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.
alester Trying to aggregate the stuff we already know, and point to what we don't www.perlfoundation.org/perl6/index....e_released 03:56
lambdabot Title: When will Perl 6 be released? / Perl 6, tinyurl.com/37k2wt
pasteling "avar" at 212.30.222.56 pasted "Progblem with Optimize::RuleConstantFold" (17 lines, 1K) at sial.org/pbot/29512 06:20
"avar" at 212.30.222.56 pasted "Progblem with Optimize::RuleConstantFold (emitted code)" (9 lines, 361B) at sial.org/pbot/29513
avar-2 perl script/kp6 --do MetaClass,Global,Emit::AstPerl -e "token opt_ws { ' ' | '' }" | perltidy <= better 06:22
ah 06:23
avar-2 argh, nothing works 07:44
pmurias avar: fixed ;) 08:06
pugs_svn r19294 | pmurias++ | [kp6] fixed bug reported today by avar++ with '' being wrongly handled by Optimize::RuleConstantFold
pugs_svn r19295 | avar++ | Don't fail when the pugs t directory isn't in ../../ 08:12
pmurias avar: if you want you could propably add RuleConstantFold to the default workflow 08:20
avar I'm testing that 08:26
pmurias i don't think a -Olevel option is needed yet 08:27
avar the current stuff is running with mp6, can't it run on kp6? 08:53
I'm wondering how much faster the kp6-kp6 parser is with the optimization but I'm not able to run it 08:54
pmurias avar: i haven't timed it but the kp6-kp6 parser seems much faster 09:04
but the kp6-kp6 visitors are horribly slow 09:05
avar how do you run it?
pmurias script/kp6-kp6.pl
i have started work on an alterantive perl5 backend Emit::Perl5V6 09:06
avar: to (re)compile kp6-kp6 you need to pass the KP6_TARGET='KP6-KP6' option to Makefile.pl 09:07
avar $ perl -I compiled/perl5-kp6-kp6/lib script/kp6-kp6.pl 09:12
KindaPerl6/Runtime/Perl6/Pair.pm did not return a true value at compiled/perl5-kp6-kp6/lib/KindaPerl6/Runtime/Perl5/Runtime.pm line 13.
pmurias avar: did you recompile kp6? 09:13
KP6_TARGET should be 'KP6-BOOT' 09:14
s/pass (.*) option/pass $1 the enviornment variable/ 09:15
pasteling "avar" at 208.78.101.240 pasted "boot failure" (211 lines, 19.7K) at sial.org/pbot/29515
pmurias the Makefile's broken it treats a perl5 file as perl6 one 09:17
want to fix it? 09:20
avar nope:)
pmurias kp6-kp6 parser is a bit broken so you won't have much use for kp6-kp6 09:21
pmurias avar: kp6 parses my declarations wrongly 09:34
pugs_svn r19296 | pmurias++ | [kp6-kp6] fixed the Makefile,updated the image
pmurias * kp6-kp6
aaron_TeeJay bah 09:53
naturally pugs requires a newer version of haskell than the one I downloaded and compiled for 48 hours last time I tried pugs 09:54
Tene What version did you use last time?
aaron_TeeJay 6.6
now I need 6.6.1
anybody have recent svn pugs and want to let me know if examples/hanoi works (without // and/or argv) ? 09:55
update www.perlfoundation.org/perl6/index....r_of_hanoi if you do
lambdabot Title: Tower of Hanoi / Perl 6
aaron_TeeJay thanks
Tene aaron_TeeJay: works for me 09:56
aaron_TeeJay which svn revision?
and with // and argv ?
can you benchmark it too?#
Tene looks like I have 18224 on that box, but I'm not certain I've recompiled with that version 09:57
Tell me what to run to produce the benchmarking info you want.
aaron_TeeJay ext/Benchmark/lib/Benchmark.pm 10:00
1000 iterations of 6 discs? 10:01
which svn revision do you have installed?
Tene 18224 10:02
aaron_TeeJay: how do you want me to run those 1000 iterations? modify hanoi.pl or run the existing hanoi.pl 1000 times, with full startup costs, or what? 10:03
aaron_TeeJay modify
no point benchmarking startup costs 10:04
thanks
Tene aaron_TeeJay: with or without the 'say' output to stdout? 10:05
aaron_TeeJay skip say 10:06
ideally, print to a file instead 10:07
but that's something else to worry about
can always improve later
thanks 10:08
1 down, 50 odd to go
Tene Running on pugs...
parrot completed it in:
real 0m3.577s 10:09
user 0m3.119s
sys 0m0.452s
aaron_TeeJay parrot?
Tene Waiting on pugs.
aaron_TeeJay what svn for parrot?
Tene Current.
pasteling "Tene" at 166.70.38.237 pasted "hanoi-bench" (13 lines, 200B) at sial.org/pbot/29516 10:10
aaron_TeeJay pugs still going 10:11
Tene Yep.
aaron_TeeJay ouch 10:12
Tene I'll kill it and re-do it, printing the iteration each time.
aaron_TeeJay try 100 instead
or even 10 10:13
Tene pugs:
real 0m22.611s
user 0m22.384s
sys 0m0.101s
parrot:
real 0m0.841s
user 0m0.694s
sys 0m0.141s
aaron_TeeJay how many iterations for pugs?
Tene 100, both with a 'say' for the iteration number. 10:14
aaron_TeeJay right
Tene also on completely different boxes
So not an actual usable comparison
I had haskell issues compiling pugs on my laptop that I didn't track down. 10:15
aaron_TeeJay thanks
:)
Tene No problem.
aaron_TeeJay at least it gives ballpark figures
Tene Yeah.
aaron_TeeJay what kind of machine for pugs vs parrot ? 10:16
Tene pugs: 1ghz, 640M ram 10:17
exp_evalbot OUTPUT[*** ␤ Unexpected "ghz"␤ expecting "_", fraction, exponent, term postfix or operator␤ at /tmp/iPdrXKywJB line 1, column 2␤]
aaron_TeeJay oops :)
Tene parrot: 1.6ghz, 2G ram
pugs: [~] <m oo se>
exp_evalbot RESULT["moose"]
aaron_TeeJay so the h/w performance wouldn't alone be the cause of the difference of that scale then :)
Tene Right. 10:18
Oh, right, I'll update parrot on that box and run it there.
aaron_TeeJay neat
can you update www.perlfoundation.org/perl6/index....r_of_hanoi when you're done
lambdabot Title: Tower of Hanoi / Perl 6
Tene has been up a bit too long.
aaron_TeeJay don't worry about it then
:)
aaron_TeeJay ought to do some Work
i.e. stuff that pays mortgage 10:19
Tene What do you want me to add?
aaron_TeeJay don't worry about it
:)
have benchmarks, machines, code - I think that page is "done"
people can always add newer benchmarks, etc
I'll look at doing hilbert curve next weekend 10:20
Tene Oh, you added stuff. I should try reading. ;)
aaron_TeeJay :)
Tene Can you point me to a hilbert impl? 10:21
aaron_TeeJay linked from hilbert page iirc ;)
but not done in perl6 yet
Tene Ahh. 10:22
aaron_TeeJay it's implemented in every language :)
except perl 6 and possibly some obscure assembly
it's another interesting recursion problem
but you can add layers of efficiency/complexity 10:23
very much a "solve on paper first" one
right - work
laters all
Tene Or a "steal from the p5 impl" one
aaron_TeeJay thanks for help
I wrote the p5 impl :) 10:24
Tene What is it you do for work, btw?
seeya.
aaron_TeeJay fuck-off-big online classifieds
so that even on a sunday morning the site is still being used by dozens of users
Tene Nice.
aaron_TeeJay which is annoying for maintainence
but lots of interesting problems in scaling and high availability 10:25
Tene Rather.
aaron_TeeJay yes - that's why I took the job 10:26
that and working from home
:)
aaron_TeeJay puts 30 mins "research" on timesheet
Tene aaron_TeeJay: parrot, 100 revs, same box as pugs: 10:41
real 0m2.007s
user 0m1.939s
sys 0m0.045s
ruoso hmm.. some irc log is lost... 17:50
too bad 17:51
LimbicRegion w 18:55
zev I'm looking to do some work on perl6. I've spent some time looking through the existing code, and want to start on some small tasks. I don't see anything in the perl6 hiveminder group. Does anyone know of anything smallish that needs working on right now? 20:14
[particle] perl 6 in general, or perl6 the parrot app? 20:15
zev perl6 the parrot app
(or things in parrot or pct that need doing for perl6) 20:16
[particle] hop on over to irc.perl.org#parrot, you're more likely to get answers there. but in general, we need help reorganizing the tests in pugs to align with the synopses
zev ok. I'll ask over there 20:19
pugs_svn r19297 | zev++ | * added myself to AUTHORS (commit bit test) 22:02
[particle] pugs: say :8<10> 22:15
exp_evalbot OUTPUT[8␤]
[particle] pugs: say :2<10>
exp_evalbot OUTPUT[2␤]
avar zev: welcome!:) 22:18
zev thanks 22:30
cognominal_ test 23:25
acmoore OK 1 23:29
markstos Is pugs still alive? pugs.blogs.com/ is fairly silent on it, and steam seems to be picking up with the parrot backend. 23:32
lambdabot Title: Pugs
[particle] markstos: pugs is all but abandoned, as nobody's made any pugs commits in a long time 23:35
markstos Is the Haskell backend not consided viable, then? I heard Audrey become ill for a while, and that was a factor. 23:36
I was just starting to learn some Haskell. :) 23:37
[particle] it's viable, as far as i know. just lost steam when audrey got Real Life 23:38
it's been waiting on a new object model, and nobody's taken up the work of merging audrey's unfinished branch with trunk
(by branch i mean svk depot somewhere) 23:39
markstos Ah, sounds like sort of an important and complex blocker, then.
markstos [particle]: thanks for the update. Maybe I'll poke at some pugs some once my Haskell gets off the ground. 23:40
cognominal_ Bonne année. Happy new year 23:43

Powered by Cro and the Raku® Programming Language.

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

Thank you!