pugscode.org/ | nopaste: sial.org/pbot/perl6 | ?eval [~] <m oo se> | We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by diakopter on 11 July 2007.
00:00 Notrix joined 00:14 ludan joined 00:26 thoughtpolice left 00:37 thestarslookdown left 00:38 KidShenck_ joined 00:40 jisom left 00:57 penk left 01:00 FurnaceBoy joined 01:02 como left 01:05 lyokato joined 01:10 Notrix left 01:12 FurnaceBoy left 01:14 kanru left 01:24 lichtkind left 01:43 agentzh joined 01:46 __Ace__ joined 01:48 lisppaste3 joined 02:00 devogon left 02:01 `nipra joined 02:23 justatheory joined 02:43 Khisanth joined 03:06 elmex_ joined 03:09 jisom joined 03:20 elmex left 03:31 justatheory left 04:04 Lunchy joined 04:14 weinig left 04:16 ludan left 04:18 Lunchy left, Lunchy joined 04:30 amnesiac joined 04:35 kane_ joined 04:53 Jamtech joined, IllvilJa joined 05:03 BinGOs left 05:06 kane__ joined 05:20 kane_ left 05:26 Jamtech left 05:27 kane_ joined 08:05 ilogger2 joined 08:06 iblechbot joined 08:32 franck___ joined 08:44 baalial joined 09:06 drupek12 joined 09:15 drupek12 left 09:52 drupek12 joined 10:15 ruoso joined 10:28 iblechbot left 10:36 nipotaway is now known as nipotan 10:53 ebassi joined 11:53 gongyiliao joined 12:12 iblechbot joined 12:15 lichtkind joined 12:16 ting joined 12:22 ruoso left 12:25 chris2 joined 12:29 ebassi is now known as ebassi|lunch, drrho joined 12:32 ruoso joined 12:33 nipotan is now known as nipotaway 12:38 jiing joined, [1]jaffa joined, Gekz joined
[1]jaffa Hi 12:39
how is parrot going?
12:42 iblechbot left
moritz_ [1]jaffa: quite well, you can expect the 0.5 release this month 12:54
[1]jaffa can it generate executables? 12:56
12:57 rindolf joined
[1]jaffa when is it going to be finished? 12:57
moritz_ no, it's a virtual machine, not a compiler
and there is no release date - it's released when it's done ;)
but the folks in #parrot on irc.perl.org can tell you more about it 12:58
[1]jaffa that channel is not very active 12:59
moritz_ oh, it is - are you sure you were on irc.perl.org? 13:00
currently 69 nicks in there
[1]jaffa no, I was not on it. 13:01
so do you use Perl 6?
moritz_ a bit, yes
and I try to hack on kp6 from time to time
[1]jaffa kp6?
moritz_ KindaPerl6, a perl 6 compiler written in Perl 6 13:02
13:03 cmarcelo joined
[1]jaffa is it complete? 13:04
moritz_ no
that's the 'Kinda' part in the name :)
[1]jaffa is the definition of Perl 6 complete?
moritz_ not yet either
large parts of the language are specced 13:05
but for example concurrency is still missing
perlcabal.org/syn/
lambdabot Title: Official Perl 6 Documentation
[1]jaffa what is the backend of Perl 6 compiler? 13:06
moritz_ there's none yet 13:08
and every compiler/interpreter that passes the test suite will be "official perl 6"
the test suite is in the pugs svn repository, and is about half way done 13:09
[1]jaffa is there a competition?
moritz_ yes
or no ;)
well, there are different approaches
and hopefully they'll meet in the middle somewhere
pmichaud++ implements perl 6 on parrot
13:10 ebassi|lunch is now known as ebassi
moritz_ fglock++, ruoso++, pmurias++ and a few others hack on kp6 (which is a way to bootstrap 6-on-6) 13:10
[1]jaffa ++?
moritz_ TimToady++ writes the Perl 6 grammar in Perl 6
[1]jaffa: I increase their karma while mentioning them ;)
@karma TimToady 13:11
lambdabot TimToady has a karma of 6
[1]jaffa are you playing a game?
moritz_ not really
it's just programmer's fun
a very crude measurement for popularity+effort 13:12
Juerd Incredibly crude and inaccurate :) 13:13
moritz_ but still fun
just for the sake of it ;)
[1]jaffa I wonder why perl6 has not been finished by now.
Juerd I think most people don't do it for the counters, but to indicate that they like certain people or their efforts.
moritz_ Juerd: did you find any tuits to restart trac?
Juerd Er
moritz_ Juerd: I couldn't find out how to do it :(
13:14 baalial left
Juerd If trac doesn't live in apache2, then I don't know where it does. 13:14
moritz_ [1]jaffa: because it's a very powerfull, complex and totally new language
[1]jaffa So?
moritz_ [1]jaffa: it tries to embrace many different programming paradigms, and that's not easy to bring together
[1]jaffa don't you have enough manpower?
moritz_ that's one of the problems, yes
would you like to help? 13:15
obra: ping
Juerd [1]jaffa: You should realise that while Perl 6 was announced in 2000, implementation began in 2005
[1]jaffa: And the design still changes a little every now and then.
[1]jaffa who is controlling the process? 13:16
Juerd [1]jaffa: Nobody.
[1]jaffa is it going anywhere?
Juerd Yes, it is
moritz_ of course ;)
Juerd [1]jaffa: It's a completely volunteer effort, so one can't simply hand out tasks and expect them to be done. 13:17
[1]jaffa why python seems to move faster....
Juerd [1]jaffa: Things take a lot of time.
[1]jaffa: Python has more funding, and doesn't try to create a wholly new language :)
moritz_ [1]jaffa: python was not a complete redisgn
Juerd [1]jaffa: Also, don't worry about Perl 6 too much. There is Perl 5, which is a really great programming language. And we won't shoot you for using Python or Ruby either :) 13:18
[1]jaffa I use Perl 5
I have written an IDE in it
Juerd Interesting 13:19
url?
moritz_ and soon there'll be perl 5.10 out
[1]jaffa I have not published it
moritz_ what language is the IDE for? perl as well? 13:20
[1]jaffa perl,c++,D
moritz_ sounds cool ;)
[1]jaffa it uses wxPerl 13:21
13:22 Torment joined
[1]jaffa the problem is if perl 6 is not released fast enough, it is not going to be popular if that matters. 13:24
there are many alternatives.
Perl6 , who is it for?
moritz_ we know, but increasing pressure on the developers doesn't speed up the process 13:25
[1]jaffa if per6 design were complete, could parrot host it?
is parrot complete enough? 13:26
moritz_ read www.perl.com/pub/a/2005/09/22/onion.html for "whom is it for"
lambdabot Title: perl.com: The State of the Onion 9
moritz_ regarding parrot - I think it's a worthy backend target right now, and improves steadily 13:27
every turing complete VM could host perl 6, so you have to ask "how well is suited" 13:28
[1]jaffa ok 13:29
moritz_ and real soon it'll implement (nearly?) all of Perl 6's object model
[1]jaffa I would like to see executable backend.
moritz_ you're welcome to write one ;)
Juerd [1]jaffa: You seem to be under the impression that Perl 5 is not used? 13:30
[1]jaffa no, it is used.
I am aware of that.
moritz_ the problem is that since perl 6 has an eval(), so every compiled exeutable that uses eval() has to have a full Perl 6 compiler linked in
[1]jaffa why is it a problem?
moritz_ it's not exactly a problem, but most people don't like that 13:31
imagine each C program having gcc linked in :-)
[1]jaffa I don!t use eval a lot
do you?
moritz_ yes
I use 'use'
which in turn calls require() 13:32
which in turn calls do
13:32 iblechbot joined
moritz_ which in turn runs eval 13:32
[1]jaffa but that is compile time thing,
if I remember correctly
moritz_ right ;)
at runtime I tend to avoid eval $string, and only use eval { ... } (which is completely different) 13:33
but there are cases where string eval is really handy
avar Perl 6 doesn't only need the compiler for eval() it needs it for all the reasons common lisp compilers need a present compiler (which includes eval)
rgs if ($feature_is_supported) { require Foo::Bar }
moritz_ right
avar inc. macros
[1]jaffa well, all requires could not included 13:34
well, all requires could be included
moritz_ you could have requires that collide
but never happen simultaniously at run time
[1]jaffa I see 13:35
moritz_ there could be disambiguation methods, but they'd be rather complicated, and not worth the trouble
rgs see alos: plugins 13:36
s/los/lso/
[1]jaffa I see 13:38
13:46 ilogger2 joined
moritz_ anyway, bashing java is easy, implementing perl 6 isn't 13:46
[1]jaffa If I use win32 , it is not going to be platform independent. 13:47
ggoebel moritz: any word on a commit bit / Trac? Whose door do I need to knock on if Juerd can't do it?
Who are the maintainers for the various bots and services that pugs and onion depend on? 13:48
moritz_ ggoebel: I pinged obra++, who is our trac guru
@seen obra 13:49
lambdabot obra is in #perl6. I last heard obra speak 15h 18m 3s ago.
[1]jaffa don't you think?
ggoebel If you use the right CPAN modules Perl isn't bad on Win32. Except for non-blocking IO...
moritz_ ggoebel: the maintainers are in the file 'SLAVES' in pugs root
[1]jaffa I mean Win32:: modules.
moritz_ [1]jaffa: if you want to access functionality that's only available on win32, there's no way around it 13:50
and there's no point in disallowing it
ggoebel mncharity: ping 13:51
We sure... if you opt in to the platform specific modules... you're code will be... platform specific!
s/we sure/well sure/ 13:52
[1]jaffa It makes me think, what does platform-independent mean?
I ported a perl code from Win32 to UNix 13:53
I had to deal with the files
how they are represented.
moritz_ with File::Spec?
[1]jaffa no
if I just use open, print, close... 13:54
Juerd What's wrong with trac on feather, by the way? It appears to work.
[1]jaffa that in self, not platform independent.
moritz_ Juerd: it doesn't work for me, proxy timeout
Juerd Weird, 'cause it's very fast for me
moritz_ which url? 13:55
Juerd dev.pugscode.org
moritz_ I try commitbit.pugscode.org/
Juerd Commitbit isn't trac
moritz_ oh damned, I always thought it was a trac plugin
then mentally rephrase all my questions with s/trac/commitbit/ 13:56
;-)
Juerd I restarted it but now it gives a 503
To restart commitbit, btw: cd /service; sudo svc -d commitbit; sudo svc -u commitbit 13:57
Ah, works now
Apparently it took some time to start - or someone else made it work in the past minute
moritz_ @tell ggoebel invitation sent. It's a custom to add your name to AUTHORS as first commit (and commit test) 14:00
lambdabot Consider it noted.
moritz_ Juerd: thanks
anybody else who wants a commit bit? ;-)
Juerd me! me!
[1]jaffa ok
guys 14:01
bye
moritz_ Juerd: /msg me your email and desired nick ;)
bye [1]jaffa, nice chat ;-)
s/, /, that was a / ;-)
Juerd moritz_: :)
14:18 ruoso joined 14:57 iblechbot joined 15:08 TJCRI joined 15:46 amnesiac joined 15:52 cognominal_ joined 15:53 bbkr joined 15:55 cognominal_ left 15:57 diakopter left, diakopter joined
obra ok. if you're hacking on Perl 6 implementation, raise your hand. 16:01
I'm going to try to get together a "state of the onion patch" 16:02
moritz_ raises hand tentativeley 16:03
btw you can ignore my previous ping ;)
obra ok. 16:04
so. there are lots and lots of questions about "what implementations exist. where are they?"
I'd like to blog at the very least a list of implementations and how far along are they.
Are you up for braindumping to me what you know? ;)
moritz_ i'd like to see that in a wiki, so it can be edited easily 16:05
ok, there's pug, which is p6 on haskell
atm only plain haskell backend works, other backends are bit rotten
there's parrot, which is the p6 VM, and many languages are implemented based on parrot 16:06
obra (important bit of data: how much of perl6 is implemented)
quibble: parrot isn't quite "The p6 vm" - it's a vm. the p6 implementation is very basic.
moritz_ pugs has much of the syntax right, lists, hashes, basic OO, basic regex, a bit IO 16:07
most control constructs work in pugs
but currently there is little pugs work, mostly because of audreyt++ being absent
did I forget anything about pugs? 16:08
obra is there anyone else who's working to maintain pugs for synopsis updates?
16:08 dlocaus joined 16:09 Aankhen`` joined
moritz_ there weere a few attempts by lumi++, masak++ 16:09
and TimToady++ did a few haskell hacks as well
but the last haskell commit is more than a month ago, if I recall correctly
obra *nod* 16:10
moritz_ PIR and JS backends compile, but fail every single test in the test suite 16:11
anything else you want to know about pugs? 16:12
obra I think that's good for pugs ;)
So. who's next?
moritz_ mp6/kp6 I'd say 16:13
mp6 is bootstrap with its perl5 backend, PIR backend exists, but doesn't bootstrapped 16:14
mp6 is pretty much frozen, except for occasional fixes or small feature additions
obra the goal of mp6 isn't to be a full implementation, right? 16:15
moritz_ right
just enough to code kp6 in it
that is, mostly regex/rules and a bit OO
dlocaus Juerd: responding to your comment, that it being a volunteer effort and one just can't hand out tasks, I think helping out would be a bit easier if the "super" strucutre of kp6 was already broken out into easily testable tasks.
16:16 BinGOs joined
dlocaus kind of black box, but it may just be my timing in coming into the project... 16:16
16:16 BinGOs left
moritz_ obra: kp6 is written in mp6, partly in kp6 as well, and has perl5 and CL runtime 16:16
obra: kp6 implements grammars, OO, basic junctions, multi methods, 16:17
obra: dan BEGIN blocks, closures 16:18
s/dan/and/
obra how much of the pugs test suite can it run?
avar The CL runtime is incomplete
moritz_ not much, because it lacks much syntactic sugar
avar And it looks like there were some regressions recently
moritz_ and the grammar has precendence problems 16:19
the plan was to get rid of thoses by swithcihng to STD.pm
obra how close is that?
moritz_ the current plan seems to be to insert another bootstrapping stage before, v6-Onion 16:20
Juerd dlocaus: Well, yes, I agree in general.
moritz_ obra: you have to ask fglock, he's working on that for quite some time...
16:20 BinGOs joined
obra *nod* 16:20
Juerd dlocaus: Not much effort goes into making collaboration easier, unfortunately.
moritz_ he implemented lazy gather/take to supporrt TimToady's STD/metholate 16:21
obra moritz_: v6-Onion? (note that Onion is the name of the implementation of Perl 6 on Parrot)
dlocaus Juerd: oi! soemthing for me to work on.
I'm currently looking at trying to solve t/kp6/internals/02-gather.t
or ... documenting something
Juerd dlocaus++
moritz_ obra: yes, but there's no better na atm
s/na/name/
I'm not too happy with that 16:22
obra but it has nothing to do with the parrot implementation, right?
moritz_ right
16:27 pmurias joined
pmurias dlocaus: t/kp6/internals/02-gather.t work here 16:27
obra any other interesting p6 hacking efforts that haven't been getting a lot of press? ;) 16:28
dlocaus pmurias: 02-gather.t works here? 16:30
works "there"
pmurias yes 16:31
do you have Coro?
dlocaus Coro? 16:32
pmurias Coro.pm from cpan
dlocaus no, I didn't know it was needed.
Installing... 16:33
moritz_ obra: there's yaph
avar llvm.org/docs/tutorial/
lambdabot Title: LLVM Tutorial: Table of Contents
moritz_ obra: but I don't know much about it
avar onion/kp6 on llvm would be neat
moritz_ oa parrot emitter ;) 16:34
pmurias mp6 is bootstraped
avar obra: since when is the p6 on parrot called onion?
16:34 IllvilJa joined
moritz_ it isn't, officially 16:34
obra avar: um. at least 6 months ago. you'd have to ask pmichaud 16:35
dlocaus I think that happened a few days ago
pmurias obra: the reason for the kp6 bottleneck is that it can bootstrap itself yet
obra it was something patrick brought up on the concall a while back. should be in the minutes chromatic posted.
avar somebody mentioned it as a /potential/ name in a mailing list
obra Patrick has been calling it onion, even if he hasn't had a lot of cycles to _hack_ on it. 16:36
dlocaus irclog.perlgeek.de/perl6/2007-10-25#i_131721
lambdabot Title: IRC log for #perl6, 2007-10-25
obra Regardless, he totally has prior claim on the name ;)
moritz_ right
obra dlocaus: that's the new thing folks here are discussing. not pmichaud's "perl 6 on parrot" 16:37
dlocaus just wants p6 out the door, before PERL goes down in history, next to bash, tcsh, and sh.
moritz_ but it would be nice to know if he _really_ plans to rename use the name
obra I promise to ask him this week
moritz_ ok, do that ;)
obra There are many names in the world. I'm sure we can come up with something else ;)
16:37 weinig joined
obra we/you/whichever. not going to force any names on anyone. just make sure we don't conflict 16:38
moritz_ and I'll try to convince everybody to name the new thing v6-Garlic ;)
dlocaus cool, ok 02-gather.t passes.
I'll document that.
obra heh
avar dlocaus: uh, bash?
rgs or "thyme"
moritz_ rgs: that's nice, actually ;)
dlocaus yes. There was an article on Perl buzz, saying that perl wasn't really that dead, and that it appeared popularly in many job offers. 16:39
moritz_ avar: bash is _so_ web 1.0 - web 2.0 uses zsh ;)
dlocaus When I see perl on craigslist.org job offers, it generally is right next to a list of shell scripting languages.
avar: When I was doing job interviews, most people think perl is a shell scripting language. It drives me nuts.
I like tcsh.
obra ..web 2.0 uses yubnub 16:40
dlocaus I'm thinking I should start a beginners.html guide. Any thoughts?
allbery_b an annoying number of people program perl as if it were a shell scripting language
dlocaus Take everything I've learned to-date (not much) and document it in docs/ 16:41
moritz_: ?
moritz_ dlocaus: just kidding ;) 16:42
dlocaus moritz_: no, I was asking about starting a beginners.html guide.
Document things like Coro.pm being needed.
pmurias dlocaus: i found the cause of the bug you looked for
dlocaus 15&17? 16:43
pmurias yes
dlocaus ok.
moritz_ dlocaus: ah yes, that's a good idea
dlocaus: I just didn't realize it was a question ;)
dlocaus oh. np :)
pmurias: did you fix the bug? 16:45
pmurias no 16:47
dlocaus pmurias: ok, are you passing this back to me? or working on it?
pmurias i'm not working on it so feel free 16:49
dlocaus ok.
pmurias dlocaus: do you know how to implement -> {return 1}? 16:53
16:55 zamolxes joined, jjore-w joined
pmurias fixing a web 2.0 bug& 16:56
16:56 pmurias left
pugs_svn r18785 | dlo++ | [KP6] A basic beginners guide to helping out in KindaPerl-6. 17:01
r18785 | dlo++ | Please comment and update.
dlocaus @tell pmurias "I don't know how to implement {return 1}
lambdabot Consider it noted.
pugs_svn r18786 | dlo++ | [kp6] added comment about prefixing with [kp6] 17:06
moritz_ recursion++ ;-)
pugs_svn r18787 | dlo++ | [kp6] Added a recent commits link 17:07
r18788 | dlo++ | added TODO list. 17:10
17:10 pmurias joined
pmurias t/kp6/56-token-sym.t i think is the easiest thing to hack on right now 17:12
lambdabot pmurias: You have 1 new message. '/msg lambdabot @messages' to read it.
pugs_svn r18789 | dlo++ | [kp6] Minor formating changes.
dlocaus pmurias: was that last message for me? 17:13
pmurias from you
dlocaus pmurias: (this one)(09:12:04) pmurias: t/kp6/56-token-sym.t i think is the easiest thing to hack on right now
pmurias yes 17:14
17:14 amnesiac left
dlocaus humm, ok. I'll work on that then. (I haven't started on grammar_17_15_bug_ilustration.t) 17:14
17:15 amnesiac joined 17:18 amnesiac left
moritz_ kp6: say "'escaping test 1'"; 17:25
exp_evalbot r18789: OUTPUT['escaping test 1'ā¤]
moritz_ kp6: say '"escaping test 2"'; 17:26
exp_evalbot r18789: OUTPUT["escaping test 2"ā¤]
moritz_ kp6: say 'escaping \' test 3';
exp_evalbot r18789: OUTPUT[escaping \' test 3ā¤]
moritz_ kp6: say "escaping \" test 3";
exp_evalbot r18789: OUTPUT[escaping " test 3ā¤]
moritz_ kp6: say "escaping \\\" test 5"; 17:28
exp_evalbot r18789: OUTPUT[escaping \" test 5ā¤]
pmurias moritz_: that bug breakes EmitPerl5 in garlic 17:29
moritz_ pmurias: which bug?
pmurias kp6: say 'escaping \' test 3'; 17:30
exp_evalbot r18789: OUTPUT[escaping \' test 3ā¤]
moritz_ ah, right
17:36 ujwalic joined
dlocaus snack & 17:38
17:42 ujwalic left
dlocaus yummm hummus and grapes! 17:48
moritz_ had "SpƤtzle", for which I don't know an English word 17:49
it's a bit like noodles, but they are thicker and contain eggs; usually served with cheese 17:50
yummy as well ;)
dlocaus opps, I need to avoid fat.
en.wikipedia.org/wiki/Sp%C3%A4tzle (Spatzle) 17:51
lambdabot Title: SpƤtzle - Wikipedia, the free encyclopedia
dlocaus yumm 17:52
looks good
moritz_ dlocaus: exactly that ;)
dlocaus wikipedia++ 17:53
moritz_ spƤtzle++ 17:54
17:54 justatheory joined 18:14 amnesiac joined 18:40 Psyche^ joined
pugs_svn r18790 | dlo++ | [KP6] Added Note about how to get around perl5's Debugger stack notice at 100 levels deep. 18:43
18:44 araujo joined 18:47 jjore-w left 18:53 cognominal_ joined 18:57 Psyche^ is now known as Patterner 19:02 ujwalic joined 19:03 ggoebel joined, barney joined
dlocaus pmurias: t/kp6/56-token-sym.t ping! 19:14
@seen pmurias 19:15
lambdabot pmurias is in #perl6. I last heard pmurias speak 1h 44m 31s ago.
dlocaus moritz_: ? 19:18
moritz_: ping
pmurias dlocaus: pong 19:20
dlocaus I've traced the problem to Runtime::Perl6;:Mutli
the Class "Multi" does not have add_token_variant implemented.
MultiToken does, however, when I tried substuting it, I got an "syntax" error 19:21
pmurias: dev.pugscode.org/browser/v6/v6-Kind...6/Multi.pm 19:22
lambdabot Title: /v6/v6-KindaPerl6/src/KindaPerl6/Runtime/Perl6/Multi.pm - Pugs - Trac, tinyurl.com/29l9ma
pmurias dlocaus: where/when do you get the syntax error? 19:23
pugs_svn r18791 | dlo++ | [kp6] Multi.add_token_variant is not implemented in Runtime::Perl6::Multi
dlocaus pmurias: one second, getting
pmurias i moved add_token_variant to MultiToken as multi token and multi sub are seperate things 19:24
dlocaus pmurias: I made the conversion from Multi to MultiToken in KindaPerl/Grammar/Sub.pm line 194 19:25
after re-compling I get this error
kitty:~/volunteer/perl6/pugs/v6/v6-KindaPerl6> !perl
perl script/kp6 t/kp6/56-token-sym.t
*** Syntax Error in Block: missing closing curly bracket }
error in Block at compiled/perl5-kp6-mp6/lib/KindaPerl6/Grammar/Sub.pm line 753, <> line 1.
19:28 BinGOs left 19:29 BinGOs joined
pmurias script/kp6 t/kp6/56-token-sym.t 19:32
gives me
not ok #...
19:33 JBoofy_ joined
dlocaus Sorry, I've updated the script to fail 19:33
19:34 BinGOs left
dlocaus reverting 19:35
reverted 19:36
pugs_svn r18792 | dlo++ | [kp6] reverted out changes
r18793 | dlo++ | [kp6] Removed contractions (English grammer) 19:37
19:39 cognominal_ left
mncharity ggoebel: pong 19:43
obra: I have to run in literally 15 min, but quickly,
In addition to current state of what exists, I'd be very interested in a characterization of the "negative space", the gaps where things are missing. 19:44
Eg, once upon a time we were stuck because only p6-hs could parse p6, its ast didn't include oo, and we were unclear how to implement oo in its backend. 19:45
and we didn't have a usable p6re engine, with which to create an alternate non-p6-hs frontend.
I'm not sure how much of all that has changed. If a .lot, we may be "unstuck, but not realizing it". 19:46
Eg, it may be that Moose is an ok oo model, the kp6 ast has the right nodes to handle the entire test suite, and we now have a fast p6re. 19:47
back to front, if Moose plus (MultiMethods-Pure or kp6 multimethods) is a right thing, what would it take to document and get development of backends on assorted platforms unstuck. 19:52
19:53 BinGOs joined
mncharity if the kp6 ast isn't quite ready for doing the whole test suite, what would it take to become so? Once is, can we do p6-hs ast <-> new kp6 ast (which would be the v2 ast we've been looking forward to for a couple of years now), conversion, and start driving backends with the non-oo entirety of t/. 19:54
do we now have a p6 rules engine which is both usable for development, and acceptably fast? if not, what kp6 tweaking would be needed to get one? are we using all the 5.latest goodness? 19:56
is STD something we think can be a next step implementation, or do we want to use it more as documentation, and do something simpler, half way inbetween existing impls and STD. 19:57
19:57 jisom joined 19:58 BinGOs left
mncharity what would be the minimum tweaking needed to get kp6 to run against the test suite? each time we spin up a new impl, progress is much faster once we can bang on t/. what would it take to get a kp6 derivative there. 19:58
19:59 BinGOs joined
mncharity timeout. bbl & 19:59
(btw, I'm told/know that I sometimes have a negative tone when doing design space analysis. I'm actually tremendously grateful and happy with fglocks heroic efforts, everyone elses work, obra doing the onion, audreyt breathtaking contribution, TimToady's perseverence, patience and insight, and so many others. P6 is a great project to have some connection to, and one I think very important for the soft. eng. and the world.) 20:02
re world, if anyone thinks I'm exagerating, OLPC has the potential for a major impact on 100 M's of kids' lives. and every day I wish p6 had been ready for primetime. 20:03
&
20:04 BinGOs left 20:09 BinGOs joined 20:11 ujwalic left
pugs_svn r18794 | avar++ | Hack to make 5 more tests pass 20:20
moritz_ re 20:26
20:36 takanori joined
pmurias moritz_: re 20:40
20:42 BinGOs_ joined, jisom left, BinGOs left, BinGOs_ is now known as BinGOs 20:43 thoughtpolice joined
pmurias avar: kp6 also does what your hack does 20:45
moritz_ sounds.. hacky 20:46
dlocaus pmurias: moritz_: well, is there anything you guys want me to work on? 20:47
pmurias dlocaus: is there anything you want to hack on?
dlocaus Documentation would probably be good. 20:48
I realize a lot of this code will be tossed once onion comes into being, but it would be good to get some of onion? documentation going? 20:49
will onion be the final release?
avar pmurias: good to know
dlocaus: some of it will be used for it at least
pmurias dlocaus: i don't know
moritz_ dlocaus: "final" sounds very... final ;) - I don't think it will be final 20:50
dlocaus mp->kp6->onion->??? ;)
pmurias onion won't toss so much code
dlocaus Well, I can refactor code if you want.
I've been working on refactoring script/kp6 (I want to test it a bit more though...)
dlocaus wants to help!!! :) 20:51
pmurias script/kp6 is very ugly
dlocaus I made it a lot less ugly :) 20:52
moritz_ dlocaus: then commit it, and let everybody else test it ;)
dlocaus I was thinking about sorting out the -st & stdout issue, but I wasn't sure that wouldn't be disrupptive to the end project.
pmurias dlocaus: i'm too confused by the MultiToken giving parse failures 20:53
dlocaus: you could try to work on kp6-kp6 20:54
dlocaus kp6-kp6?
pmurias wait
pugs_svn r18795 | dlo++ | [kp6] I have refactored script/kp6, I bottomed out a lot of the exit codes, attempted to clean up the documentation, while preserving the core of the code, please revert if I screwed up, and please send me the command that got refactored out dlocaus #perl6 irc.freenode.net
dlocaus I saw src-script/kp6-kp6, but I didn't know what to make of it. 20:55
moritz_ dlocaus: kp6-kp6 is what you get when you compile kp6 with kp6 instead of mp6
dlocaus crriick!
moritz_: doesn't kp6 using mp6 to compile kp6-kp6?
moritz_ now we're in that bootstapping discussion again ;) 20:56
dlocaus aye, that was the "crashing" sound in my head again.
moritz_ ussually when you compile kp6, you use mp6 for most of it, and kp6 for Runtime::Perl6::* 20:57
20:57 the_dormant joined
moritz_ now what we call kp6-kp6 is when you compile all of kp6 with kp6 20:57
it compiles iirc, but it doesn't run yet
dlocaus moritz_: I'm not sure I'm ready for that, I'm still working on learning the perl6 language. 20:58
ok, so we use kp6 w/ mp6 to compile kp6 into perl5 code? 20:59
moritz_ dlocaus: in which step do you mean? normal kp6-mp6?
pmurias in kp6-kp6 we don't use mp6 at all 21:00
moritz_ right
because kp6 is a fully compiled perl5 image
dlocaus wouldn't it be
perl script/kp -o kp6.pl src-script/kp6-kp6
pmurias yes 21:01
dlocaus where kp6.pl is a completely perl5 code.
moritz_ dlocaus: exactly
dlocaus: so no mp6 involved anymore
dlocaus urk!
moritz_ which is why we want the bootstrapping ;)
dlocaus You mean that everyone down wind of us, has to use the ::DISPATCH code?
avar no
dlocaus the perl5 code produced by script/kp6 is a bunch of ::DISPATCH calls... 21:02
lisppaste3 avar pasted "kp6 breakage" at paste.lisp.org/display/50339
21:02 drupek12 joined
dlocaus avar: ok, I'll check into that 21:02
avar oh it's because of a svn conflict 21:03
works now
avar likes lisppaste
They have a xml-rpc interface unlike sial.org so I can do M-x lisppaste-region in emacs easily:) 21:04
lisppaste3 dlocaus annotated #50339 with "it worked for me... " at paste.lisp.org/display/50339#1
avar yeah, my bad
dlocaus avar: ok
moritz_ avar: sial has a command line interface
they provide a client script, I think it's perl
avar I know, but the xml-rpc stuff I can do in pure-elisp 21:05
dlocaus moritz_: pmurias: I'm going to try and come up with an document that explains how the code works. IE, who does want and where. If I'm lucky, I'll get "why" in there too. :) 21:07
21:08 baalial joined
dlocaus The "big picture" of kp6. Which, is something I haven't been able to get a hold of in my head yet. 21:08
21:08 pjcj joined
moritz_ not either 21:08
21:10 Aankhen`` left
lisppaste3 avar pasted "What's the 't' in ast & perl5 output all about?" at paste.lisp.org/display/50343 21:24
dlocaus what does CompUnit mean?
CompileUnit?
moritz_ yes 21:25
avar "compilation unit"
moritz_ which is a module, package or class
dlocaus ok 21:28
is that "compilation unit" Compiled? or need to be compiled?
avar It really means it can be stand-alonein some aspects, check out -Ccl for a better exmple than -Cperl5 21:29
@tell fglock "What's the 't' in ast & perl5 output all about?" at paste.lisp.org/display/50343
lambdabot Consider it noted.
moritz_ are you talking about the AST?
avar The general thing, where does this t come from?
but yes, the ast
pmurias avar: you want to compile 't/kp6/21-oo.t' 21:34
?
i think you put the -e by mistake 21:35
avar haha, right
thanks:)
pmurias the polite response to "thanks" is? 21:36
moritz_ "you're welcome" 21:37
pmurias thanks ;)
moritz_ or "yw" for short
21:39 barney left
pmurias dlocaus: the syntax errors in token proto come from inside COMPILER::begin_block($bind) 21:40
how are multi methods inherited as normal ones? 21:42
dlocaus pmurias: looking 21:44
pmurias: as for the multi methods part, I don't know. I'm still trying to grasp the big picture part. 21:45
pmurias * s/inherited/inherited,/ - in p6 i mean (not in kp6) 21:47
moritz_ pmurias: could you rephrase your question?
I'd think "just like any other method"
21:48 Jamtech joined
moritz_ just the type of the invocant changes to that of the child class 21:48
pmurias is multi method inheritance the same as only method inheritance?
moritz_ yes, I think so
pmurias so one would have to add multi method visit($node) {callsame} at the end in a multi method based visitor 21:51
for example in Perl5Token 21:52
21:52 weinig left, weinig joined
pmurias dlocaus: you might like fixing Emit::Perl6 (i don't think there are any hidden pitfalls in that task) 21:56
although you like writing a code overview more, do it 21:57
22:00 drupek1253 joined
avar urgh, proto objects 22:00
dlocaus pmurias: what does Emit::Perl6 need fixing? 22:02
I'm attempting to find that compilation error... (it would help if I could read perl6)
;)
avar what error? 22:05
pmurias Emit::Perl6 should turn the ast back into valid perl6 code
avar should kp6->emitperl6->.p6->kp6 produce the same AST ? 22:06
dlocaus Humm, ok. I think I'm some where in the middle of the this bug, but I'll check into p6->kp6->p6.
pmurias avar: yes 22:07
sleep& 22:12
22:12 pmurias left 22:14 drupek12 left
dlocaus is class Main { ... } a requirement for perl6? 22:15
moritz_ no
dlocaus thanks
moritz_ it's a requirement for mp6 though, iirc 22:16
mncharity dlocaus: a "wrap your head around it" intro (blog post?) to kp6/mp6 would be nifty. 22:17
dlocaus mncharity: aye, I'll be working on that.
I've finally found what was causing the MultiToken problem, or rather, I've found how to "properly" invoke it. 22:18
proto token xyz { };
The compilier doesn't like Empty brakets
proto token xyz { 1; };
this works, sort of.
well, it compiles
mncharity *caveat, _I'd think_ would be nifty. My head is as yet unwrapped, so I'm hypothesizing.
dlocaus mncharity: some of the problem with modifing this code, is you don't know where it came from. 22:20
dev.pugscode.org/browser/v6/v6-Kind...Grammar.pm 22:23
lambdabot Title: /v6/v6-KindaPerl6/compiled/perl5-kp6-mp6/lib/KindaPerl6/Grammar.pm - Pugs - Trac, tinyurl.com/25fh8q
dlocaus I was stuffing my head into this file.
mncharity that first step of getting familiar with an unfamiliar project is often a challenge / barrier to entry.
dlocaus That code, I think is perl6 code produced by mp6.
mncharity re "stuffing my head into this file", hmm, you might try the reverse...
dlocaus sorry, is perl5 code produced produced by mp6, which I think is a translation from perl6. 22:24
mncharity: I don't think I can pull the reverse off.
The file is just too complicated.
avar perl5.9.5 script/kp6 -Cast t/kp6/21-oo.t <= fails 22:26
moritz_ avar: does it work with perl 5.8? 22:27
dlocaus I have to pick up my dog (teeth cleaning) & 22:32
mncharity dlocaus: re complicated, for mechanically generated code it helps if you approach it systematically. eg, if you don't have an editor which can show you just the top level sub names, then grep them out, and start an analysis doc. for boilerplate, consider bulk s///ing it away. And then block by block, go through either adding comments, or discarding code for clear pseudo code. Eventually get a handle on whats going on. And start t
dlocaus mncharity: go ahead, I'll read up on the log when I get back :) 22:33
22:33 iblechbot left
pugs_svn r18796 | avar++ | [kp6] this print_perl() never worked on installations without Perl::Tidy 22:33
avar fixed it
moritz_: yes
22:34 baalial left
mncharity actually, given that Grammar.pm is generated, first step should probably to look at the file it was generated _from_, no? 22:35
avar eek! 22:41
dlocaus: Do your manual changes to compiled/perl5-kp6-mp6/lib/KindaPerl6/Visitor/Emit/Perl5.pm need to be kept? 22:42
dlocaus: That "free unreferenced scalar" thing is a bug in the perl interpreter, it's fixed in the 5.8.9 branch and 5.10
pugs_svn r18797 | avar++ | [kp6] Emit some comments to make ::Method() easier to read in the emitted code. this overwrites some MANUAL CHANGES to the emitted file by dlocaus which probably don't need to be there 22:44
avar caps for the changelogs
22:45 weinig left 22:46 weinig joined 23:03 bbkr left 23:07 bbkr joined 23:09 TJCRI left 23:14 Jamtech left, the_dormant left 23:30 BinGOs left
dlocaus avar: don't worry about it 23:32
avar: its just some code I put in there to help pmurias / fglock find a bug. 23:33
23:33 amnesiac left
dlocaus avar: I can add it in if the bug hasn't gone away 23:34
and I'm away again, pick up my dad at the airport.
23:35 Limbic_Region joined 23:51 fglock 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!