»ö« Welcome to Perl 6! | perl6.org/ | evalbot usage: 'perl6: say 3;' or rakudo:, niecza:, std:, or /msg camelia perl6: ... | irclog: irc.perl6.org | UTF-8 is our friend!
Set by sorear on 25 June 2013.
lizmat grrrrrrr 00:00
java.lang.NullPointerException
in (src/Perl6/Optimizer.nqp:532)
calling it a night: last->done, fail->catch on taps will have to wait 00:01
sleep&
00:02 rurban1 left
jnthn sleeps also...gotta be up in 5 and a bit hours... :/ 00:05
&
japhb wonders what p6store is
00:10 woolfy left 00:13 woolfy joined 00:23 eternaleye joined 00:34 stevan_ left 00:35 stevan_ joined 00:42 rurban1 joined, thou left
dwarring rn: grammar G1 {rule TOP {[ x ]**2..3}}; say ~G1.parse("x x x x x x"); 00:45
camelia rakudo-parrot 180fca, rakudo-jvm 180fca: OUTPUT«x x x ␤»
..niecza v24-98-g473bd20: OUTPUT«Use of uninitialized value in string context␤ at /home/p6eval/niecza/lib/CORE.setting line 1389 (warn @ 5) ␤ at /home/p6eval/niecza/lib/CORE.setting line 269 (Mu.Str @ 15) ␤ at <unknown> line 0 (ExitRunloop @ 0) ␤ at /tmp/BMJgypOUu3 line 1 …»
dwarring r: grammar G2 {rule TOP {:my $narity = 2..3; [ x ]**$narity}}; say ~G2.parse("x x x x x x");
camelia rakudo-jvm 180fca: OUTPUT«===SORRY!=== Error while compiling /tmp/zODuXtac0I␤Quantifier quantifies nothing␤at /tmp/zODuXtac0I:1␤------> 2 {rule TOP {:my $narity = 2..3; [ x ]**⏏$narity}}; say ~G2.parse("x x x x x x");␤ expecting any …»
..rakudo-parrot 180fca: OUTPUT«===SORRY!=== Error while compiling /tmp/hDqxi4p2zg␤Quantifier quantifies nothing␤at /tmp/hDqxi4p2zg:1␤------> 2 {rule TOP {:my $narity = 2..3; [ x ]**⏏$narity}}; say ~G2.parse("x x x x x x");␤ expecting a…»
dwarring r: grammar G3 {rule TOP {:my $narity = 2..3; [ x ]**{$narity}}}; say ~G3.parse("x x x x x x"); 00:46
camelia rakudo-jvm 180fca: OUTPUT«===SORRY!=== Error while compiling /tmp/jIa66xgG9d␤Block case of ** quantifier not yet implemented␤at /tmp/jIa66xgG9d:1␤------> 3 {rule TOP {:my $narity = 2..3; [ x ]**⏏{$narity}}}; say ~G3.parse("x x x x x x"␤…»
..rakudo-parrot 180fca: OUTPUT«===SORRY!=== Error while compiling /tmp/Csn36D9DeB␤Block case of ** quantifier not yet implemented␤at /tmp/Csn36D9DeB:1␤------> 3 {rule TOP {:my $narity = 2..3; [ x ]**⏏{$narity}}}; say ~G3.parse("x x x x x x"…»
dwarring trying to compose regex ranges on the fly... 00:47
dwarring thinking G3 should do it, but its nyi
00:49 stevan_ left 00:50 btyler joined
dwarring or is G2 closest? 00:52
00:59 stevan_ joined 01:03 stevan_ left 01:06 huf_ joined 01:07 huf left
dwarring r: grammar G4 {rule TOP {:my $n = 2;:my $m = 3; [ x ]**{$n..$m}}}; say ~G4.parse("x x x x x x"); 01:09
camelia rakudo-parrot 180fca: OUTPUT«===SORRY!=== Error while compiling /tmp/_zsUGWWi40␤Block case of ** quantifier not yet implemented␤at /tmp/_zsUGWWi40:1␤------> rule TOP {:my $n = 2;:my $m = 3; [ x ]**⏏{$n..$m}}}; say ~G4.parse("x x x x x x")…»
..rakudo-jvm 180fca: OUTPUT«===SORRY!=== Error while compiling /tmp/r5kBg20KJp␤Block case of ** quantifier not yet implemented␤at /tmp/r5kBg20KJp:1␤------> rule TOP {:my $n = 2;:my $m = 3; [ x ]**⏏{$n..$m}}}; say ~G4.parse("x x x x x x")␤…»
01:09 btyler left 01:14 tgt left
dwarring it may not be pretty, but I thinks there's a work around... 01:17
grammar G5 {rule TOP {:my $n = 2;:my $m = 3;:my $i; [ x <?{$i++ < $m}> ]* <?{ $i >= $n}>}}; say ~G5.parse("x x x x x x");
r: grammar G5 {rule TOP {:my $n = 2;:my $m = 3;:my $i; [ x <?{$i++ < $m}> ]* <?{ $i >= $n}>}}; say ~G5.parse("x x x x x x"); 01:18
camelia rakudo-parrot 180fca, rakudo-jvm 180fca: OUTPUT«x x x ␤»
dwarring not pretty, but it'll do for the moment
01:19 raiph left 01:22 raiph joined 01:31 jferrero left, jferrero joined 01:41 eternaleye left 01:58 benabik joined 01:59 Vlavv` left
BenGoldberg r: grammar G5 {rule TOP {:my \N = 2;:my \M = 3;:my $i; [ x <?{$i++ < M}> ]* <?{ $i >= N}>}}; say ~G5.parse("x x x x x x"); 02:05
camelia rakudo-parrot 180fca, rakudo-jvm 180fca: OUTPUT«x x x ␤»
BenGoldberg r: grammar G5 {rule TOP {:my \N = 2;:my \M = 3;:my \i; [ x <?{i++ < M}> ]* <?{ i >= N}>}}; say ~G5.parse("x x x x x x"); 02:06
camelia rakudo-parrot 180fca, rakudo-jvm 180fca: OUTPUT«===SORRY!===␤No such method 'ast' for invocant of type 'NQPMu'␤»
BenGoldberg r: grammar G5 {rule TOP {:my \N = 2;:my \M = 3;:my \i = 0; [ x <?{i++ < M}> ]* <?{ i >= N}>}}; say ~G5.parse("x x x x x x");
camelia rakudo-jvm 180fca: OUTPUT«Cannot modify an immutable value␤ in sub postfix:<++> at gen/jvm/CORE.setting:4265␤ in regex TOP at /tmp/WNSZ9ZRf3V:1␤ in method parse at gen/jvm/CORE.setting:12302␤ in block at /tmp/WNSZ9ZRf3V:1␤ in any eval at gen/jvm/stage2/NQPHLL.nqp:108…»
..rakudo-parrot 180fca: OUTPUT«Cannot modify an immutable value␤ in sub postfix:<++> at gen/parrot/CORE.setting:4276␤ in sub postfix:<++> at gen/parrot/CORE.setting:1739␤ in regex TOP at /tmp/O2jSuaI8xi:1␤ in method parse at gen/parrot/CORE.setting:12333␤ in block at /…»
02:14 Vlavv` joined 02:23 rurban1 left 02:42 rurban1 joined 03:08 raiph left 03:15 eternaleye joined 03:18 raiph joined 03:37 davecc left 03:38 daveec left 03:41 daveec joined, davecc joined 03:58 stevan_ joined 04:02 jnap left 04:03 fridim_ joined 04:09 preflex_ joined, ChanServ sets mode: +v preflex_ 04:10 rurban1 left, preflex left, preflex_ is now known as preflex 04:12 FROGGS left 04:16 stevan_ left, stevan_ joined 04:32 rurban1 joined 04:39 eternaleye left 04:42 eternaleye joined 04:48 daveccc joined 04:51 daveccc left 04:56 berekuk joined 04:57 kaleem joined 04:58 BenGoldberg left 05:07 berekuk left 05:15 berekuk joined 05:22 odoacre joined 05:23 [Sno] left 05:29 berekuk left 05:36 kaleem left 05:46 berekuk joined 05:48 Entonian joined
moritz good morning, #perl6 05:48
05:54 Entonian left 06:03 Entonian joined 06:05 Entonian left 06:07 berekuk left 06:08 Entonian joined, Entonian left 06:10 Entonian joined, Entonian left 06:14 Entonian joined 06:17 Entonian left 06:23 Entonian joined 06:24 Entonian left 06:31 [Sno] joined 06:40 kaleem joined 06:41 Entonian joined 06:42 Entonian left 06:43 Entonian joined 06:46 rindolf joined 06:51 Entonian left 06:52 Entonian joined 06:53 SamuraiJack joined, rindolf left 06:54 Entonian left, rindolf joined, Entonian joined 06:57 rindolf left, rindolf joined 06:58 rurban1 left 07:01 Entonian left 07:06 Entonian joined, Entonian left 07:07 Entonian joined, Entonian left 07:08 Entonian joined, rindolf left, rindolf joined, Entonian left 07:09 Entonian joined 07:12 Entonian left, Entonian joined 07:14 Entonian left, Entonian joined, FROGGS joined
FROGGS japhb: I think it is about storing the value of one container in another 07:17
07:19 lizmat_ joined
FROGGS morning moritz 07:20
07:22 lizmat left 07:23 hummeleB1 joined 07:24 Entonian left 07:28 rurban1 joined 07:30 kivutar left 07:33 rurban1 left 07:35 PZt left 07:36 hummeleB1 left 07:37 hummeleB1 joined 07:40 berekuk joined 07:48 fridim_ left 07:55 berekuk left 08:02 yves left 08:14 nnunley left 08:15 zakharyas joined 08:31 rurban1 joined 08:35 rurban1 left 08:56 dmol joined
timotimo o/ 09:09
dwarring: it may be safer to compare the count of a named match group after the match rather than using ++ inside, because backtracking may bite you
hoelzro damn, I always miss late-night messages =/ 09:11
(good morning, #perl6)
masak good morning hoelzro, #perl6 09:13
hoelzro ahoy masak 09:14
FROGGS o/ 09:16
timotimo: I think that might be due to your patch: 09:17
t/spec/S05-modifier/perl5_5.rakudo.parrot (Wstat: 0 Tests: 102 Failed: 3)
Failed tests: 5, 7, 9
timotimo i think so, too. good thing i didn't merge it into nom.
FROGGS timotimo: I am on nom
timotimo oh?!
FROGGS (using nqp HEAD) 09:18
timotimo that's bad
which one do you think it is? the performance tweaks? 09:19
FROGGS something after rakudo 2013.10-235-g180fcac and nqp 2013.10-254-gaf22bef
my nqp-p is 2013.10-257-gc293015 09:20
rakudo is the same version
so it only can be either bdbb63a19ff37e09dd7ac15fa6a93c56f92876f6 or c2930155bd14d95e93576c9cc7e5e2209050c2e4 09:21
timotimo i'll test the former, you test the latter?
FROGGS k
umm 09:22
it can't be c29...
because that is for jvm only
timotimo oh, right
i'm testing the other commit anyway
FROGGS I am testing af22bef4a382ba5a4eb887a407ff154c4de2d73c, which should fix the regression 09:23
timotimo good 09:24
it could be that $I12 is used for something else and is required to stay unchanged throughout other regex ops
09:25 raiph left
timotimo last night before bed i was thinking: findcclass probably has to rescan the whole file up to each newline for each of the newlines 09:25
if there's multibyte characters in it 09:26
that would explain the crazy slowness of the lineposcache in actions and grammar
09:29 xinming joined 09:30 iSlug joined
timotimo now i thought if findcclass (and findnotcclass and friends) would cache what string they were last used on, what byte position they were at and what character position it corresponds to, it could perhaps immediately jump very close to the last newline it found rather than scanning where the "begin" character is at in the string 09:31
09:32 dakkar joined, rurban1 joined, xinming_ left 09:34 xinming left, xinming joined 09:36 rurban1 left
FROGGS reverting it fixed it btw 09:38
timotimo: does is store any other information that you might need to cache? 09:39
moritz timotimo: I don't think the caching is very benefitial, because the regex engine uses substr so often that it won't have many cache hits 09:45
timotimo but i'm talking only about findcclass 09:56
FROGGS I think moritz meant that findcclass will already get a substr 09:58
so it doesn't have to rescan that much
timotimo i'm testing just the $I12 part reverted of that commit
oh, it will?
but that's kinda wasteful
FROGGS I dunno
I am just a mediator :o) 09:59
timotimo i don't have time to implement that cache until ... hmm ... in about 5 hours? 10:01
FROGGS timotimo: it is not a bad thing to implement (experimental) stuff after a release :o) 10:02
because it is always nice to have a module test from the codebase of the release 10:03
timotimo yes, true.
FROGGS that's one reason I made the spawn changes right after the release 10:04
timotimo yeah, the partial revert fixes the perl5 regexes 10:09
dalek p: 4aea5a1 | (Timo Paulssen)++ | src/vm/parrot/QAST/Compiler.nqp:
Partially revert "tiny performance tweaks in the parrot regex compiler"

This reverts parts of commit bdbb63a19ff37e09dd7ac15fa6a93c56f92876f6.
Apparently $I12 is required to not change its value in here.
FROGGS cool 10:12
timotimo++ 10:13
timotimo i don't think i deserve positive karma for this :P 10:16
moritz oh yes you do
admitting and fixing your own mistake is a very important trait
in the long term much more important than getting things right in the first place, which you can't always achieve 10:17
FROGGS true :o) 10:18
timotimo mhh okay :) 10:21
10:22 xinming left 10:23 fhelmberger joined, xinming joined 10:32 Mouq left, rurban1 joined 10:33 Mouq joined
woolfy Saturday 23 November: Nordic Perl Workshop in Copenhagen, Denmark. Sponsored by the Perl 6 Community! Yay! act.yapc.eu/npw2013/index.html 10:34
10:34 SamuraiJack left 10:36 ssutch left
FROGGS woolfy++ # thank you! 10:36
woolfy Thanks to the Perl 6 Community. Great people!
10:37 rurban1 left
timotimo I'll be afk for a bit now 10:50
10:52 kivutar joined 10:54 SamuraiJack joined 10:59 tgt joined 11:00 felher joined 11:04 Ulti_ is now known as Ulti
FROGGS lizmat_: latest rakudo + latest nqp passes p-spectests 11:49
11:49 xinming left
moritz imagines the p-fleet flying by 11:51
11:51 kivutar left, xinming joined
tadzik by a p-class planet 11:51
FROGGS (class M)++, even if the weather is not very nice today 11:54
feels more like class L atm 11:55
tadzik it's funny because M in class M can stand for Moar :P
FROGGS :o)
coincidence? NO! 11:56
12:09 FROGGS left
timotimo should WINNER go into the new compiler release? or should it keep the winner sub that takes lists of pairs 12:15
?
12:16 stevan_ left 12:17 colomon left
timotimo is going to look into making winner * work again as expected 12:20
12:22 tipdbmp left 12:28 stevan_ joined
timotimo is browsing the irclogs to figure out when he started being involved with perl6 12:29
12:29 tipdbmp joined 12:31 stevan_ left
moritz irclog.perlgeek.de/perl6/2012-01-25#i_5048758 around here, it seems 12:35
timotimo that's the conclusion i've come to, as well 12:36
12:36 FROGGS[mobile] joined
timotimo my actual contributions started about 10 month ago. 12:36
(there were a few typed exception commits 1 year 8 months ago, though)
FROGGS[mobile] I am here for 13 months now 12:39
lizmat_ is finally awake again after fitness, swimming, a nice breakfast and a nap that took a little longer than expected 12:42
12:42 lizmat_ is now known as lizmat
woolfy A lot longer... 12:42
lizmat FROGGS: thank you
timotimo hi liz :) 12:43
12:44 iSlug left
lizmat hi timotimo! 12:51
timotimo i'm bringing WINNER up to speed with the timeout for "wait". do you think it should go in the release or not? 12:52
lizmat yes, in the release please
timotimo since rakudo-jvm isn't that important for star yet, i don't think it'd be terrible
how much time do i have left to make it work and pretty? :) 12:53
lizmat it's fluid still, but the more people start using it now, the better
I'm a little behind on my own schedule, so I'll probably cut the release sometime after 11pm tonight
so about 9 hours from now
timotimo ah, that sounds doable
lizmat still want to see something of Copenhagen by daylight 12:54
FROGGS[mobile] daylight? why that? :p 12:56
timotimo what is daylight? 12:57
lizmat It's that thing between nights :-)
timotimo oh, you mean when the evil day star is out?
lizmat or on, depending on pov 12:58
timotimo :) 12:59
12:59 huf_ left 13:01 huf joined
lizmat hmmm... seems like the nqp release guide doesn't mention anything about jvm 13:06
moritz yes, too old for that 13:07
lizmat should I just ignore jvm while cutting the release
moritz yes
lizmat I mean, it's one tarbal in the end, right
ok
I will :-)
13:10 MikeFair___ joined 13:11 MikeFair_ left
FROGGS[mobile] the only thing would be that you have parrot specific tests 13:12
err, spectests
lizmat gotcha.. will see if I can do something to test the jvm case as well and adapt release guide accordingly 13:13
now off for some tivoli&
FROGGS[mobile] or does it include all and does the skipping and fudging? 13:14
moritz: can you tell?
dalek rl6-roast-data: b2e337f | coke++ | / (4 files):
today (automated commit)
 13:17
timotimo good, 0 failing tests 13:18
13:19 huf left
moritz FROGGS[mobile]: iirc it includes all, and simply uses t/harness to run the tests, which does the fudging and skipping 13:19
[Coke] woolfy++ 13:20
13:20 huf joined 13:21 benabik left
[Coke] S02-types/deprecations.rakudo.jvm aborted 27 test(s) 13:21
those are the only rakudo failures we're seeing in yesterday's run.
(the parrot stuff got cleaned up, moritz++)
dalek kudo/WINNER: 528c078 | (Timo Paulssen)++ | src/ (2 files):
fixes for wait $n and winner *
 13:22
moritz I cleaned up parrot stuff? I can't remember :-)
timotimo if anybody wants to play with the winner syntax and find the (probably) lots of cases where it doesn't work as planned yet ... ^ 13:24
[Coke] moritz: "don't run these jvm only tests on parrot" 13:25
13:25 berekuk joined
moritz ah 13:25
13:34 cooper left 13:35 rurban1 joined, berekuk left 13:36 cooper joined, cooper left, cooper joined 13:39 rurban1 left 13:40 xenoterracide left, colomon joined
moritz also has a clean p-spectest o/ 13:41
13:44 kaleem left
timotimo huh, the negated character range test on jvm still fails. why? 13:52
13:54 spider-mario joined
timotimo in the repl it works. weird. 13:55
13:56 stevan_ joined 14:00 berekuk joined
FROGGS[mobile] sideeffects 14:01
timotimo that would be pretty terrible
14:03 colomon left
FROGGS[mobile] dies it work in a -e too? 14:05
does* 14:06
timotimo yes, it does
FROGGS[mobile] hmmm 14:07
Well, reduce the test until it works :o)
(sorry for stating the obvious, but I am not at my pc) 14:08
14:08 kshannon left 14:09 kshannon joined
moritz timotimo: fwiw all the nqp qregex tests pass on nqp-j for me 14:14
14:15 berekuk left 14:17 nnunley joined
Woodi hallo today :) 14:27
found PMC like that: pastebin.com/Dk9JFwyx 14:28
14:28 berekuk joined 14:29 xinming left, kaare_ joined, benabik joined 14:31 xinming joined
timotimo moritz: that was a rakudo-j test from S05-mass/rx.t 14:32
moritz timotimo: oh 14:33
diakopter . 14:34
14:34 PacoAir joined, berekuk left 14:45 btyler joined 14:47 yves_ joined, bluescreen10 joined
dalek kudo/WINNER: ba8b9fb | (Timo Paulssen)++ | src/Perl6/Actions.nqp:
listify the expression for winner

this makes single-argument winner work.		 14:52
14:56 ajr joined, ajr is now known as Guest8202 14:57 szabgab joined, Guest8202 is now known as ajr_, berekuk joined
szabgab hi, I just compiled Rakudo * on OSX , perl6 -v works, but running panda I get: Could not find Shell::Command 14:57
timotimo hey szabor! 14:59
tadzik szabgab! \o/
szabgab hi both!
timotimo no, wait. gabor, right?
szabgab tadzik: you still owe me a picture of your T-shirt! 15:00
timotimo i'm too lazy to write hundreds of tests for WINNER now
szabgab timotimo: that's a better way to address me, but szabgab is also fine :)
tadzik szabgab: oh, I thought you want the t-shirt itself :) 15:01
I thought I've replied to your email, hrm
szabgab you promised that too, but I'd be ok witha picture of it :) 15:02
tadzik I was sure I asked you for a size you want :)
szabgab but mostly I just want panda to work for me :)
tadzik I should be able to help with both :) 15:03
szabgab let's start with panda
straight out of 2013.10 star it tells me Could not find Shell::Command 15:04
have I missed a step?
moritz szabgab: do you have the transcript of your 'make install' run?
szabgab not any more
but I can run it again 15:05
moritz :( sadpanda
that might help, yes
szabgab oh and hi moritz 15:06
paste?
moritz szabgab: use gist.github.com
and welcome back szabgab 15:07
szabgab I got a few errors
eg.
Can't stat modules/shell-command/lib: No such file or directory
that might be a clue :)
same for modules/file-find/lib:
and modules/perl6-psgi/lib: 15:08
but it is still runing
this is stderr: gist.github.com/szabgab/7583213 15:09
15:11 thou joined
szabgab moritz: and now stdout is also pasted 15:11
moritz $ tar tzf rakudo-star-2013.10.tar.gz |grep modules/|grep -i shell|wc -l 15:12
0
no wonder it can't stat it
so, apparently the tarball simply doesn't contain module :( 15:13
wtf
seems there are 4 module directories missing from the tar ball 15:14
file-find, perl6-File-Tools, perl6-psgi, shell-command 15:15
szabgab++
15:16 rindolf left
moritz oh, perl6-File-Tools doesn't belong in there, so effectively 3 modules missing 15:16
szabgab: so you probably need to bootstrap your own panda (git clone it, and then run perl6 bootstrap.pl), or wait for the next star release
sorry about that
szabgab ok, thanks, then I can stop banging my head agains this wall 15:17
15:18 berekuk left
moritz szabgab: I've opened github.com/rakudo/star/issues/34 for you 15:19
szabgab thanks
moritz huh, that's weird. When I recreate the release locally, the files are included 15:24
15:24 kaleem joined
szabgab so I tried to install panda from github 15:25
at the end of the bootstrp it told me to include /Users/gabor/work/rakudo-star-2013.10/install/lib/parrot/5.9.0/languages/perl6/site/bin in the path
but there is no such directory 15:26
actually panda was installed as /Users/gabor/work/rakudo-star-2013.10/install/lib/parrot/5.9.0/languages/perl6/site/panda
15:26 stevan__ joined, stevan__ left
szabgab err, not even that 15:26
szabgab checking it now
oh, I see it first told me build stage failed for File::Find: Failed building lib/File/Find.pm 15:28
moritz tadzik++ wanted to debug that further, I believe
tadzik I'd love to 15:29
15:29 stevan_ left
szabgab in any case, regarding those directories, it told me to include the .../bin directory, but instead of that I had an empty panda directory! at that level 15:29
but I believe there is a problem before that
tadzik yeah :(
15:30 jnap joined
szabgab this is a problem, because I wrote a blog post on how to get started with rakudo * but I cannot post it as my script is in perl 6 using HTML::Template and I cannot install it :( 15:30
15:31 kivutar joined
szabgab is there a templating system that comes with * ? 15:31
tadzik I think Template::Mojo does 15:32
szabgab No Pod found in ... Template/Mojo.pm
tadzik oh, shame on me
szabgab I remember,I started to write docs for that 15:33
tadzik oh, awesome!
szabgab that was the last time my Yak ran away :)
and I could not finish shaving it :(
15:33 berekuk joined
szabgab let me see if I can find the left over from that time 15:34
15:41 raiph joined 15:48 rurban1 joined 16:05 berekuk left
FROGGS[mobile] tadzik: I am going to add an S11 compliant installer (hook) to panda 16:08
if you know any current problems that I need to be aware of, please tell 16:09
the will coexist different installers btw 16:10
16:11 berekuk joined
FROGGS[mobile] there* 16:12
16:16 rurban1 left, rurban1 joined
Woodi hey, found nice PRAM/XMT advertisement :) www.umiacs.umd.edu/users/vishkin/XM...dJelly.pdf 16:17
moritz suggests relabeling the jars 16:18
Woodi umm... one by one or the other way ? ;) 16:25
16:26 szabgab left 16:29 berekuk left, [Sno] left 16:39 berekuk joined 16:43 berekuk left 16:45 zakharyas left, lizmat_ joined 16:48 lizmat left 16:57 ajr joined, ajr_ left, ajr is now known as Guest13120, Guest13120 is now known as ajr_ 17:03 Rotwang joined 17:10 pzu joined 17:12 pzu left 17:14 kaleem left
moritz by outsourcing. Don't care how *they* do it :-) 17:17
17:17 berekuk joined 17:21 berekuk left
timotimo No such method 'has_value' for invocant of type 'Promise' - huh 17:24
dalek kudo/WINNER: 39c85ba | (Timo Paulssen)++ | src/vm/jvm/core/asyncops.pm:
promise has "has_result", not "has_value".
 17:27
timotimo nobody has been testing my stuff for me! :(
TimToady there is no has_result in S17 17:29
timotimo ah, it's gone
it was still here locally
dalek kudo/WINNER: 3e53dd5 | (Timo Paulssen)++ | src/vm/jvm/core/asyncops.pm:
use Promise.Bool instead of has_result/value.
 17:30
timotimo force-pushed over it :P
my invoke_right isn't correct >_> 17:32
oh, in fact, it odesn't use invoke_right here.
17:34 logie left
dalek kudo-star-daily: 951eaac | coke++ | log/ (5 files):
today (automated commit)
 17:36
timotimo great! lots of deprecations gone! :) 17:40
17:41 FROGGS joined 17:42 [Sno] joined
timotimo huh, now the test hangs 17:43
it calls last, but the loop does not terminate 17:45
that's kind of not so good
i don't think i can fix that to make the last lexotically last out of the map :( 17:47
if it's dynamically inside the loop that's inside winner
can i put a CATCH block around the invocation that throws last/next/... further out? or something like that? 17:48
17:49 thou left, dakkar left 17:51 Entonian joined
TimToady someone needs to straighten out all that labeled loop stuff that P5ers take for granted 17:51
17:51 benabik left
timotimo yeah :( 17:51
[Coke] r: HELL: say "here"; goto HELL; 17:52
camelia rakudo-parrot 180fca: OUTPUT«===SORRY!=== Error while compiling /tmp/Skd9U8kTFq␤Undeclared name:␤ HELL used at line 1␤Undeclared routine:␤ goto used at line 1␤␤»
..rakudo-jvm 180fca: OUTPUT«===SORRY!=== Error while compiling /tmp/0jNk4JKQ6H␤Undeclared name:␤ HELL used at line 1␤Undeclared routine:␤ goto used at line 1␤␤»
TimToady alas, it's one of those not-so-glamorous jobs
np: HELL: say "here"; goto HELL; 17:53
hmm
n: HELL: say "here"; goto HELL;
[Coke] Might be nice if camelia standardized "while compiling /tmp/..." so the output would collapse.
camelia niecza v24-98-g473bd20: OUTPUT«(timeout)here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here␤here…»
Rotwang hi 17:54
17:54 ajr_ left
timotimo yes, it would 17:54
[Coke] (or, used the same temp file for both.)
TimToady pugs: HELL: say "here"; goto HELL;
camelia pugs: OUTPUT«here␤*** No such subroutine: "&goto"␤ at /tmp/_UrEDOy7H3 line 1, column 19-28␤»
TimToady n: OUTER: loop { INNER: loop { say "here"; last OUTER } } 17:55
camelia niecza v24-98-g473bd20: OUTPUT«here␤Unhandled exception: Nominal type check failed in binding '$x' in 'last'; got OUTER, needed Any␤ at /home/p6eval/niecza/lib/CORE.setting line 0 (last @ 1) ␤ at /tmp/Ex4gymBHW3 line 1 (ANON @ 4) ␤ at /tmp/Ex4gymBHW3 line 1 (ANON @ 5) ␤…»
TimToady wow
[Coke] wow good or wow bad
TimToady n: OUTER: loop { INNER: loop { say "here"; OUTER.last } }
camelia niecza v24-98-g473bd20: OUTPUT«here␤Unhandled exception: Unable to resolve method last in type OUTER␤ at /tmp/rHVmusgV75 line 1 (ANON @ 4) ␤ at /tmp/rHVmusgV75 line 1 (ANON @ 5) ␤ at /tmp/rHVmusgV75 line 1 (mainline @ 6) ␤ at /home/p6eval/niecza/lib/CORE.setting line 4…»
timotimo ~@(Publish.for(1..5).Channel) just hangs :\ 17:56
17:56 kivutar left
TimToady what does it mean to coerce a Supply to a Channel anyway? 17:58
there's something deeply suspicious about Supply.for anyway
FROGGS jnthn: is it possible (for you) to outline the implementation of goto and last LABEL and friends? 18:01
timotimo deeply suspicious? 18:03
FROGGS jnthn: so, in a month when S11 is done :o) 18:07
18:10 rurban1 left 18:11 Entonian left, rurban1 joined 18:15 fhelmberger_ joined 18:19 fhelmberger left, thou joined 18:20 fhelmberger_ left
TimToady it just seems to me that Supply.for is already just a channel in disguise 18:21
depending on whether the events get scheduled serially or in parallel
it's a supply that wakes up "when tapped", so it cannot easily have a second tap added, if the 1..5 are delivered instantaneously to the first tap 18:22
perigrin "Sir?" "One TAP only." 18:24
dalek ast: 3554f38 | (Timo Paulssen)++ | S17-concurrency/ (2 files):
skip hanging tests :(
TimToady if your 'more' tap is named 'a', does it call a(1),a(2),a(3),a(4),a(5) in parallel? or is there some implicit serialization intended but not specced? 18:25
if the latter, it really is just a channel
dalek kudo/WINNER: 2622c63 | (Timo Paulssen)++ | src/vm/jvm/core/ (3 files):
hotfix for winner swallowing last.
18:25 ajr_ joined
timotimo i have more confidence in this code now than i had earlier today, but i'm still not sure it's right enough to be released
lizmat_ is back but about to go out to see a wizard 18:26
18:26 lizmat_ is now known as lizmat
TimToady if the parallel call is intended, it seems a rather difficult way to implement a hyper or race 18:26
lizmat will start release process in about 4 hours
TimToady so I don't know what .for is intended to be used for 18:27
timotimo anybody want to write some tests for winner? i've gotta run again for a bit.
lizmat fwiw, it feels more like a demo of possibly functionality then actual useful functionality atm 18:28
I'll write tests for winner, but *after* the release
18:31 xinming left, xinming joined
lizmat off to see a wizard& 18:33
Rotwang can I somehow suppress deprecated warnings? 18:34
I know that the point of deprecated warnings is that they should be fixed, but I get these warnings from core perl6 library 18:35
18:37 SamuraiJack left
TimToady a CONTROL block is supposed to be able to intercept warnings, but I don't know how implemented that part is 18:37
there's a specced 'quietly' statement prefix that I'm pretty sure is unimplemented
I don't know if there's a way to suppress classes of warnings yet at the source 18:39
that would be problematic in any case if the warnings are coming from a different lexical scope, assuming such warning suppression is dynamic and not lexical 18:40
er, other way around
CONTROL is the intended way to modify warnings dynamically 18:41
18:42 spider-mario left
TimToady I guess the primary intent of Supply.for is to translate the blocking of a lazy list into a schedule of events, so 1..5 is a bad example, since it likely won't block 18:43
18:46 spider-mario joined, PZt joined
TimToady otoh, something like Supply.for(lines) can block while doing I/O, but if the lines are dispatched in parallel, I'm not sure the 'race' semantics are worth much, unless you're just counting occurrence of something and don't care about the result order 18:46
if that's the case, then Supply.for is mostly only good for lazy lists whose laziness is timed by some real events, and that doesn't do significant batching of those events while processing the lazy list 18:47
timotimo rotwang, the message is probably wrong and the warning comes somewhere from your code 18:48
Rotwang Method exists (from EnumMap) called at: 18:49
/usr/lib/parrot/5.9.0/languages/perl6/lib/NativeCall.pm6, line 81
Please use the :exists adverb instead.
TimToady and if it's intended as the way to implement or hyper semantics, then it doesn't solve the fundamental problem that the list is lazy, and yields values one at a time 18:50
hyper and race need to primitively be able to split lists in the middle, not head/tail
moritz Rotwang: that warning is fixed in the git repo of NativeCall
Rotwang cool, thanks
TimToady (this is Guy Steele's insight into what's wrong with most FP languages)
Oh, another problem with Supply.for, "If the iteration at some point produces an exception, then the C<catch> 18:53
callable will be invoked."
how is it supposed to catch an exception in a lazy list without a context in which to install a CATCH?
well, I guess it's thrown at the point the next value is pulled out of the list, is all, so nevermind 18:54
TimToady is obviously in one of those days where he thinks out loud too much... 18:55
19:02 nebuchadnezzar left, tgt left 19:03 nebuchadnezzar joined
Rotwang is there any way to pass unsigned char or whatever in NativeCall? 19:04
19:04 tgt joined 19:10 eternaleye left, kivutar joined
moritz star: use NativeCall; my uint8 $x 19:12
camelia ( no output )
moritz Rotwang: that's what I'D try
19:15 xinming left
Rotwang moritz: I've glanced through NativeCall.pm6 and found nothing similar, but it actually works 0__o, perl6 is a land of wonder for me 19:16
moritz p: my uint8 $x; say $x
camelia rakudo-parrot 180fca: OUTPUT«0␤» 19:17
moritz Rotwang: it's probably implemented and mapped at the NQP level
19:17 xinming joined
Rotwang moritz: ok thank you 19:18
timotimo if nobody objects, i'll merge WINNER into nom. 19:28
TimToady is probably going to rename fail/catch to quit/quit 19:33
to give more implication that the supply is terminating early
diakopter quitter
timotimo i had not implemented fail yet and catch is currently just a stub action that does nothing special 19:34
it'll probably result in strange errors if i don't put in an explicit NYI, though.
19:35 kivutar left
timotimo should i rename that catch to quit right now? 19:35
19:35 bluescreen10 left
dalek ecs: 194620a | larry++ | S17-concurrency.pod:
change fail/catch to quit/quit
 19:39
TimToady yes :)
timotimo :) 19:40
TimToady thanks for your patience, or in the absence of that, your persistence :)
TimToady promised not to change S17 again in the next five minutes
timotimo :)
TimToady *ses 19:41
moritz 4 and counting
TimToady well, anyone else might change it in that time :)
diakopter buhleted
dalek kudo/WINNER: c3c3e54 | (Timo Paulssen)++ | src/Perl6/ (2 files):
catch => "quit", NYI for quit blocks
 19:42
19:42 ssutch joined
nwc10 3 and conting 19:42
er, counting
19:42 spider-mario left
diakopter also, conting 19:43
19:43 beastd joined
TimToady note the fail method is now the quit method 19:44
timotimo i'm not responsible for that part :) 19:47
dalek ecs: 97fc551 | larry++ | S17-concurrency.pod:
add vim line
timotimo i only do winner.
19:48 spider-mario joined
TimToady good thing you don't implement the vim part either :) 19:48
nwc10 "this hour in Perl 6..."
timotimo perl6 plugins for vim? :)
moritz timotimo: re objections to merging, I have none except my general dislike of last-minute-before-release patching 19:50
timotimo lizmat asked me to develop this in not-a-branch from the start :P
i'll go ahead then. 19:51
dalek kudo/nom: b76c38c | (Timo Paulssen)++ | src/ (3 files):
implement winner syntax and WINNER sub.
rakudo/nom: 3129cb7 | (Timo Paulssen)++ | src/vm/jvm/core/ (2 files):
rakudo/nom: update usages of winner in the setting
19:51 dalek left, dmol1 joined
timotimo sorry, dalek :( 19:52
EXCESS FLOODINATE~
19:52 dmol left, dalek joined, ChanServ sets mode: +v dalek
timotimo github.com/perl6/nqp/pulse/monthly - oh, i did lots of stuff this month! :D 19:53
github.com/rakudo/rakudo/pull/226 - what about this? 19:59
moritz timotimo: +1 on the idea, -1 on merging until the spectest failures are fixed 20:01
timotimo sure. 20:02
20:03 crab2313 joined 20:06 telex left 20:08 telex joined 20:09 DarthGandalf left 20:15 DarthGandalf joined
Rotwang r: 'use NativeCall;my $file="./libfoo";sub bar is native($file) { * }' 20:16
camelia ( no output )
moritz that's just a string literal :-) 20:21
Rotwang: that code won't work; $file is initialized at run time, but the 'is native(...)' trait is evaluated at compile time
Rotwang that's good to know 20:22
moritz best to use constant libfile = ...; ... is native(libfile) { * }
Rotwang so constant works also without a sigil [; 20:24
20:25 kivutar joined 20:35 masak left, masak joined 20:47 FROGGS[mobile] left 20:55 REPLeffect left, REPLeffect joined 20:57 Mouq left 20:58 awwaiid left, awwaiid joined
[Coke] wishes he could get some time from work to hack on p6. 20:59
21:15 kaare_ left
timotimo okay so making before smarter for simple contents and removing scan before anchors... what else could i do... 21:16
measure constant things width and generate scans for literals could be doable...
removing redundant scans directly after the nfa ran seems to have been safe but not that helpful for performance... i should come up with sane benchmarks for that 21:17
21:23 spider-mario left
raiph timotimo: profiling? 21:24
timotimo: (in response to your "what else could i do" and because i don't recall why you aren't focusing on profiling) 21:30
21:33 ssutch left
timotimo profiling parrot is *hard*. 21:37
21:45 dmol1 left 21:47 PacoAir left 21:50 crab2313 left
[Coke] timotimo: parrot has a builtin --profile that generates kcachegrind output. 21:51
timotimo yeah, that's still *hard* :(
[Coke] which part? generating the output, getting it into kcachegrind, or understanding it once you're there? 21:52
I can help with the first 2.
timotimo understanding it
[Coke] what about profiling java 21:54
timotimo never tried that 21:55
[Coke] might be easier, esp. if you use visualvm
dwarring timotimo: good catch, will refactor 22:12
22:26 arcane joined
timotimo i caught what? 22:28
ah, that. 22:31
where am i supposed to get jvisualvm from? 22:32
22:32 Rotwang left
[Coke] it's part of the distro 22:33
(in the latest JDKs)
timotimo i don't have it :(
[Coke] do you have a javac ?
timotimo yes 22:34
i have openjdk7-jdk
[Coke] you need the oracle jdk to get jvisualvm
timotimo oh
is that Java SE Development Kit 7? 22:35
[Coke] very likely.
timotimo ah, "standard edition"
[Coke] (we're an oracle shop at dayjob. I frequently forget there are other javas, sorry. )
as opposed to EE.
timotimo ah
[Coke] (which includes tons of webappy stuff, and more)
timotimo how do i make sure my installed-to-home jdk will be used for everything? 22:36
[Coke] I think the biggest issues with jvisualvm is that our classes don't have friendly java names. I think you can probably work out a mapping from the mangled names to the p6 ones. (by invoking .WHAT.Str, mebbe). I haven't done that kind of modification, though. 22:37
make it first in your path?
timotimo OK sure.
[Coke] If you get stuck on anything, I'm happy to chip in. 22:38
22:38 xenoterracide joined
timotimo builds nqp and rakudo with oracle jdk 22:39
22:40 arcane left 22:42 ssutch joined
[Coke] *kermitarmflail* 22:46
->
timotimo huh?
it seems like die_s is taking most of the compile time in the setting, haha 22:47
that's just wat.
22:48 stevan_ joined 22:52 lawrmcfa joined 22:57 BenGoldberg joined
diakopter timotimo: wat. 22:58
23:01 dmol joined 23:02 lawrmcfa left
timotimo i better get some rest now. 23:02
23:04 colomon joined 23:05 xinming left 23:06 xinming joined 23:23 eternaleye joined 23:38 kivutar left 23:44 rurban1 left 23:48 hummeleB1 left

Powered by Cro and the Raku® Programming Language.

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

Thank you!