnavy 14:23 amnesiac joined 14:26 theorb joined, cjeris joined 14:45 b2gills joined 14:55 daxim joined 14:56 Eidolos joined 15:02 unobe is now known as unobe_away
beppu masak: thanks for the coro example. 15:02
15:06 fglock joined 15:14 chylli left
masak beppu: no problem. could I also recomment the t/ directory? it's a good read 15:20
moritz's logs only go back to earlier this year 15:22
are there older logs somewhere else?
15:22 ashleyb joined
fglock masak: colabti.de/irclogger//irclogger_logs/perl6 15:23
lambdabot fglock: You have 1 new message. '/msg lambdabot @messages' to read it.
Title: #perl6 irc log
masak fglock: thx -- those were the ones I was looking for 15:24
Google does not do an excellent job at turning them up
fglock np
lunch & 15:25
diakopter pugscode.org/ | nopaste: sial.org/pbot/perl6 | ?eval [~] <m oo se> | We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/ 15:27
diakopter moritz: How about a link from your logs to the colabti logs? 15:27
masak what is the relation, really, between the colabti logs and moritz's logs? 15:28
why are there two sets of logs?
(not that I mind) 15:29
mncharity 15:31
diakopter well, moritz's logs were set up (I think) because the colabti logger was down for a time 15:32
and it's become a fun project in and of itself (to tweak the output)
beppu masak: I'll take a look in t/ for more coro examples. thank you.
mncharity @tell fglock I made some notes after you left yesterday. Just let me know if/when you would like them. 15:33
lambdabot Consider it noted.
diakopter moritz: even better: ;) a tool to scrape/import the prior logs from colabti to yours ;)
15:35 afraid joined, afraid left
beppu ?eval async { 5 } 15:43
15:43 evalbot_r16970 is now known as evalbot_r16971
evalbot_r16971 undef 15:43
beppu ?eval async { say 5 }
evalbot_r16971 OUTPUT[5ā¤] undef
masak beppu: I meant syntax examples in general in t/, not just for coro 15:46
that's the way I learn new perl6 syntax nowadays :)
that, and the synopses
15:54 penk joined 16:10 pmurias joined
pmurias hi 16:10
fglock: ping
16:10 toshinori joined
diakopter moritz: yet another wishlist item for your logger: a way to get/set a cookie of browser's TZ? or even javascript conversion of all the hours based on what TZ the browser reports? :) 16:12
16:26 b2gills left 16:27 kolibrie joined, Juerd joined, audreyt joined 16:28 orafu joined, cernd joined, franck__ joined 16:29 Yaakov joined 16:30 charsbar_ joined 16:31 toshinori joined, stevan__ joined, rho joined, yves joined, edenc joined, rintaro joined, orafu joined, audreyt_ joined, jql joined, arguile_ joined, OuLouFu joined, dactyli joined
pmurias fglock: i'm toying around with using SDL from kp6 16:31
dactyli oO
moritz diakopter: number 1 and two are noted on my virtual TODO list, the TZ thing sounds like much work (and not very much use)
pmurias dactyli: Oo
should i implement the use Foo:from<perl5> syntax? 16:32
fglock pmurias: hi
lambdabot fglock: You have 2 new messages. '/msg lambdabot @messages' to read them.
16:32 macli joined
fglock hmm 16:33
I think this belongs to kp6
mp6 is not required to support perl 5
pmurias fglock: hi
fglock mncharity: pong - I'd sure like to read your notes 16:34
16:34 cernd joined, bonesss joined
pmurias i want to add it to kp6; i'm using sdl from a script written in kp6 16:35
fglock pmurias: cool
there is some special compilation magic related to 'use'
since it's executed (linked) at compile-time 16:36
but use-perl5 can probably do without that 16:37
and it would probably execute at both compile- and run-time
pmurias what i need the from bit is emit special code to wrap the module into a kp6 object 16:39
fglock pmurias: I think the most likely way to implement it right now is to write a bridge module by hand
that is, a perl5 module with a MOP-compatible api 16:40
one possible way to do this,
pmurias i'll checkin what i already have
fglock is to write a kp6 module and compile it to perl5; then add your custom code to the compile code by hand 16:41
s/compile/compiled code
ok
pugs_svnbot r16972 | pmurias++ | trunk:
: kp6: a wrapper for perl5 modules
diff: dev.pugscode.org/changeset/16972
lambdabot Title: Changeset 16972 - Pugs - Trac
pmurias fglock: doing it now, but having kp6 emit it is more usefull & fun 16:42
fglock looks good
how about implementing eval( '...', lang:<perl5> ) 16:43
I think this is the specced way to mix code; another option is { use v5; ... } but I'm not sure if this is specced 16:44
6-in-6 will end up requiring a perl5 parser anyway - but it's not a priority right now 16:46
pmurias a p5 parse would be a bit difficult IMHO :) 16:47
fglock it would be fun
pmurias it would be great to have perl5 quasiquoting 16:48
added the :from<perl5> to the grammar
can i extend the ast?
and is using adding a native native method to Value classes okay or should it be _native? 16:50
fglock the places to modify are Traverse.pm and Visitor::EmitPerl5, I think
not sure about 'native'
hmm
there should be no 'native' class, because it's not OO 16:51
I think
pmurias p5landish would be more accurate 16:52
fglock yes, let's look if there is something about that in the synopsis 16:53
pmurias the low speed of mp6 allows one to think/look around a lot ;) 16:55
16:55 b_jonas joined
fglock I think the way to go is to box the data to p6land: p5str-to-Str, p5num-toNum, ... - but this takes a lot of effort 16:56
re low speed: the kp6 regex engine might be somewhat faster, but the runtime will be even slower 16:57
mncharity fglock: not really notes to read. more "I haven't forgotten everything since yesterday". ;) so whenever you want comments... 17:00
fglock pmurias: I'm about to start writing the plan for kp6 phase 4 (improve the runtime) - would you like to discuss that?
mncharity: sure, is now ok?
mncharity sure, one sec 17:01
pmurias fglock: sure
fglock pmurias: I think the most important target right now is to add support for multis 17:02
this requires a working type system; finishing the Code, Signature, Capture classes
and also Array and Hash, which are barely started 17:03
mncharity ok, I guess three main thoughts,
I like the log file system idea. to get access to match objects during matching, you'll need something like
a match object, and a log delta, both local()ized, and when you access the match object it does a self deepcopy, and update from log, and zeroing of log. 17:06
fglock pmurias: it would be great if it would be possible to write these classes in mp6
mncharity which I think is what you said earlier?
that was first thought
fglock mncharity: one sec, I'll point you to the code
mncharity second is re blocks, third re 5.12 and rentrancy
fglock Visitor::EmitPerl5Regex line 154 - implements getting the $/ value in the current regex scope 17:08
mncharity blocks is basically "blocks are hell". I don't see any way to do them without doing PPI (is that right?) code analysis.
looking...
fglock hm? blocks are compiled from p6 code - why PPI? 17:09
pasteling "evalbot_r16972" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26219 17:10
pugs_svnbot r16973 | pmurias++ | trunk:
: kp6:
: native => p5landish
: TODO - prefix:<++> implemented
: :from<perl5> parsed & recorded in ast
diff: dev.pugscode.org/changeset/16973
lambdabot Title: Changeset 16973 - Pugs - Trac
fglock re scope - Match::from_global_data automatically detects unmatched 'create' ops, and then return the current scope in the top of the result stack 17:11
mncharity re 154, so how do you handle writing to the Match object? 17:12
re 'blocks are compiled from p6 code', ah, that makes it simpler. I was thinking p5 code. but if you already have the code ast, then doing things like "return" are straightforward. 17:13
fglock it's currently a copy - the plan is to implement this either with overloading, or pointers, or ... 17:14
'return' is working :)
mncharity :)
[particle] can someone point me to where string increment is implemented? 17:15
fglock [particle]: in which Perl 6? :) 17:16
mncharity re 'implement this', right, the scheme I mentioned was an "or...". Though it has the downside that if anyone stashes a pointer into the Match struct, it may dark-magically no longer be part of the match struct shortly thereafter. :(
[particle] any one where it's working :)
preferably an implementation where the underlying vm does not have a builtin string increment 17:17
fglock [particle]: there was a working impl in the old pugs-p5 backend; not sure where is it stored now - looking
mncharity: ideally, the match changes should be recorded as ops, which is more or less what you said, I guess 17:19
because this make it backtrackable
mncharity well, you could either make the match changes recorded as ops, or modify the concept so match state is a composite... no, that won't work. 17:22
it's going to take a rather magical Match proxy to deal with code, during matching, being able to stash parts of the Match tree and modify them later. or maybe not. I guess it might just be a generic tree proxying problem. 17:23
fglock [particle]: in svn.openfoundry.org/pugs/misc/old_p...6/Value.pm there is a Perl6::Value::Str implementation of 'decrement' - you can have an idea on how to do increment
lambdabot tinyurl.com/2custl
17:24 jedai joined
pmurias fglock: you want to implement those (Array,Hash..) classes in non-oo kp6? 17:24
fglock mncharity: the way I'm taking care of this during compilation is that all STOREs are recorded; I could use the same thing for runtime, even if it is a bit costly 17:25
[particle] fglock++
mncharity ok, third thought was re rentrancy. is this intended to be 5.12+ only? ie, dependent on the underlying p5 engine being reentrant? 17:26
avar you mean 5.10?
fglock pmurias: yes, if possible - because it's good for portability
mncharity re 5.10, um, whatever was in blead a month or few ago.
fglock mncharity: it seems to work right now; the prototype implements a recursive call; 17:27
I wonder if it works because I'm using regex-eval, which is not really eval'ing the same code, but compiling new code
works in 5.8 17:28
mncharity prior to 5.mumble, if an regexp has a code block which tries to use a regexp, that second regexp doesn't work.
it doesn't matter what the regexp's were, simply that call stack hits two of them. 17:29
rephrased, for clarity, prior to the p5 regexp being modified to be rentrant, if your call stack has two regexp calls, the second one isn't working. Hmm, I don't recall at the moment whether the first one gets smashed too. I think not, but I'm fuzzy. 17:31
pugs_svnbot r16974 | pmurias++ | trunk:
: kp6:
: emiter recognises "use" of perl5 modules
: a basic sdl example works!
diff: dev.pugscode.org/changeset/16974
lambdabot Title: Changeset 16974 - Pugs - Trac
fglock see svn.pugscode.org/pugs/v6/v6-KindaPe...subrule.pl line 272; works under 5.8
lambdabot tinyurl.com/29lv29
fglock well, it seems to work :)
mncharity very doesn't. :) looking...
fglock it's tested at line #348 17:32
[particle] pmurias: that's the use syntax now... trailing :from<foo> without whitespace?
mncharity re 272, that's just inlining a regexp into a regexp. that works, though obviously if the regex is self recursive, that has problems. but if instead of eval(), that said eval() if "a" =~ /a/ (to use a trivial example of the code block _itself trying to use a regex_), then it wouldnt work. 17:35
fglock pmurias++ 17:36
pmurias [particle]:S11
[particle] pmurias: just found it, thanks. and nice work! pmurias++
ah, there's a syntax bug in the spec 17:37
pmurias fglock: how/where should Wrap.pm be named/placed?
mncharity let's see, what's the underlying issue...
fglock mncharity: yes, that may be a problem - I need to test it 17:38
pmurias: probably under Runtime/Perl5/ ?
[particle] @tell TimToady S11:379 s/ver/:ver/ 17:39
fglock pmurias: that's worth a blog :)
17:40 charsba__ joined
lambdabot Consider it noted. 17:42
mncharity the objective is to create a regex from other regexen. for some cases, you can simply inline the regexen. and for efficiency, when you can, you seem likely to want to. for other cases, you can't, but you could get away with python/ruby/5.mumble-like named regex fragments. I'm not sure this is worth bothering with until later. Because in the general case,
17:42 elmex joined
[particle] why is lambdabot so slow? 17:42
fglock mncharity: re when you can't - see line #240 - using a method as a matcher 17:44
mncharity you actually have to handle recursion. if you are supporting p5, the only way to get p5 local() in embedded code blocks to work properly, is to actually recurse, since there is no other way to fiddle with local(). if you are just doing p6, then you have control, so you can use a stack machine, or whatever. then the question is whether you're targeting
fglock is supposed to cover the corner cases, if it doesn't kill perl
17:45 ruoso joined
pmurias fglock: got to get it to support a bit more p5 land interaction first ;) 17:46
mncharity a version of p5 which has recursive regexp support or not. if yes, life is simple, and you should be able to just inline code blocks, and have the right thing happen. But if not, if you want it to work on 5.8, then when you encounter a subrule or code block, you can't use a real regular expression anywhere above it in the regexp ast tree. Ie, given (b*<foo>)+
the b* can be a native re, but the () and + have to be implemented by you. 17:47
17:47 barney joined
mncharity actually, b* can only be an native re with some hoop jumping (ie, to backtrack, you have to hand the native b* re a shortened version of the target string). which may well not be worth it for simple b*, but might be for more complex expressions. 17:49
fglock mncharity: hmm - it seems to work; I wonder if you'd try it out (use svn.pugscode.org/pugs/v6/v6-KindaPe...en-test.pl as a start point)
lambdabot tinyurl.com/2hf4xk
fglock I'll try it here
mncharity oh, I guess there was also a fourth set of thoughts, re inheritance and methods access. but looking... 17:50
fglock inheritance is tested in the prototype, but it is not tested in kp6 yet 17:51
17:51 charsbar joined
mncharity re token-test, there don't seem to be any instances of non-top-level subrules. 17:52
err, non-(top-level and non-recursive) subrules
token-test is an example of something which can be inlined. 17:53
let's see, if $<rep> := (.) is changed to $<rep> := (<word>) ... 17:55
pasteling "fglock" at 201.54.129.80 pasted "backtracking p5rx+ subrule" (25 lines, 360B) at sial.org/pbot/26224
fglock mncharity: it worked 17:56
mncharity and $/<word> got set? 17:57
fglock sure
mncharity ponders...
fglock do you have a kp6 checkout? 17:58
mncharity ok, so what's happening is that the log approach is letting you flatten that into a single inline re. so, a case where this won't be possible is... 18:01
re kp6, yes, but I haven't run it, so rather than fighting/learning, I was waiting for there to be a pressing need...
...won't be possible is... well, either recursion token word { q <word>* } or a code block anywhere which tries to itself do a re match. 18:05
18:05 Psyche^ joined, Psyche^ is now known as Patterner 18:06 fglock__ joined
mncharity the first can't be inlined (I don't think), but does work with a named named sub-expressions extension (like (?P<word>)q(?<word>)*) or somesuch, I forget the syntax). 18:06
pmurias fglock: where is Scalar _dispatch defined? 18:07
mncharity but the second can't be done unless you either have a regexp engine capable of reentrancy, or you can inline not just the pattern, but the surrounding code as well. (ie, p5 op codes, which then looks a lot like having a regexp engine capable of reentrancy;) 18:08
or you handle the recursion yourself 18:09
rather than trying to embed it in an re
avar: yes, it looks like I meant 5.10 :) thanks 18:13
my suggestion would be to keep it simple, and target 5.9/5.10. 18:14
18:15 rintaro joined
mncharity then implementing a subrule call is as simple as (?{ high_Im_a_subrule_call() }) (assuming the match context is passed in locals) 18:15
s/high/hi/
18:16 edenc joined
mncharity questions, comments, thoughts? 18:17
18:19 b_jonas joined, toshinori joined, stevan__ joined, rho joined, yves joined, audreyt_ joined, jql joined, arguile_ joined 18:20 audreyt_ joined
fglock__ rehi 18:20
pmurias rehello
fglock__ pmurias: one sec 18:21
18:21 thoughtpolice joined
mncharity is unsure what flgock_* did/didn't see... 18:22
fglock__ MOP.pm line 394; it points to the implementation in line 78
mncharity: backlogging 18:23
18:25 devogon joined
fglock__ about recursion - this is easy to test; about a re inside a block, I'll try that too 18:26
pmurias fglock: i want to create an kp6 wrapper object around a p5 one 18:27
fglock__ re using kp6 - there are no crazy dependencies, it should be straightforward
pmurias: adding a dispatch key should do it, I think 18:28
pmurias i get a no such method in Class error 18:31
pugs_svnbot r16975 | pmurias++ | trunk:
: kp6: Wrap renamed
: a problem with wraping objects
diff: dev.pugscode.org/changeset/16975
lambdabot Title: Changeset 16975 - Pugs - Trac
18:31 iblechbot joined
pmurias works fine if i FETCH the value, so i think Scalar does some filtering of it's own 18:31
mncharity fglock__: ok. inheritance now, or inheritance later? 18:32
pmurias will take a break and listen to bbc a bit... 18:33
fglock__ mncharity: I added the recursive example to the repo - it just works
pmurias: hmm
mncharity: sorry? 18:34
the way inheritance works right now is by eval'ing the regex variable in the grammar's namespace; this is supposed to be refined later 18:35
18:36 b_jonas joined
fglock__ hmm - fixing my commit 18:37
18:37 jisom joined
fglock__ mncharity: svn.pugscode.org/pugs/v6/v6-KindaPe...recurse.pl 18:38
pugs_svnbot r16976 | fglock++ | trunk:
: [kp6] added examples/token-test-recurse.pl
diff: dev.pugscode.org/changeset/16976
lambdabot Title: Changeset 16976 - Pugs - Trac
tinyurl.com/2xfp2e
18:39 rho joined 18:41 barney joined
mncharity re recursive works, that's rather unexpected. how about word { a<word>*b } ? 18:41
fglock__ trying 18:42
mncharity re inheritance, refined later, ok :)
pugs_svnbot r16977 | fglock++ | trunk:
: [kp6] v6/docs/p6regex-on-p5regex.txt TODO updated with mncharity++ questions
diff: dev.pugscode.org/changeset/16977
lambdabot Title: Changeset 16977 - Pugs - Trac
18:43 weinig joined
fglock__ mncharity: no problem !!! 18:44
mncharity perl -we 'sub f{ print "hi$1" if "b" =~ /(b)/ } "=~ /(a)(?{f()})/; print $1' => Out of memory! 18:47
err, that's not right,
18:48 jhorwitz joined
mncharity perl -we 'sub f{ print "hi$1" if "b" =~ /(b)/ } "a" =~ /(a)(?{f()})/; print $1' => Out of memory! 18:48
v5.8.8
audreyt "hiba"
v5.9.5
fglock__ kp6's version: qr/(?{ local $GLOBAL::_M = [ $GLOBAL::_M, 'create', pos(), \$_ ]; $GLOBAL::_M2 = $GLOBAL::_M; })(?:(?:a(?:(??{ eval '$_rule_word' })(?{ local $GLOBAL::_M = [ $GLOBAL::_M, "named_capture_to_array", "word" ]; }))*b))(?{ local $GLOBAL::_M = [ $GLOBAL::_M, 'to', pos() ]; $GLOBAL::_M2 = $GLOBAL::_M; })/x
:)
audreyt !!!
audreyt greetings 18:49
mncharity yes. :) thus my suggestion to target 5.9/5.10. It has a reentrant regex engine. or at least claims to.
fglock__ missed you so much :)
audreyt (no, my liver is not yet fully allowing me to hack. but finally some irc'ing.)
mncharity: yeah, 5.10 has built-in grammar support and recdescent et al
would make performance competitive, or at least comparable
mncharity grammar support ?!? oh my 18:50
18:50 weinig joined
mncharity audreyt: you realize this is the second time I haven't been around for months, stop by, and you're here. as far as I can tell, you're here all the time :) 18:51
audreyt /(?<NAME>(?&NAME_PAT))(?<ADDR>(?&ADDRESS_PAT))
(?(DEFINE)
(?<NAME_PAT>....)
(?<ADRESS_PAT>....)
)/x
supported natively, apparently fast enough, all thanks to dmq++
(the ?<> is named capture.)
mncharity ah, right. I like the DEFINE concept. must nicer than "it must preceed". named capture too?? nifty. 18:52
audreyt possesive quantifiers, regex verbs like , (*PRUNE), (*MARK), (*SKIP), (*COMMIT)(*FAIL) and (*ACCEPT) 18:53
and more :)
avar yeah, a lot of that stuff could make emitting p5rx a lot faster:)
plus maybe a re::engine that uses perl's engine with another interface that would be easier for an emitter to use
mncharity has anyone gone through and looked to see what p6 regexp concepts -don't- easily map to the new stuff? to create a verb shopping list? :) 18:54
fglock__ mncharity: I couldn't find out a way to create a Match object from that, yet 18:55
avar named captures are bound to numbered ones in p5
(for one) 18:56
mncharity bound how? ie, since the p6 is being emitted, could one do a simple renaming/numbering to keep them clear of each other? 18:57
avar $ perl5.9.5 -E '"abc" =~ /(?<bee>b)/; say $1' 18:58
b
pugs_svnbot r16978 | audreyt++ | trunk:
: * Makefile.PL: Do not remove the trailing .1 from 6.6.1
: for $ghc_version, so we can distinguish 6.6.1 from 6.6.0.
diff: dev.pugscode.org/changeset/16978
lambdabot Title: Changeset 16978 - Pugs - Trac
fglock__ mncharity: how about (.)* - capture to array
18:58 weinig_ joined
mncharity but big picture, it looks like it would be much easier to target 5.9/5.10 than 5.8. then the question becomes, could 5.9 be tweaked to make it even easier. 18:58
fglock__ hmm
audreyt renaming seems feasible if you emit (?<#0>) for $0 18:59
avar <#0> is invalid syntax, it woulh have to be ?<__0> or something 19:00
audreyt sure...
fglock__ there are several things that are useful for 6rx-on-5rx, but I think we still need to keep the Match building code 19:01
mncharity re capture to array, yeah, something like that might be the gotcha. some subtle difference in how repetition or somesuch was handled, which forced one to bail out entirely. or at least for some class of regexp's, which still means paying the cost of doing a fullish implementation. :/
audreyt ...but at least such an enging can be swapped in lexically... 19:02
pugs_svnbot r16979 | audreyt++ | trunk:
: * inc/Module/Install/Pugs.pm: Since src/ now contains
: no .h and .c files anymore, remove the -I and -i lines.
diff: dev.pugscode.org/changeset/16979
lambdabot Title: Changeset 16979 - Pugs - Trac
fglock__ it may make sense to have 2 implementations; PCR does that
mncharity re Match, I wonder if the p5 5.9 folks could be talked into having a Match object... :) 19:03
avar you can make one yoursel
[particle] mncharity: 5.10RC1 is out, so don't expect any 5.9 changes now
avar search.cpan.org/~avar/re-engine-PCR..._01/PCR.pm
lambdabot Title: re::engine::PCR - Pugs::Compiler::Rule regex engine - search.cpan.org
[particle] ...they'd have to be 5.11. 19:04
mncharity re rc1, ah, ok, ah well, but nifty it's coming out
avar all this stuff can be changed post 5.10 since you can plug in another rx engine
including a modified 5.10 engine
[particle] mncharity: in case you didn't know, perl devel versions are 5.<odd> and release versions are 5.<even>
mncharity yup, tnx :)
so yes, we would still need to do the Match object. 19:05
avar why do you need a match object? 19:06
$+[] and $-[] are ->to and ->from at least
fglock__ avar: for p6-in-p5
avar Yes, but whan specific features
mncharity p6 has one :)
fglock__ avar: for representing submatch trees, for example 19:07
mncharity separate named and numbered captures, which are themselves ... what fglock__ said
$2 is a Match object in p6, which may itself contain other Match objects. rather than being a string. (which I assume in 5.10 it still is, but I wouldn't bet on anything at this point:) 19:08
fglock__ audreyt: I wonder if you'd have some time to discuss p6-in-p6 - maybe at Vienna, or before? 19:10
audreyt at Vienna perhaps 19:11
I'm still constrained by incessant sleep
e.g., now
audreyt fades back to background
fglock__ sings a lullaby 19:12
mncharity fglock__: so I guess my one-liner is suggesting you target 5.10, and use explicit (?{ subrule() }) recursion rather than trying to inline everything. 19:13
audreyt: fare and be well...
fglock__ mncharity: yes, it's a possible optimization
19:14 lichtkind joined
fglock__ mncharity: however, I'm more likely to follow on to the next project phase, and optimize later 19:15
gotta finish this project :) 19:16
mncharity err, I think of inlining as the special case optimization for speed. which won't work in general. so correctness is at issue, rather optimization
but I gather I haven't persuaded you. I do suggest checking at least token word { a<word>*b } 19:17
fglock__ hmm - I tried that, it worked
mncharity hmm.
fglock__ the code I pasted above 19:18
mncharity looking...
fglock__ should I commit the code?
mncharity i don't see the paste... 19:20
pugs_svnbot r16980 | fglock++ | trunk:
: [kp6] examples/token-test-recurse2.pl - a recursive regex example
diff: dev.pugscode.org/changeset/16980
lambdabot Title: Changeset 16980 - Pugs - Trac
fglock__ please do: $ perl kp6-perl5.pl --perl5rx < examples/token-test-recurse2.pl | perl -I lib5regex -I lib
from directory: v6/v6-KindaPerl6/ 19:21
mncharity doing...
19:22 ashleyb joined
mncharity It also seems to match aaababb 19:24
which is a no no
fglock__ testing 19:25
heh - look again 19:26
a(a(ab)(ab)b)
mncharity lol 19:27
fglock__ I should say mncharity++ # looking very hard for bugs 19:29
mncharity :) 19:30
mncharity tries to read compiler output...
fglock__ use perltidy - though it doesn't do regexes 19:31
you'll see several (?: ... ) that could be optimized out 19:32
diakopter sweet.... 19:33
mncharity so the key idea is that it is not in fact inlining, or at least not statically, but instead doing just-in-time inclusion via the eval()...
pugs_svnbot r16981 | diakopter++ | trunk:
: please disregard this test.
diff: dev.pugscode.org/changeset/16981
lambdabot Title: Changeset 16981 - Pugs - Trac
fglock__ mncharity: yes; perl5.9.x would be able to do real inlining 19:34
pugs_svnbot r16982 | diakopter++ | trunk:
: deleting my test file. :)
diff: dev.pugscode.org/changeset/16982
lambdabot Title: Changeset 16982 - Pugs - Trac
fglock__ and true recursive calls 19:35
diakopter announcement to #perl6: we have been donated a license to use Repos.se, a browser-interface to subversion. It is installed transparently at svn.pugscode.org/pugs/ You can edit, rename, copy, move, delete, replace(upload), and view the changelogs of any file or a directory (and all its contents, deeply), all through a web browser. Permissions should work as they do via DAV. 19:37
mncharity returns from pacing in thought 19:39
very nifty.
fglock__++
fglock__ mncharity: :) 19:40
diakopter: it looks very nice
pugs_svnbot r16983 | avar++ | trunk: 19:41
: woohoo
diff: dev.pugscode.org/changeset/16983
lambdabot Title: Changeset 16983 - Pugs - Trac
avar so, uh, it works 19:42
diakopter: what is it written in? 19:43
pugs_svnbot r16984 | avar++ | trunk:
: deleting test directory
diff: dev.pugscode.org/changeset/16984
lambdabot Title: Changeset 16984 - Pugs - Trac
diakopter obfuscated/encrypted php
mncharity and even if it turns out one can't for some reason throw exceptions between blocks of a regexp, one could fake it. and you can "succeed" even if the actual regexp doesn't, so you can assign to the 'string being matched', even if p5 doesn't let you. and, at least on 5.10, you can do continuations down through subrule if you need to.
pasteling "evalbot_r16982" at 193.200.132.135 pasted "Pugs build failure" (6 lines, 278B) at sial.org/pbot/26229 19:44
19:44 bonesss is now known as bones`OFF
mncharity that's sweet 19:44
fglock__ mncharity: that's too much info in one sentence - I'm still reading :) 19:45
mncharity lol
fglock__ re succeed - yes, you just need to close the lexical scope in the global operation stack 19:47
re throw exceptions, I think it's possible, but I need to write down an example to test 19:49
re continuations, not sure what you mean? 19:50
ashleyb audreyt: are you at TimToady's this month? (I recall that you were planning to do that)
diakopter ashleyb: audreyt didn't make that trip, nor to oscon this year, it seems. :( still, !health hinders. 19:52
19:52 barney joined, Aankhen`` joined
diakopter Aankhen``: backlog! 19:52
Aankhen`` Sir yes sir! 19:53
19:54 Shangrila joined
Shangrila hi 19:54
mncharity re continuations, just wondering how one might deal with pure code, non-re subrules, if it turned out one needed to for some reason. 19:55
s/how/whether one coud
19:55 justatheory joined
diakopter I'll add the various plugins later... gtg4now 19:56
mncharity so a lot hinges on the regexp engine actually behaving correctly, in the face of this unusual exercise it's being given. historically that's been problematic. eh, but that's what 5.10.1 is for. 19:57
pmurias Shangila: hi
mncharity oh, silly me, you are emitting the code blocks which would hypothetically write to the Match, which makes the Match side of things much simpler. 19:59
fglock__ mncharity: there is a non-re subrule test in the prototype
yes
mncharity fglock__: ok, new one-liner: looks good ;) 20:00
spinclad .oO{ o-tori tan westu hƔl! }Oo. 20:01
Aankhen`` diakopter: ā€¦was there a reason why you wanted me to backlog? 20:02
mncharity so the leverage comes from having control of the code that's embedded (by virtue of emitting from p6, rather than accepting arbitrary p5), and doing infrastructure with (?{}) blocks, rather than external code, which lets you use (??{}) for inclusion. and 5.10 is available for when a code block needs to itself use a regexp. 20:06
fglock__ "emitting the code blocks which would hypothetically write to the Match" - I wonder how to explain that to an audience 20:07
mncharity it's just a variant on "having control of the code that's embedded (by virtue of emitting from p6, rather than accepting arbitrary p5)". 20:08
so you don't have to... oh, wait.
well, it's not as good as I first thought. you can when the code block _isn't_ writing to $/, and can tell when the top-level of the block itself _is_, but if the block calls a sub(), handing it the Match, you may have to be pessimistic, perhaps do a diff after the call, since you can't be sure whether the Match has been written to or not. 20:10
s/you can when/you can tell when 20:11
fglock__ the plan (so far) is to log the STORE operations in the MOP 20:12
mncharity in the arbitrary p5 case, unless you do code analysis, if there's a user specified code block, you have to assume they are reading and writing the match tree, and have everything set up for that.
fglock__ I need this anyway for BEGIN block post-processing 20:13
mncharity ah, ok
ok, back to hacking. that was fun. thanks! :D 20:15
fglock__ re the sub() case, $/ is lexical - only OUTER::<$/> can change the match result (not properly implemented, yet)
mncharity: thank you :)
20:20 sk joined 20:21 Blwood joined
pugs_svnbot r16985 | Aankhen++ | trunk: 20:21
: * [irclog] more abbrs.
diff: dev.pugscode.org/changeset/16985
pmurias fglock__: any clues why the Scalar inteferes with dispatch to my wrapper and how to fix it? 20:25
fglock__ pmurias: (probably) Scalar is a container; you need to dispatch to the value 20:26
20:28 weinig joined 20:30 drupek12 joined
fglock__ pmurias: if that's the problem, the way I found to fix it is to always call FETCH, since FETCH on a value is a no-op 20:32
pmurias it works that way 20:33
but what interest me what exactly is Scalars->{_dispatch} 20:34
you pointed me to Scalars->{_dispatch_VAR} before
fglock__ Scalar->{_dispatch} implements the Scalar own methods, like STORE 20:36
pmurias and _dispatch_VAR?
fglock__ Scalar->{_dispatch_VAR} implements dispatching on the value's methods
hmm - oops 20:37
or is it the other way
pmurias the emitted code does it the other way around 20:38
fglock__ _dispatch_VAR should implement VAR($x) 20:39
sorry - too many implementation details :P 20:40
$x = 42 should compile to VAR($x).STORE( 42 ) 20:41
pmurias in terms of emitted code into $x->{_dispatch_Var}($x,...)? 20:47
with ... being "STORE",42 20:48
20:48 Schwern joined
pmurias what i'm thinking of shouldn't VAR($x).STORE compile to $x->{_dispatch} 20:50
($x,"STORE") 20:51
20:52 fglock joined
pmurias and $x.method to ($x->{_dispatch}($x,"FETCH"))->{_dispatch}($x->{_dispatch}($x,"FETCH"),"method") 20:53
fglock yes, unless $x->{_roles}{auto_deref}
which calls FETCH automatically 20:54
pmurias i see 20:55
fglock hmm - no, $x.method is: $x->{_dispatch}( $x, 'method', ) 20:57
pmurias fglock: squashed the bug :) :) :) 20:58
pugs_svnbot r16986 | pmurias++ | trunk: 20:59
: kp6: squashed the containers dispatcher used on value bug
diff: dev.pugscode.org/changeset/16986
lambdabot Title: Changeset 16986 - Pugs - Trac
fglock cool - sorry for the confusion, I wrote this code long ago ( > 1 month ) :) 21:00
oh, it was modifying self 21:01
nice catch
21:05 [particle] joined 21:13 fglock___ joined 21:15 rindolf joined
rindolf Hi all! 21:16
Can anyone help me with a Prolog problem? erxz.com/pb/3527
lambdabot Title: freenode_web paste from "rindolf" at 71.6.194.243...
21:21 fglock___ left
pmurias good night 21:22
Aankhen`` rindolf: I'm pretty sure this was #perl6 last time I checked.
rindolf pmurias: night.
pmurias rindolf: got a prolog book at home 21:23
what do you want to achive?
rindolf pmurias: I'm trying to solve exercise 6.4 from here - www.coli.uni-saarland.de/~kris/lear....exercises 21:24
lambdabot Title: 6.3 Exercises, tinyurl.com/ysouw7
pmurias i'll look into it tommorow as it's too late here for thinking& 21:26
21:26 pmurias left 21:28 weinig joined 21:31 drupek12 joined 21:35 Schwern joined 22:03 weinig joined 22:04 marmic joined 22:21 weinig joined 22:22 unobe_away is now known as unobe 22:28 fglock joined 22:40 cjeris left 22:41 Blwood joined 23:01 thorat joined 23:04 Schwern joined 23:09 weinig joined 23:16 stevan_ joined 23:30 weinig joined 23:36 weinig joined 23:45 theorbtwo joined 23:47 Limbic_Region joined
Limbic_Region what's the svn incantation to see the diff of a file from the current version to the previous revision? 23:49
diakopter svn diff filename 23:50
Limbic_Region that's not working 23:51
by not working, I mean it returns immediately - no output
diakopter do you mean diff the previous revision to the current revision? 23:52
I thought you meant you'd modified the file
Limbic_Region no, I modified the last checkin
I just checked out and saw that it had been modified in the current version 23:53
so now I want to see the diff from the file in the HEAD with PREV
I figured it out 23:55
svn dff file -rPREV 23:56
unfortunately, it shows every single line as being changed
*shrug*
23:56 weinig joined, Blwood joined

Powered by Cro and the RakuĀ® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!