Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6, 2007-09-05

Perl 6 | Reference Documentation | Rakudo

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

All times shown according to UTC.

Time Nick Message
00:23 pugs_svnbot r17674 | lwall++ | [STD] Trying out new <+ws> notation; tracing improvements
00:23 pugs_svnbot diff: http://dev.pugscode.org/changeset/17674
00:23 lambdabot Title: Changeset 17674 - Pugs - Trac
00:38 Blwood joined #perl6
01:18 drupek1260 joined #perl6
02:00 sunnavy joined #perl6
02:17 REPLeffect joined #perl6
02:20 hcchien joined #perl6
02:26 dduncan joined #perl6
02:54 justatheory joined #perl6
03:08 Jamtech joined #perl6
03:27 thoughtpolice joined #perl6
03:41 TimToady_ joined #perl6
03:43 amnesiac joined #perl6
04:09 shachaf joined #perl6
05:13 masak joined #perl6
05:38 devogon joined #perl6
05:38 masak morning :)
05:40 agentzh morning, masak
05:40 masak morning, agentzh
05:40 Aankhen`` joined #perl6
05:45 Ashizawa joined #perl6
05:47 agentzh left #perl6
05:50 lambdabot joined #perl6
05:57 REPLeffect_ joined #perl6
05:57 agentzh joined #perl6
05:58 lumi joined #perl6
06:01 moritz_ good morning *
06:01 masak morning, moritz_
06:36 weinig joined #perl6
06:36 jisom joined #perl6
06:36 weinig joined #perl6
06:51 elmex joined #perl6
07:25 iblechbot joined #perl6
07:27 dduncan left #perl6
07:37 marmic joined #perl6
07:44 masak joined #perl6
07:45 franck___ joined #perl6
07:56 amnesiac joined #perl6
07:56 masak can anyone tell me a little of what the plans for 6.28.0 are?
07:58 moritz_ http://dev.pugscode.org/roadmap says "objects"
07:58 lambdabot Title: Roadmap - Pugs - Trac
07:58 masak right
07:58 moritz_ which would be mostly audreyt++'s MO branch
07:58 masak ah
07:59 masak tell me about that
07:59 masak it's something local on audreyt++'s computer?
07:59 moritz_ yes
08:00 moritz_ wait a second...
08:00 moritz_ http://moritz.faui2k3.org/irclog/out.pl?channel=perl6;date=2007-05-17#i_18608
08:01 lambdabot Title: IRC log for #perl6, 2007-05-17, http://tinyurl.com/29p8sk
08:01 michaelkohwj joined #perl6
08:01 masak oh, so she basically has the MO branch ready?
08:02 moritz_ seems like, yes
08:02 masak almost four months ago...
08:06 michaelkohwj hi all, is this the right place to express opinons about perl 6? (as oppossed to development discussion)
08:06 masak michaelkohwj: sure
08:07 moritz_ michaelkohwj: depends on the opinions ;)
08:07 wolverian michaelkohwj, only if they're correct opinions...
08:07 moritz_ "p6 is bloatware and dead" is something we hear too often to think it's still fun ;)
08:07 pugs_svnbot r17675 | sunnavy++ | more "my" tests
08:07 pugs_svnbot diff: http://dev.pugscode.org/changeset/17675
08:07 masak moritz_, wolverian: don't be silly, we need all the opinions, not just the correct ones
08:07 lambdabot Title: Changeset 17675 - Pugs - Trac
08:08 masak moritz_: we need people come here and tell us that so that we can change their minds before they tell too many other people
08:08 moritz_ masak: #perl6 eq brain washing? ;-)
08:08 wolverian masak, it was a joke, I just forgot the :)
08:08 masak for starters, how can it be bloatware if it's dead? :)
08:08 moritz_ michaelkohwj: go ahead ;)
08:08 masak wolverian: I gathered. but still
08:08 michaelkohwj ok
08:08 wolverian masak, so, please, opinions are great.
08:09 wolverian er
08:09 wolverian that was to michaelkohwj
08:09 michaelkohwj wolverian: thanks, i didn't want to go OT
08:09 masak moritz_: brainwashing in a very broad sense, yes. but so is everything else
08:09 moritz_ brainwashing++
08:10 moritz_ erm
08:10 moritz_ forget it ;)
08:10 michaelkohwj i think that sigils in perl 6 should remain the same as perl 5- yes, you guys want to change them to make perl more english-like, but-
08:10 michaelkohwj 1. this will make life hard for people porting code to perl 6
08:10 michaelkohwj 2. what's so difficult about perl 5 sigils? the sigil stands for "what type of variable do you want to get?"
08:11 masak michaelkohwj: what do you mean by "sigils"? $@%& ?
08:11 michaelkohwj i.e. @array1 = (0) x 10;
08:11 masak ok
08:11 michaelkohwj in per 6: to access element 0, use @array1[0]
08:11 masak yes
08:11 michaelkohwj in perl 5, $array1[0]
08:11 masak I, for one, like the change
08:11 michaelkohwj see, in perl 5, the $ says that i want a scalar
08:12 masak I understand the idea with the p5 way
08:12 moritz_ michaelkohwj: r1 no. 1: that can be automated for 99% of all cases
08:12 wolverian michaelkohwj, I think it's a bit too late to change that. did you find the motivation for the change in the relevant apocalypse? I'm sure TimToady_ will fill in if the reasons aren't documented yet very well.
08:12 masak I think the reasons are fairly well documented
08:12 moritz_ michaelkohwj: re no. 2: I don't think it's difficult, but beginners tend to have problems with error messages...
08:13 moritz_ they have a $foo[3], and the compiler complains about @foo not being declared, but they never even typed @foo!
08:13 masak michaelkohwj: given that the p6 way of treating sigils is already more powerful and consistent, I suggest you try it for a while if you haven't already
08:15 michaelkohwj TimToady_: sigil
08:15 michaelkohwj oops, sorry
08:15 masak :)
08:16 masak michaelkohwj: do you have Pugs installed?
08:16 moritz_ http://dev.perl.org/perl6/doc/design/apo/A02.html contains a bit of explanation
08:16 lambdabot Title: Apocalypse 2: Bits and Pieces - perl6:
08:16 moritz_ though just a little bit, really ;)
08:17 michaelkohwj masak: nope..
08:17 masak michaelkohwj: you should. it's a blast
08:17 moritz_ #pugs say "We have evalbots as well"
08:18 exp_evalbot OUTPUT[We have evalbots as well␤]
08:20 michaelkohwj guys, see the latest xkcd, its a coincidence that it's about perl 6 too
08:21 rgs is it ? :)
08:21 michaelkohwj rgs: if you intepret it as such
08:21 michaelkohwj 'that way
08:21 moritz_ I didn't recognize it
08:21 moritz_ and I'm pretty much Perl 6 aware, I think
08:22 michaelkohwj oh no
08:22 michaelkohwj oops
08:22 * michaelkohwj blushes
08:22 michaelkohwj ignore the above
08:23 moritz_ /ignore the above
08:28 masak :)
08:28 michaelkohwj i misread that
08:28 michaelkohwj never mind
08:29 masak I guess the existence of audreyt++'s MO branch means I shouldn't spend time hacking the OO parts of Pugs
08:29 moritz_ michaelkohwj: would like to help with Perl 6, by any chance?
08:29 moritz_ s/would/would you/
08:29 masak moritz_: aw, just give him the commit bit :)
08:30 michaelkohwj moritz_: how? i just started programming in perl a week ago
08:30 masak michaelkohwj: /msg your email address, and we'll send you the privs
08:30 michaelkohwj privs
08:30 michaelkohwj ?
08:31 masak access to the repository of pugs and other things p6
08:31 masak that's write access, by the way
08:31 michaelkohwj masak: not like i know how to do compiler/intepreter creation
08:32 masak some of us do, most don't
08:32 moritz_ michaelkohwj: I wrote a summary of what can be done here: http://www.perlmonks.org/?node_id=628746
08:32 lambdabot Title: Getting Involved with Perl 6
08:32 masak moritz_++
08:33 sunnavy moritz_++
08:34 michaelkohwj masak: alright then, ill try out the non-compiler-programming stuff
08:34 masak michaelkohwj: cool
08:36 sunnavy michaelkohwj: welcome, :-)
08:37 masak hm, http://commitbit.pugscode.org is not responding
08:38 michaelkohwj masak: yes
08:38 masak michaelkohwj: "yes", as in "yes, it is?"
08:39 masak not here, it isn't
08:39 michaelkohwj masak: yes, it is not responding
08:39 masak ah, ok :)
08:40 masak everyone, how to add a committer without commitbit.pugscode.org?
08:40 moritz_ I don't think that's possible atm
08:41 masak that's terrible news
08:41 moritz_ I already asked obra++ if there's a command line interface, but that doesn't seem to be the case
08:41 michaelkohwj masak: its ok
08:41 masak :(
08:41 masak michaelkohwj: I'm going to have to stall indefinitely on your commit bit, it seems
08:41 masak but I'll keep the email address you sent me
08:42 masak and since I'll be hanging here almost daily, I'll know when commitbit is back up
08:42 masak one of these days you'll get an email with commit rights :)
08:43 masak in the meantime, just hand patches to someone else (hardly an optimal solution, I know)
08:44 masak who hosts http://commitbit.pugscode.org? why is it down? how many people are working on getting it back up?
08:44 moritz_ masak: it's hosted on feather, and the situation will improve with the hardware upgrade Juerd++ promised recently
08:45 masak moritz_: eh.
08:45 masak why not now?
08:46 moritz_ don't know - http://moritz.faui2k3.org/irclog/out.pl?channel=perl6;date=2007-09-03#i_95978
08:47 lambdabot Title: IRC log for #perl6, 2007-09-03, http://tinyurl.com/23aag5
08:47 moritz_ hey, feather's load is < 4 ;-)
08:47 masak what's sslh?
08:48 moritz_ that's a daemon that automagically accepts https and ssh connections on the same port
08:49 moritz_ just to fool corporate firewalls :(
08:49 masak is that a req for commitbit?
08:49 moritz_ no
08:49 zamolxes joined #perl6
08:49 masak so, what's lacking that needs a hardware upgrade?
08:50 moritz_ scroll up a bit ;)
08:50 bsb left #perl6
08:52 masak ah, "<obra>: the problem is that feather is fucked up and has horrible performance for large file scans. (like what sqlite does all the time)"
08:52 masak so, we've basically locked in our meta-committer system in a database on a machine that is currently too slow to run it, without any alternative means of reaching/changing the data? :/
08:54 moritz_ masak: seems like, yes
08:54 masak that's amazingly short-sighted of us. not like us at all
08:55 masak I just wish there was some sort of plan of action
08:55 moritz_ masak: you mean besides waiting for hardware? ;_9
08:55 masak that's not a plan
08:56 masak is obra the one to ask about the commitbit system? there must be some way of accessing the data
08:56 masak I mean, the web server does it
08:58 ruoso joined #perl6
09:08 polettix joined #perl6
09:10 andara joined #perl6
09:13 michaelkohwj masak: that's ok, i'm still learning perl =)
09:13 masak michaelkohwj: I just think it's a bad state of affairs for things to be in, that's all
09:14 masak #perl6 really thrives on inviting people to collaborate
09:14 masak michaelkohwj: but by all means, start by downloading pugs
09:14 masak and taking a look at it
09:15 cognominal_ joined #perl6
09:15 michaelkohwj masak: ok
09:15 masak michaelkohwj: http://www.perlfoundation.org/perl6/index.cgi?download_perl_6
09:15 lambdabot Title: Download Perl 6 / Perl 6
09:18 michaelkohwj lambdabot: pugs
09:24 masak @google pugscode
09:24 lambdabot http://run.pugscode.org/
09:24 lambdabot Title: Run Perl 6 Now -- in your browser!
09:37 iblechbot joined #perl6
09:45 pmurias joined #perl6
09:46 pmurias hi
09:46 masak pmurias: hi
09:49 pugs_svnbot r17676 | ruoso++ | [kp6] Merging the changes from lib-kp6-mp6 to lib-kp6-kp6 -r17629:HEAD
09:49 pugs_svnbot diff: http://dev.pugscode.org/changeset/17676
09:51 michaelkohwj masak: cool! it has a shell-like command history recall with the arrow keys!
09:56 michaelkohwj ignore the dead link in that site at the bottom, i emailed the webmaster
09:58 moritz_ michaelkohwj: which dead link?
09:59 masak looks like the "Perl 6" link has an extra space in it
09:59 michaelkohwj yes, but thats a small issue
09:59 masak michaelkohwj: thanks for noticing and notifying the webmaster
10:00 masak michaelkohwj++
10:01 michaelkohwj np
10:02 moritz_ it seems to be fixed in the repo, only comments contain 'perl 6'
10:04 michaelkohwj gtg home cya
10:04 pugs_svnbot r17677 | moritz++ | [runpugs] fixed links (though only in comments)
10:04 pugs_svnbot diff: http://dev.pugscode.org/changeset/17677
10:04 lambdabot Title: Changeset 17677 - Pugs - Trac
10:05 masak wonder if the page needs to be updated manually somehow?
10:05 moritz_ maybe
10:07 masak in fact, the "development version" featured on the page is quite old, too: r17041
10:08 riffraff joined #perl6
10:11 riffraff ciao
10:11 masak y0
10:12 moritz_ masak: that's... uhm.. just 400 revisions behind
10:12 fglock joined #perl6
10:12 pmurias fglock: hi
10:13 moritz_ but there weren't many pugs changes in the mean time iirc
10:13 fglock hi
10:13 pmurias i'll add a :p5 option to tokens to enable porting of Runtime::Perl5::Grammar to p6
10:14 fglock pmurias: i'd rather fix it to work in p6 land
10:14 fglock Runtime::Perl5::* is meant as p6-land code
10:15 pmurias it could be moved
10:15 pmurias where should backend specific files in p6 live?
10:16 masak moritz_: that may be, but it's an indication that the page is not updated automatically
10:17 fglock pmurias: Runtime::Perl5::* is ok, but it needs to use the kp6 calling convention (tokens are using mp6 calling convention)
10:20 pmurias the tokens in Runtime::Perl5::Grammar, the ones writen in kp6 don't?
10:21 fglock the ones writen in kp6 should be generated correctly already
10:22 felipe joined #perl6
10:23 pmurias k
10:28 pmurias how should i modifiy the ast?
10:28 Juerd Announcement: SPPN (Dutch foundation for the promotion of Perl) has also sponsored feather's hardware upgrade.
10:28 cognominal__ joined #perl6
10:28 pmurias is a p5 attr on Token ok?
10:28 moritz_ SPPN++
10:29 fglock pmurias: why do you need to modify the ast?
10:29 Juerd They enable an even better upgrade (harddisks, even better CPU), and generously offered to compensate my time
10:30 Juerd I'm ordering the new components today
10:30 wolverian Juerd, nice :)
10:30 allbery_b I just got a propfind failed: can't connect to http://svn.pugscoode.org error.  expected?
10:30 pmurias token newline :p5 {\n}
10:30 Juerd The bad news is that feather will be offline either tomorrow or the day after tomorrow :)
10:30 allbery_b (i.e. is thie discussion of feather h/w upgrade related?)(
10:30 pmurias has to be represented as something
10:30 allbery_b pretend I spelled pugscode right
10:31 pmurias OTOHO i might emit is as a subroutine call
10:31 moritz_ allbery_b: that happens from time to time (and yes, it is ;)
10:31 allbery_b tried to update my existing tree
10:31 allbery_b ah, ok
10:31 pmurias &newline := p5token("\n")?
10:31 Blwood_ joined #perl6
10:31 Juerd allbery_b: feather is the shared Perl 6 development server, and is heavily overloaded to the level that new account requests are currently denied even.
10:31 pmurias fglock: makes sense?
10:31 allbery_b ah
10:31 fglock hmm
10:32 Juerd allbery_b: Its hardware will be upgraded, and the shared part is turned into a virtual server ("feather1") on top of a Xen domain-0 ("feather0"). Some services, like runpugs, can be moved to their own virtual machines ("feather2", ...) so they no longer interfere with the rest.
10:33 fglock i'd rather keep this as p5 code, with kp6 calling convention
10:34 ruoso fglock, that kinda hard to write and maintain
10:34 moritz_ harder than two branches of kp6? ;-)
10:34 ruoso writing a small subroutine that encapsulates most of it is usually easier
10:34 ruoso moritz_, heh
10:35 pmurias fglock: why would you prefere to keep it p5 code?
10:37 fglock because it's just a few lines of code, i think it doesn't need the syntax sugar
10:38 fglock and &newline := p5token("\n") is not much portable
10:38 fglock and kp6 doesn't recognize "\n"
10:39 pmurias that can me fixed
10:39 fglock ok
10:39 pmurias i think most possible backends understand p5 regex
10:39 pmurias fglock: i'm doing it on Ast level
10:40 pmurias s/.*//
10:40 pmurias it's actually &newline := p5token('\n')
10:41 fglock i wonder if it could break perl5rx
10:41 fglock possibly not
10:43 pmurias fglock: why?
10:43 pmurias would it break
10:45 pugs_svnbot r17678 | ruoso++ | [kp6] trying to visualize a PAD-centered compiler/interpreter
10:45 pugs_svnbot diff: http://dev.pugscode.org/changeset/17678
10:45 lambdabot Title: Changeset 17678 - Pugs - Trac
10:45 fglock pmurias: i think it is ok to implement this in kp6-kp6
10:46 fglock but if you need an AST transform, please add it as a separate visitor
10:48 wolverian isn't kp6-kp6 = 0?
10:49 fglock it's NaN
10:49 wolverian ah
10:49 ruoso wolverian, that I don't know, but kp6-mp6 + kp6-kp6 = kp6
10:49 wolverian interesting
10:52 ruoso but that's probably the reason fglock was trying to give the right-hand operator above a different name
10:52 ruoso AlmostPerl6 ;)
10:52 ruoso LooksLikePerl6
10:52 ruoso or something like that
10:52 wolverian operand?
10:52 ruoso yep yep
10:53 ruoso sry
10:53 wolverian somethingalongthelinesofperl6
10:53 wolverian resemblesperl6ifyouturnitupsidedownanddon'tlooktooclosely
10:53 wolverian too long, maybe.
10:53 ruoso rp6iytiudadltc
10:54 ruoso yeah... bad acronym too
10:54 ruoso fglock, the PAD centered thing may even provide a nice way to do the typemappings between p6 and p5
10:55 fglock ruoso: how?
10:55 ruoso as the parent pad would know one of its children is a p5 pad, and the child pad would know that its parent is a p6 pad
10:56 ruoso a macro language like XS would be needed anyway...
10:56 ruoso but it's possible to do it using pads that way
10:57 ruoso a XS-like macro language in p5
10:57 ruoso that sounds scary
10:58 pmurias ruoso: you want this for use Foo:from<perl5>?
10:59 ruoso pmurias, not exactly
10:59 ruoso but yes, that would be a side effect
11:00 ruoso I'm thinking about the compiler architecture
11:04 pmurias seen your diagram
11:04 buubot pmurias: Sorry, I haven't seen your
11:04 pmurias i think i understand
11:05 pmurias i don't think an XS like language would be needed
11:05 pmurias as p5 and p6 types corespond closly
11:06 ruoso pmurias, do they?
11:07 ruoso you can't even "shift" the arguments from p6 in p5?
11:07 ruoso !
11:07 ruoso s/?/!/
11:07 ruoso and after that, you cant use print $var;
11:07 ruoso because the $var is a hash
11:07 ruoso even if it's a string
11:09 pmurias ruoso: they are diffent at the internal level
11:09 ruoso ahm?
11:10 pmurias s/diffent/different/
11:10 ruoso they are different at the p5 level
11:10 pmurias at the implementation level
11:10 ruoso yep
11:10 ruoso so...
11:10 pmurias you have still to do the conversion
11:11 ruoso and that's what the XS-like macro language would help to do
11:11 pmurias but you can do it automaticly
11:11 ruoso but the question is if you want to
11:11 ruoso because that implies in even more overhead
11:13 ruoso fglock, the PAD thing is nice because it handles lexical grammar changes
11:13 ruoso fglock, (which I just realized it was what you was saying before)
11:13 * ruoso took some time to get that...
11:16 ruoso lunch &
11:25 pugs_svnbot r17679 | fglock++ | [kp6] make forcerecompile
11:25 pugs_svnbot diff: http://dev.pugscode.org/changeset/17679
11:37 danga joined #perl6
11:45 buetow joined #perl6
11:51 supery joined #perl6
11:57 lexkde joined #perl6
11:58 lexkde joined #perl6
12:00 qmole joined #perl6
12:23 mr_ank joined #perl6
12:29 xinming joined #perl6
12:36 franck____ joined #perl6
12:37 pmurias fglock: regarding implementing the :P5 modifier in kp6-kp6, does kp6-kp6 work right now?
12:38 pmurias or should i just port the changes when it starts to work
12:38 pmurias ?
12:39 franck____ joined #perl6
12:40 pugs_svnbot r17680 | pmurias++ | [kp6] - initial :P5 modifier implementation
12:40 pugs_svnbot diff: http://dev.pugscode.org/changeset/17680
12:40 lambdabot Title: Changeset 17680 - Pugs - Trac
12:43 franck___ joined #perl6
12:49 dmq32 joined #perl6
12:52 michaelkohwj joined #perl6
13:01 jhorwitz joined #perl6
13:15 fglock pmurias: re kp6-kp6, there are still some things missing
13:16 fglock match_p5rx should return a p6 Match; there is some possible reusable code in PCR that collects p5 match results
13:17 fglock http://svn.pugscode.org/pugs/perl5/Pugs-Compiler-Rule/lib/Pugs/Compiler/RegexPerl5.pm
13:17 lambdabot http://tinyurl.com/2d3mqm
13:18 fglock i think we should implement named parameters before bootstrapping
13:19 fglock so that .new() works properly
13:19 cmarcelo joined #perl6
13:19 fglock a possible simplification is to allow either named or positionals (but not mixed)
13:20 michaelkohwj joined #perl6
13:21 michaelkohwj left #perl6
13:23 Muixirt joined #perl6
13:29 daxim joined #perl6
13:45 TJCRI joined #perl6
13:53 |Jedai| joined #perl6
13:56 vel joined #perl6
13:57 pmurias fglock: i agree that we should implement named parameters now
13:58 fglock pmurias: i'm adding it to the grammar
13:59 pmurias k
14:02 cognominal_ joined #perl6
14:03 pmurias fglock: i'll modify CAPTURIZE to use a capture if it
14:03 pmurias 's passed as the only parameter
14:03 fglock ok
14:04 pmurias which compiler branch are you changing?
14:04 pmurias kp6-mp6 or kp6-kp6?
14:05 fglock kp6-mp6, will merge later
14:06 pmurias ok
14:06 pmurias i'll work on that one too
14:07 lisppaste3 joined #perl6
14:11 pugs_svnbot r17681 | fglock++ | [kp6] minor changes to grammar; allows pairs as parameters
14:11 pugs_svnbot diff: http://dev.pugscode.org/changeset/17681
14:11 lambdabot Title: Changeset 17681 - Pugs - Trac
14:11 fglock oops - i lost token_P5, reverting
14:12 drrho joined #perl6
14:12 wolverian holy crap, these windows computers have perl4 installed
14:13 xinming joined #perl6
14:21 moritz_ wolverian: and I thought Perl 5.6 was old ;)
14:21 fglock help
14:21 buubot fglock: Available Plugins:  oeis jeval spell cpan shorten rt topic squeeze rbeval acronym change managementspeak reverse join translate bibleit core imdb insult tv help echo eval convert jseval mangle restart quit deparse part pyeval reload_plugins goobooblink heap_test rss
14:22 wolverian moritz_, '93 :)
14:22 fglock i don't know how to revert my changes
14:22 moritz_ I don't know either, but you can try to svk up -r $last_revision, and then commit again
14:22 fglock moritz_: i tried, but it doesn't commit
14:23 moritz_ hrmpf
14:23 fglock "No targets to commit."
14:23 moritz_ maybe if you add small changes? (whitespaces)
14:23 fglock it seems i can only revert locally
14:23 pmurias svk merge
14:23 moritz_ that's an ugly workaround, I know
14:23 cognominal__ joined #perl6
14:23 pmurias can be used for reverting changes
14:24 pmurias svk merge -r N:N-1
14:25 pmurias for reverting changes at revision N
14:27 wolverian wow, svk is slow on feather
14:27 wolverian 'svk help' took a minute to run
14:27 wolverian using all available cpu
14:32 fglock "svk merge -r HEAD:18575" invokes the svk help
14:34 rgs "svk merge -r HEAD:18575 . ."
14:36 pmurias svk merge -r 8048:8047 //pugs/v6/v6-KindaPerl6 also worked here
14:38 pmurias fglock: assuming single captures are only created by the new calling convetion makes some tests fail
14:38 pmurias * convention
14:38 pmurias i need something like func(\(...),'newcc') is it ok?
14:40 Juerd wolverian: Everything is currently slow on feather
14:40 Juerd I've just, 10 minutes ago, ordered the components for feather v2
14:41 fglock pmurias: i don't understand - how did it break?
14:42 Juerd Core2Duo E6550 2330 MHz, 4 GB RAM, SATA 160 GB in RAID 1
14:42 Juerd On a nice Intel mainboard
14:42 fglock pmurias: i'm not changing the calling convention yet
14:42 wolverian Juerd, well, vim starts instantly
14:42 Juerd wolverian: vim's in cache :)
14:42 wolverian you'd think svk is too..
14:42 wolverian (even with all the modules)
14:43 Juerd wolverian: The numerous small files svk has to use are constantly pushed out of the buffer, I guess
14:43 wolverian yeah.
14:43 Juerd wolverian: Don't forget that feather used to have 2 GB RAM, and that 1 GB died.
14:43 Juerd 1 GB, shared with 10 people, isn't much :)
14:44 Juerd 320 MB of swap space is in use right now
14:44 pmurias fglock: captures passed with the old calling conventions are mistaken for the new one
14:44 Juerd I'll probably assign 2,5 GB to feather1
14:44 Juerd 0,5 GB to feather0, feather2, and feather3 each.
14:44 wolverian ouch, I didn't know that
14:45 Juerd 3 can run all the public eval stuff
14:45 fglock pmurias: i'm not sure what you mean, what's old and new
14:45 Juerd 2 can be used for important stuff like the main SVN repos.
14:45 Juerd Only feather1 stays as it is: semi-public access, and 0,2,3 will be more restricted.
14:46 Juerd diakopter: Are you available to migrate things across servers in a week or so?
14:46 xinming_ joined #perl6
14:49 pmurias fglock: the new calling convention is the one with all parameters passed in a capture, the old one is all parameters passed as positionals in @_
14:49 pugs_svnbot r17682 | fglock++ | [kp6] recompiled grammar
14:49 pugs_svnbot diff: http://dev.pugscode.org/changeset/17682
14:49 lambdabot Title: Changeset 17682 - Pugs - Trac
14:49 fglock pmurias: that's a painful change
14:49 pmurias so we aren't going to do it?
14:50 fglock hmm - i think it needs some migration plan
14:52 pmurias all the p6 can be easily migrated, what will cause problems is the p5 stuff
14:53 rgs pesky p5 stuff
14:53 Juerd Heh
14:57 Blwood joined #perl6
14:58 pmurias fglock: i don't know if the change is worth it
14:59 fglock how about push the named args as the last parameter to @_
14:59 pmurias fglock: it could be uses as a temporary solution
15:00 fglock yes
15:00 pmurias after the boostrap will have less p5 code
15:00 [particle] is kp6 grammar trying to mirror STD?
15:01 fglock [particle]: not yet
15:01 [particle] okay, that's the way nqp took, and supports named params now
15:01 fglock we need kp6 to do that
15:01 [particle] ah, i see. named params aren't native kp6
15:02 pmurias [particle]: parsing named params isn't the problem
15:05 TimToady joined #perl6
15:08 * pmurias reading a book&
15:19 fglock pmurias: i've got an implementation
15:19 pugs_svnbot r17683 | fglock++ | [kp6] fixed Ast for positional args
15:19 pugs_svnbot diff: http://dev.pugscode.org/changeset/17683
15:19 lambdabot Title: Changeset 17683 - Pugs - Trac
15:20 fglock s/positional/named/
15:24 ruoso pmurias, p5 modifier is still not merged to kp6-kp6, right?
15:25 amnesiac joined #perl6
15:33 pmurias ruoso: not yet
15:34 pmurias ruoso: do you need it
15:37 pmurias ?
15:51 diakopter Juerd: maybe
15:51 Juerd diakopter: Would be great
16:10 TimToady .oO(Captures are pretty close to anonymous typeglobs...)
16:14 fglock i wonder if captures could be pads
16:17 TimToady sounds heavy
16:17 pugs_svnbot r17684 | fglock++ | [kp6] initial change to CAPTURIZE
16:17 pugs_svnbot diff: http://dev.pugscode.org/changeset/17684
16:17 lambdabot Title: Changeset 17684 - Pugs - Trac
16:17 justatheory joined #perl6
16:23 xinming_ joined #perl6
16:25 snearch joined #perl6
16:25 ruoso pmurias, it would be nice to keep both as close as possible
16:33 TimToady joined #perl6
16:35 TimToady joined #perl6
16:35 pugs_svnbot r17685 | fglock++ | [kp6] CAPTURIZE understands named args
16:35 pugs_svnbot diff: http://dev.pugscode.org/changeset/17685
16:35 lambdabot Title: Changeset 17685 - Pugs - Trac
16:36 diakopter There's that word again, 'heavy'. Why are things so heavy in the future?
16:41 ruoso diakopter, because premature optimization is the root of all evil? ;)
16:42 ruoso BTW... it looks like the version after kp6 will be p6 in p6 already
16:42 ruoso completely p6 i mean
16:42 ruoso no p5 emitting at all
16:42 wolverian cool!
16:46 zamolxes joined #perl6
16:47 araujo joined #perl6
16:48 pmurias ruoso: how do you plan to do it wholly with p6?
16:51 pmurias s/with/in/
16:53 ruoso pmurias, I'm drawing a diagram to explain it
16:53 pmurias ok
16:57 barney joined #perl6
16:57 * ruoso just notices that it's not quite a diagram, just a flow of ideas using boxes and arrows
17:02 pugs_svnbot r17686 | ruoso++ | [kp6] writing down ideas of how to write p6-on-p6 based on kp6
17:02 pugs_svnbot diff: http://dev.pugscode.org/changeset/17686
17:02 lambdabot Title: Changeset 17686 - Pugs - Trac
17:03 kaether joined #perl6
17:06 pmurias ruoso: ah, so you plan to interpret AST
17:06 ruoso yep
17:06 ruoso but not in kp6...
17:06 ruoso only in the other version
17:07 ruoso because kp6 will the door to provide alternative runtimes
17:07 ruoso or "vm"s if you wish
17:07 pmurias i had thought of something more evil, i'll write when i get back
17:07 pmurias visiting a friend&
17:19 jisom joined #perl6
17:27 riffraff joined #perl6
17:29 elmex joined #perl6
17:33 pugs_svnbot r17687 | fglock++ | [kp6] implemented %_
17:33 pugs_svnbot diff: http://dev.pugscode.org/changeset/17687
17:33 lambdabot Title: Changeset 17687 - Pugs - Trac
17:36 pugs_svnbot r17688 | fglock++ | [kp6] fixed a test
17:36 pugs_svnbot diff: http://dev.pugscode.org/changeset/17688
17:36 lambdabot Title: Changeset 17688 - Pugs - Trac
17:36 spinclad joined #perl6
17:39 Psyche^ joined #perl6
17:39 jamesmck joined #perl6
17:50 iblechbot joined #perl6
17:50 pugs_svnbot r17689 | fglock++ | [kp6] fixed last commit
17:50 pugs_svnbot diff: http://dev.pugscode.org/changeset/17689
17:50 lambdabot Title: Changeset 17689 - Pugs - Trac
17:58 cognominal_ joined #perl6
18:06 REPLeffect joined #perl6
18:10 xinming joined #perl6
18:11 pugs_svnbot r17690 | fglock++ | [kp6] refactored Array and Hash out of MOP
18:11 pugs_svnbot diff: http://dev.pugscode.org/changeset/17690
18:11 lambdabot Title: Changeset 17690 - Pugs - Trac
18:17 * ruoso merging all changes from lib-kp6-mp6 to lib-kp6-kp6
18:19 * ruoso is always impressed by the time it takes to compile kp6-kp6
18:21 weinig|zZz joined #perl6
18:21 renormalist joined #perl6
18:23 pugs_svnbot r17691 | fglock++ | [kp6] force recompile
18:23 pugs_svnbot diff: http://dev.pugscode.org/changeset/17691
18:23 lambdabot Title: Changeset 17691 - Pugs - Trac
18:24 * ruoso probably needs to recompile again
18:24 pugs_svnbot r17692 | ruoso++ | [kp6] merging changes from kp6-mp6 to kp6-kp6. (r17676:HEAD). all tests pass in both runtimes
18:24 pugs_svnbot diff: http://dev.pugscode.org/changeset/17692
18:24 lambdabot Title: Changeset 17692 - Pugs - Trac
18:24 * ruoso recompiling again
18:24 * [particle] updates to look at kp6
18:25 * fglock is always impressed by ruoso patience
18:25 [particle] nothingmuch: ping
18:25 nothingmuch [particle]: pong
18:25 TimToady it takes pugs a minute or two to recompile STD; I hate to think how long STD on STD would take...
18:26 rindolf joined #perl6
18:26 ruoso I have been thinking...
18:27 ruoso and I still can't see an easy way to p5 and p6 to interoperate
18:27 ruoso we'll need a XS-like glue
18:27 ruoso between them
18:28 ruoso the semantics are too much different...
18:28 ruoso fglock, that temp.pm thing in kp6-mp6 compilation is really necessary?
18:29 ruoso it prevents the use of make -j
18:29 ruoso s/-j/-j2/
18:29 fglock i think so, otherwise you end up modifying a lib that is in use
18:29 ruoso so, I'll think I'll hack it to use a temp name
18:30 ruoso instead of a fixed "temp.pm" name
18:30 fglock ok
18:31 drupek1260 joined #perl6
18:31 pmurias ruoso: the temp.pm file used in kp6-mp6 compilation is used to avoid creating empty files
18:32 pmurias so that you don't have to svk revert to recompile if there is a syntax error
18:32 pmurias it used to be without it
18:32 ruoso pmurias, I'll create a file named the same as the original with a .temp appended
18:32 ruoso instead of a fixed temp file
18:35 pugs_svnbot r17693 | ruoso++ | [kp6] recompilation after merge
18:35 pugs_svnbot diff: http://dev.pugscode.org/changeset/17693
18:36 renormalist irc looks like you solved kp6-in-kp6, is that true?
18:37 devogon joined #perl6
18:38 fglock renormalist: it is still missing a few runtime tweaks, as it no longer uses the mp6 calling convention internally
18:41 ruoso yep.. kp6-kp6 still doesn't run
18:41 pmurias ruoso: not perltidy'ing everything might speed the compliation up
18:41 pugs_svnbot r17694 | ruoso++ | [kp6] allows -j2 in make
18:41 pugs_svnbot diff: http://dev.pugscode.org/changeset/17694
18:41 lambdabot Title: Changeset 17694 - Pugs - Trac
18:41 ruoso pmurias, I think we can remove that already
18:42 ruoso we're having less problems with the generated code
18:42 nothingmuch ?w 61
18:42 lambdabot Maybe you meant: web1913 what where where+ wiki wikipedia wn world02 . v
18:43 fglock perltidy results in more readable diffs
18:43 ruoso in fact
18:44 nothingmuch except for the diff where you started using it ;-)
18:44 ruoso that's a good reason
18:44 nothingmuch consistency makes for readable diffs
18:44 * ruoso hails to the goodness of wdiff
18:45 ruoso but -j2 already helps a lot in the time it takes
18:45 ruoso considering you have a dual core processor
18:46 * pmurias dosn't
18:46 * renormalist had to revert Test::Harness to stable
18:48 [particle] rats, the kp6 makefile isn't windows friendly
18:48 [particle] s/windows/nmake/
18:48 ruoso [particle], oops
18:48 [particle] $(patsubst ...), $(wildcard ...) etc
18:49 ruoso but at this moment you don't need to issue make
18:49 [particle] i'm trying with 'make' now
18:49 fglock [particle]: make'ing is optional
18:49 xdg joined #perl6
18:49 fglock it just p5 code
18:49 ruoso svn contains the runnable code as well
18:50 ruoso [particle], unless you have a sh-like shell, it won't work in windows
18:50 ruoso there are also shell constructs in the kp6-mp6 compilation
18:50 [particle] btw jhorwitz has made great progress with mod_perl6:  http://nopaste.snit.ch:8001/11083 and http://www.smashing.org/sandbox/perl6-handlers/counter
18:52 fglock yay!
18:52 fglock is it based on mod_parrot?
18:52 renormalist [particle]: what does mod_perl6 run on? pugs? kp6?
18:54 jhorwitz parrot's perl6 compiler on top of mod_parrot
18:54 * jhorwitz awakens from two years of slumber on #perl6
18:55 [particle] i wonder what jhorwitz will do after coffee and breakfast....
18:55 pmurias how can one have an error at line 0?
18:55 pmurias eval {...} called at lib-kp6-kp6-p5//KindaPerl6/Runtime/Perl6/Compiler.pm line 0
18:55 buubot pmurias:  ''ERROR: syntax error at (eval 212) line 1, near "{..."
18:55 jhorwitz [particle]: that's nap time!
18:55 ruoso pmurias, you're trying to run kp6-kp6
18:55 ruoso right?
18:55 pmurias yes
18:56 ruoso yep... that's really weird
18:56 ruoso it does say line 0
18:56 ruoso but I think that's from kp6 itself, not from pelr
18:56 ruoso perl
18:57 ruoso so, it's not so weird
18:57 pmurias the backtrace is generated by perl
18:58 pmurias by Carp::confess
18:58 ruoso hmmm.. the global variable initialization is broken again
18:58 ruoso in kp6-kp6
18:59 fglock jhorwitz: congrats!
18:59 pmurias ruoso: what should i do if i want to help with kp6-kp6
18:59 pmurias ?
18:59 jhorwitz fglock:  thanks.  pmichaud++ and particle++ for lots of help catching up with parrot
19:00 ruoso pmurias, port Pad.pm to kp6 calling convention
19:00 pmurias ok
19:00 ruoso and all other files in docs/KP6-Bootstrap-Notes.txt
19:01 fglock pmurias: you can now use named parameters in .new() - this should make things easier
19:01 fglock hmm - i need to reimplement .new
19:01 ruoso hmmm... globals are broken in both runtimes
19:01 ruoso intersting
19:02 pmurias global being
19:02 pmurias ?
19:02 pugs_svnbot r17695 | ruoso++ | [kp6] setting a global variable is broken in both runtimes
19:02 pugs_svnbot diff: http://dev.pugscode.org/changeset/17695
19:02 lambdabot Title: Changeset 17695 - Pugs - Trac
19:03 pmurias ruoso: it Global.pm
19:03 pmurias being a pain
19:03 pmurias it turn Main::foo into GLOBAL::Main::fo
19:03 pmurias o
19:03 ruoso but, is the variable being initialized somewhere?
19:03 fglock looking
19:05 pmurias irc eats a my 's' chars
19:05 pmurias *all
19:05 pmurias ;)
19:06 fglock where can i read more about BUILD() vs. new() vs. bless() ?
19:07 ruoso it's not being initialized at all
19:07 ruoso pmurias, the translation to GLOBAL::Main::foo kinda makes sens
19:08 ruoso but the problem is that it doesn't initializes it anyway
19:08 ruoso I made a hack other time
19:08 ruoso I could do again
19:08 fglock ruoso: yes, the AST transformations are ignoring globals
19:08 ruoso which is to test inside the DISPATCH_VAR if $_[0] is undef
19:08 ruoso and in this case, define it...
19:08 fglock this may mask real errors
19:08 fglock i'd rather fix it
19:09 fglock though globals are autovivified in p6 too, right?
19:09 ruoso the problem with that is that you never know when to initialize the global
19:09 ruoso I think so
19:09 fglock so i think it is ok to autovivify then
19:10 rlpeacock joined #perl6
19:13 Grrrr joined #perl6
19:14 pmurias ruoso: i think Pad.pm could be ported to perl6
19:14 ruoso fglock, which is the correct way of doing the variable initialization to a container with undef in p5 land?
19:15 ruoso I'm doing a $_[0] = ::DISPATCH( $::meta_Container, 'new' );
19:15 ruoso and it doesn't seems to be ok
19:15 fglock ruoso: see the initialization of 'my' vars in compiloed code
19:16 fglock ::DISPATCH( $::Container, 'new' )
19:16 fglock oops
19:16 fglock $::Scalar
19:16 renormalist lib-kp6-mp6/KindaPerl6/Runtime/Perl5/Compiler.pm   does not have a finishing   1;  at the end. Is that wanted?
19:17 renormalist I have compile errors due to this, but not always, eg. not when I do a complete make
19:20 ruoso hmmm
19:20 ruoso fglock, not enough
19:21 ruoso Undefined subroutine &main:: called at lib-kp6-mp6-p5/KindaPerl6/Runtime/Perl5/MOP.pm line 711
19:21 fglock hmm - you may need to assign $::Undef to it
19:21 pmurias renormalist: you can probably add '1;'
19:23 pmurias fglock: you shouldn't have to
19:23 ruoso pmurias, in fact... it already is undef by default
19:23 fglock ruoso: i'll fix it
19:23 pmurias FETCH returns ::DISPATCH($::Undef,'new',0) if you have p5 undef
19:24 pmurias i remember modifing it to do so
19:25 ruoso fglock, I think it needs to be done in a AST transform
19:25 ruoso to add the auto-vivifying code in the code itself
19:26 pmurias well we could have :: call a method on the package object
19:26 pmurias it would come in handy for :from<perl5> for non-oo modules
19:28 fglock hmm - i don't think i can fix it right now - i'm tired :P
19:29 fglock we need a new implementation of namespaces
19:29 ruoso fglock, can we hold it for the version after kp6 bootstrap?
19:30 fglock yes, if you can work around globals
19:33 fglock home &
19:48 danga joined #perl6
19:51 pugs_svnbot r17696 | pmurias++ | [kp6-kp6] Runtime/Perl6/Grammar.pm
19:51 pugs_svnbot r17696 | pmurias++ | translation to p6 - dosn't compile or run yet
19:51 pugs_svnbot diff: http://dev.pugscode.org/changeset/17696
19:51 lambdabot Title: Changeset 17696 - Pugs - Trac
19:58 REPLeffect joined #perl6
19:59 xdg joined #perl6
20:00 weinig|zZz joined #perl6
20:03 lichtkind joined #perl6
20:05 lichtkind moritz_: hihi
20:07 pmurias lichtkind: hellohello
20:09 lichtkind pmurias: hello, so glad to be in perl channel since weeks
20:11 renormalist lichtkind: why were you absent? vacation?
20:12 lichtkind renormalist: pilgrim in spain
20:13 Patterner no yapc:eu?
20:14 lichtkind Patterner: no german perl workshop was enough for me
20:16 szbalint yapc:eu was nice :)
20:17 lichtkind szbalint: yea tinita is currently explaining it :)
20:32 elmex joined #perl6
20:40 elmex joined #perl6
20:51 pmurias sleep&
20:52 pmurias why do i have to sleep when i start having most fun hacking on something? ;)
20:57 xinming joined #perl6
21:26 pugs_svnbot r17697 | renormalist++ | - just a finishing ;1 , no more, no less
21:26 pugs_svnbot diff: http://dev.pugscode.org/changeset/17697
21:26 lambdabot Title: Changeset 17697 - Pugs - Trac
21:27 elmex joined #perl6
21:33 Belaf joined #perl6
21:46 TJCRI joined #perl6
22:21 justatheory joined #perl6
23:10 Limbic_Region joined #perl6
23:13 polettix joined #perl6
23:48 mncharity joined #perl6

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

Perl 6 | Reference Documentation | Rakudo