»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.
00:00 AW3i left
dalek d: e347792 | Mouq++ | STD.pm6:
Generalize quote escapes to include starter

  (Not actually significant here, but still is important since the other
impls base their grammars off of STD.pm6)		 00:08
TimToady someone needs to update the feature matrix for moar concurrency 00:09
dalek atures: 8fe71d2 | jnthn++ | features.json:
Updates for MoarVM concurrency progress.
 00:13
TimToady thanks 00:14
TimToady is busy (not) writing his talk :) 00:15
Mouq Weird… I got a segfault and now I can't get it again… 00:16
00:27 hoverboard left
Mouq .tell FROGGS github.com/rakudo-p5/v5/pull/1 00:29
yoleaux Mouq: I'll pass your message to FROGGS.
00:45 rurban joined
jnthn sleep & 00:48
Mouq 'night jnthn! 00:49
00:55 telex left, xenoterracide left, colomon joined, colomon left 00:56 rurban left, telex joined
Mouq r: say hash set <a b c> 00:57
camelia rakudo-jvm dfc7d7: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at gen/jvm/CORE.setting:9066␤ in sub hash at gen/jvm/CORE.setting:9391␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-moar dfc7d7: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at src/gen/m-CORE.setting:9123␤ in sub hash at src/gen/m-CORE.setting:9448␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-parrot dfc7d7: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at gen/parrot/CORE.setting:9085␤ in sub hash at gen/parrot/CORE.setting:9411␤ in block at /tmp/tmpfile:1␤␤»
Mouq r: say set(<a b c>).hash
camelia rakudo-jvm dfc7d7: OUTPUT«("b" => Bool::True, "c" => Bool::True, "a" => Bool::True).hash␤»
..rakudo-parrot dfc7d7, rakudo-moar dfc7d7: OUTPUT«("a" => Bool::True, "b" => Bool::True, "c" => Bool::True).hash␤»
Mouq r: say %(set <a b c>)
camelia rakudo-jvm dfc7d7: OUTPUT«("b" => Bool::True, "c" => Bool::True, "a" => Bool::True).hash␤»
..rakudo-parrot dfc7d7, rakudo-moar dfc7d7: OUTPUT«("a" => Bool::True, "b" => Bool::True, "c" => Bool::True).hash␤»
Mouq Bug? 00:58
00:58 hoverboard joined
BenGoldberg r: my $s = set <a b c>; say hash $s; 01:01
camelia rakudo-jvm dfc7d7: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at gen/jvm/CORE.setting:9066␤ in sub hash at gen/jvm/CORE.setting:9391␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-moar dfc7d7: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at src/gen/m-CORE.setting:9123␤ in sub hash at src/gen/m-CORE.setting:9448␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-parrot dfc7d7: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at gen/parrot/CORE.setting:9085␤ in sub hash at gen/parrot/CORE.setting:9411␤ in block at /tmp/tmpfile:1␤␤»
Mouq r: my %h = set <a b c> 01:02
camelia rakudo-parrot dfc7d7: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at gen/parrot/CORE.setting:9085␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-moar dfc7d7: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at src/gen/m-CORE.setting:9123␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-jvm dfc7d7: OUTPUT«Odd number of elements found where hash expected␤ in method STORE at gen/jvm/CORE.setting:9066␤ in block at /tmp/tmpfile:1␤␤»
01:03 xenoterracide joined 01:08 xenoterracide left 01:13 rurban1 left 01:15 colomon joined 01:18 kurahaupo left 01:26 rurban joined 01:28 wbill joined 01:32 wbill left 01:40 denisboyun left 01:44 araujo joined 01:50 dayangkun left
dalek ast: b1b60df | (Andrew Egeler)++ | S32-io/IO-Socket-INET. (2 files):
Add socket test to cover recent moarbug
 01:53
retupmoca jnthn: moar patch now tested! 01:54
01:58 colomon left 02:18 colomon joined 02:26 xragnar_ joined, xragnar is now known as Guest67462, Guest67462 left, xragnar_ is now known as xragnar
timotimo r: say 1 ~~ Nat 02:27
camelia rakudo-parrot dfc7d7, rakudo-jvm dfc7d7, rakudo-moar dfc7d7: OUTPUT«===SORRY!=== Error while compiling /tmp/tmpfile␤Undeclared name:␤ Nat used at line 1. Did you mean 'Rat'?␤␤» 02:28
02:32 hiatobr joined
cognominal geeksta.net/geeklog/exploring-expre...-messages/ 02:40
02:41 xenoterracide joined
retupmoca so I want to compile and distribute a .so file (that will be used via NativeCall) along with my perl 6 module 02:43
is there a standard way to do this?
Mouq retupmoca: ./bin ? 02:44
Maybe in ./lib
But I do know that Panda copies files in ./bin
retupmoca yeah, but my module needs to be able to call out to it with sub ... is native('mylib') 02:45
idk if that'll look in ./bin
Mouq :/
02:46 lustlife joined 02:49 klapperl_ joined 02:53 klapperl left
Mouq → bed 02:54
timotimo → timotimo 02:55
03:00 Mouq left
cognominal who said Perl is slow? Perl 6 is caught speeding : twitter.com/SciencePorn/status/449...48/photo/1 03:13
03:14 punter left 03:16 colomon left 03:42 raiph left 03:57 kurahaupo joined
dalek osystem: e6cd767 | (Andrew Egeler)++ | META.list:
Add Auth::PAM::Simple
 04:04
04:11 kurahaupo left 04:12 xenoterracide left 04:31 hoverboard left, BenGoldberg left 04:32 hoverboard joined 04:34 hoverboard left 04:35 araujo left 04:47 zby_home joined 05:07 zby_home left 05:11 REPLeffect left 05:21 darutoko joined 05:24 REPLeffect joined 05:48 kaare_ joined 06:07 SamuraiJack joined 06:42 FROGGS joined 06:44 thou left 07:03 ggoebel11118 left, ggoebel11118 joined
FROGGS thanks Mouq! 07:19
yoleaux 00:29Z <Mouq> FROGGS: github.com/rakudo-p5/v5/pull/1
FROGGS jnthn: when we autogenerate a proto, then the traits like hidden_from_backtrace are left out... is that intentional? 07:31
07:52 sfsdfd joined 07:53 sfsdfd left 07:58 kaare_ left 08:02 Rounin joined
FROGGS m: say lines 08:06
camelia rakudo-moar dfc7d7: OUTPUT«(timeout)»
FROGGS :o(
08:06 woolfy joined 08:16 denisboyun joined 08:19 woolfy left, SamuraiJack left, IllvilJa left 08:22 ssutch joined, ssutch_ left
Rounin Hm... I was worried about stack space for recursive functions yesterday, since Rakudo apparently runs Perl on the JVM 08:27
But now I've had a program do nothing but recurse for 2.5 minutes, and it's still running, so that seems OK
I guess it does optimize tail calls
It leaks memory though 8| Perhaps it's just my system being very generous with memory 08:28
FROGGS yeah, I think we had another case where something leaked using perl6-j 08:29
but I cannot remember what is was
Rounin Ah well... 7 minutes of recursing at 100% CPU, and it's only using 5.9% of 8GB... Should be OK I suppose :P 08:32
FROGGS not too bad, aye :o) 08:33
08:40 thou joined 08:45 thou left 08:59 Rotwang joined 09:02 JimmyZ left 09:11 xfix left
dalek p: 83f5f9d | (Tobias Leich)++ | tools/build/MOAR_REVISION:
bump moar rev to get public real_data helper
 09:12
kudo/nom: 79eb79e | (Tobias Leich)++ | / (2 files):
make Rakudo_List a structure of a P6opaque's body

This body can be replaced when it needs to grow, e.g. when we mix a role with an attribute into a list.		 09:14
09:17 xfix joined
FROGGS ummm, there seems to be a lot of build system stuff todo for r-j* and r-m* 09:21
09:37 colomon joined 09:52 kaare_ joined
Rounin Agh.. I cut my finger making salad... Now I have only 9 fingers to try Perl 6 with! 09:53
FROGGS >.< 09:54
09:54 kurahaupo joined
FROGGS then perhaps use the texas style operators like >> or » 09:54
(easier to type)
Rounin :DD Thanks for the suggestion! 09:55
FROGGS :o)
09:55 dmol joined
Rounin I'll use only half of the keyboard ... my uiop = 9 - 8 09:55
Ach, $ is on the lefthand side DX
jnthn FROGGS: (*) Are you looking at the tristar branch, or master? 09:56
FROGGS master
damn
:P
jnthn :P
OK, there's a bit less work to do in Tristar ;)
FROGGS thanks, that saves me some work :o)
10:02 JimmyZ joined 10:05 iolympian1 joined
FROGGS jnthn: does it make sense to build for 2014.03? or should be build for HEAD? 10:05
10:06 rindolf joined
jnthn FROGGS: Should be aiming at released versions for * really 10:08
FROGGS k
I am building 2014.03 atm, and will have a look at the module tests 10:09
jnthn OK.
I guess it's possible we'll need point releases...
FROGGS lib/Digest/MD5.pm===SORRY!===
Error while compiling op p6store (source text: "$md5_hex = Q:PIR {\n .local pmc f, g, str\n str = find_lex '$str'\n ..."): To compile on the MoarVM backend, QAST::VM must have an alternative 'moar' or 'moarop'
jnthn Yeah, that's exactly what it says... 10:10
FROGGS precomp of lib/Perl6/TypeGraph/Viz.pm seems to hang
yeah, I am just pasting this so I know what modules needs attention, not that something is wrong on our side :o) 10:11
hmmm, and make modules-test seems to hang/do nothing 10:13
10:13 kaare_ left
jnthn Hmm 10:13
Not sure if the modules-test stuff got updated for Moar or not. 10:14
for multi-backend that is
10:24 AW3i joined 10:37 kivutar left
dalek ar/tristar: 7ef34d1 | (Tobias Leich)++ | / (2 files):
support modules-test for all backends
 10:45
10:50 colomon left 11:03 kurahaupo left 11:10 Alula joined 11:12 spider-mario joined 11:23 iolympian1 left
FROGGS failing modules in modules-test-m: gist.github.com/FROGGS/0ea5537eb675588baaa2 11:32
I'll build not for -j and later for -p
jnthn FROGGS++ 11:34
Hmm, that Pod::To::HTML thing was fixed post-release...
I *think*
FROGGS I think so too 11:36
perhaps we can checkout an older version of Pod::To::HTML 11:37
jnthn Well, or try it against HEAD...
Non-zero wait status: 139
139?
FROGGS hmmm, that rings a bell... 11:38
11:38 Rotwang left 11:39 LLamaRider joined 11:46 LLamalessRider joined 11:51 kivutar joined 11:55 Rotwang joined
dalek ar/tristar: a5ffea1 | (Tobias Leich)++ | tools/build/Makefile.in:
fix modules testing by using correct -j/-p runner
 12:11
12:28 LLamaRider left 12:29 LLamalessRider left
FROGGS star-test-modules build and test errors: gist.github.com/FROGGS/0ea5537eb675588baaa2 12:40
jnthn FROGGS++
looking
Damn, that LWP one is a release blocker... 12:45
And r-j is in way, way, worse shape than r-m o.O
FROGGS well, the weekend has just begun :o) 12:46
jnthn: so, r-j* has priority? 12:47
jnthn Well, fxing the r-p regression for LWP::Simple probably has to come first... 12:49
FROGGS k
dalek ast: 1589ef2 | moritz++ | S06-macros/unquoting.t:
fudge moar macro test (RTed)
moritz fwiw t/spec/S17-concurrency/channel.t aborts after test 12 here, with "Aborted (core dumped)" 12:50
FROGGS moritz: on my box too I think
jnthn Hm...how many tests are there? 12:51
moritz (though I can't find the core dump)
FROGGS at least, one of the S17 tests files failed on my box... I don not have the output anymore
the lwp-simple precomp problem seems to happen when we load MIME::Base64 12:55
looks like the fix is already in... checking 12:56
ahh, I am using MIME::Base64 HEAD where I probably should not 12:58
moritz jnthn: fwiw S17-concurrency/channel.t hangs here after test 12 on the JVM 13:00
and that seems to be at the end of the test file 13:01
(and it doesn't consume CPU while hanging) 13:02
FROGGS m: EVAL "pir::load_bytecode__vs('MIME/Base64.pbc');"
camelia rakudo-moar 79eb79: OUTPUT«===SORRY!===␤Error while compiling op call: To compile on the MoarVM backend, QAST::VM must have an alternative 'moar' or 'moarop'␤»
FROGGS m: EVAL "pir::load_bytecode__vs('MIME/Base64.pbc');"; say "alive"
camelia rakudo-moar 79eb79: OUTPUT«===SORRY!===␤Error while compiling op call: To compile on the MoarVM backend, QAST::VM must have an alternative 'moar' or 'moarop'␤»
FROGGS m: EVAL 'pir::load_bytecode__vs("MIME/Base64.pbc");'; say "alive"
camelia rakudo-moar 79eb79: OUTPUT«===SORRY!===␤Error while compiling op call: To compile on the MoarVM backend, QAST::VM must have an alternative 'moar' or 'moarop'␤»
FROGGS m: try EVAL 'pir::load_bytecode__vs("MIME/Base64.pbc");'; say "alive"
camelia rakudo-moar 79eb79: OUTPUT«alive␤»
FROGGS m: try EVAL "pir::load_bytecode__vs('MIME/Base64.pbc');"; say "alive" 13:03
camelia rakudo-moar 79eb79: OUTPUT«alive␤»
FROGGS m: if 0 { EVAL "pir::load_bytecode__vs('MIME/Base64.pbc');" }
camelia ( no output )
FROGGS I thought this will still put the call into the ast 13:04
dalek ast: 46d6d67 | moritz++ | S29-conversions/ord_and_chr.t:
(re)fudge ord_and_char.t for rakudo
 13:05
FROGGS okay, the pure perl MIME::Base64 works 13:08
moritz also LWP::Simple with pure perl MIME::Base64? 13:15
FROGGS at least I can compile it
moritz m: 1 = 2 13:18
camelia rakudo-moar 79eb79: OUTPUT«No such method 'STORE' for invocant of type 'Int'␤ in block at /tmp/Bx4qdWfHqC:1␤␤»
moritz r: 1 = 2
camelia rakudo-jvm 79eb79: OUTPUT«(timeout)»
..rakudo-moar 79eb79: OUTPUT«No such method 'STORE' for invocant of type 'Int'␤ in block at /tmp/tmpfile:1␤␤»
..rakudo-parrot 79eb79: OUTPUT«Cannot modify an immutable value␤ in block at /tmp/tmpfile:1␤␤»
dalek ast: 068ed16 | moritz++ | S32-exceptions/misc.t:
moar-fudge S32-exceptions/misc.t
 13:22
13:25 logie joined 13:31 logie left 13:32 kurahaupo joined 13:33 logie joined 13:34 kurahaupo_mobile joined 13:37 kurahaupo left 13:40 ssutch left, ssutch joined 13:42 ssutch_ joined, ssutch left 13:44 kurahaupo_mobile left 13:45 kurahaupo joined
retupmoca FROGGS: FWIW, MIME::Base64 head works fine on moarvm 13:52
LWP::Simple will compile and run tests (it fails a couple, but they run and a bunch pass)
FROGGS retupmoca: I am using MIME::Base64 HEAD and MoarVM/nqp/rakudo 2014.03 13:53
and right now I try to make it work on parrot 6.1.0
I am going to try MIME::Base64+rakudo@parrot HEAD now 13:56
14:04 raiph joined
raiph jnthn has updated the compiler/features matrix data; what updates the web page? 14:05
jnthn raiph: As far as I know, there's some con job set up to do it automatically 14:09
*cron
14:13 xenoterracide joined
moritz 12,42 * * * * sh ~/update-features.sh 14:16
dalek atures: 6d4d025 | moritz++ | README:
update REAMDE

  * mention cron job
  * mention web page		 14:21
FROGGS retupmoca: LWP::Simple on rakudo@parrot HEAD does not work 14:26
bbiab
14:28 logie left 14:34 dmol left 14:36 dmol joined 14:37 FROGGS[mobile] joined, raiph left
kurahaupo it's run twice an hour, not once an hour. 14:39
retupmoca FROGGS: Does the precompile fail or do the tests fail? 14:40
kurahaupo assuming we're talking about moritz'z crontab entry
[Coke] rakudo-moar build failed on my older mac.
moritz kurahaupo: "at leat" 14:41
[Coke] gist.github.com/coke/9855725
moritz *least
[Coke]: looks like you either need to clean out some installed files, or rebuild nqp and parrot after rebuilding moar 14:42
commit eb2ba0792b166ebd2febcb2d99e2bb047972ca56 14:43
Author: Jimmy Zhuo [email@hidden.address]
Date: 2014-03-29 20:33:19 +0800
s/MVMP6opaque_real_data/MVM_p6opaque_real_data/g
sounds like it might cause that very problem
[Coke] That's highly plausible. checking. 14:44
moritz afk
FROGGS[mobile] retupmoca: precomp of LWP::Simple fails 14:45
retupmoca FROGGS[mobile]: I'll take a look once I get a rakudo built on all three backends 14:46
FROGGS[mobile] for parrot the problem is the eval of the load_bytecode call 14:47
strip the eval and it works for parrot
retupmoca *sigh*
FROGGS[mobile] indeed
retupmoca the eval was the only way I could get it to work on moar
the easiest way to fix this is to just strip out the PIR bytecode and make this a pure-perl only module 14:49
FROGGS[mobile] cant we require the ::PIR xor ::Perl and export the subs at runtime?
jnthn Could also do it in a BEGIN block...
retupmoca that's what I was doing before, but IIRC the require doesn't precompile
timotimo this is one of the current weak spots in rakudo (or maybe even the spec), IMO
jnthn require is runtime, but you can do it at BEGIN time.
retupmoca yes, I was doing a require at BEGIN time 14:50
but it never worked when precompiled
timotimo for example, how do i best handle checking if the user has libunibilum installed and if not fall back to curses and still get the benefits of precompilation?
FROGGS[mobile] m: require Test <&is> 14:56
camelia ( no output )
FROGGS[mobile] I would do that 14:57
retupmoca This is what it was when the require was in there: github.com/perl6/Perl6-MIME-Base64...Base64.pm6 14:58
that worked for parrot and jvm if it wasn't precompiled, 14:59
but it worked for 0 backends when precompiled
(it would give an error 'Missing or wrong version of dependency ...PIR.pm6' or something for parrot 15:00
15:01 hoverboard joined, raiph joined
FROGGS[mobile] yeah, I had that too when I played around with BEGIN+require 15:01
[Coke] nope. removed install, realcleaned, reconfigured, failed. 15:02
I ran with --gen-nqp=master --gen-moar=master --backends=moar 15:03
retupmoca does anyone think removing ::PIR completely for now is a bad idea? 15:04
FROGGS[mobile] retupmoca: in the previous code you used the symbols like M::B::PIR, but I propose to import subs, and with the pasted snippet these subs are known at compile time
JimmyZ COke, what's the error? 15:05
retupmoca FROGGS[mobile]: so a runtime require instead of BEGIN time?
FROGGS[mobile] yes
retupmoca ok, I'll try that
FROGGS[mobile] if that works out we can try to put it in a begin block 15:06
[Coke]: moar master cannot be used atm for rakudo 15:07
jnthn JimmyZ: Did you update Rakudo's usage of the function you renamed? 15:08
FROGGS[mobile] it needs to be aligned to the changes in moar, like the renaming of the C function moritz++ mentioned
JimmyZ I didn't... 15:09
Doesn't it need bump version? 15:10
jnthn JimmyZ: Yes, will have to bump vesions along with the change. 15:11
[Coke] FROGGS[mobile]: danke. 15:13
FROGGS[mobile] problem is just that many folks (me included) expect that HEAD/HEAD/HEAD works
15:16 dmol left, dmol joined
[Coke] I was asking for trouble and I got it. no worries. 15:17
raiph moritz++ # cron info/doc/fix 15:26
15:27 [particle] joined
retupmoca does star use panda to install it's modules? or does it have it's own precompilation system? 15:31
jnthn It's own thing 15:35
retupmoca will it use a custom Build.pm like panda will?
jnthn: or could it be taught to? 15:36
because using require for this will still cause problems with panda and star trying to compile the PIR.pm6 15:37
but with a Build.pm I can tell panda to never even look at anything parrot-specific
but that requires star to use Build.pm as well 15:38
jnthn retupmoca: In the long run, I think Star should use Panda, but that's a bigger task...
FROGGS[mobile] we could special case it like we did for perl6-debug 15:41
jnthn True
FROGGS[mobile] but that should not be the first choice
jnthn Though the easy solution is just to inline the Q:PIR bit in a heredoc
And check if we're on Parrot, and if so eval that instead. 15:42
Caching it.
Instead of trying to be clever with modules. :)
15:42 zakharyas joined
retupmoca jnthn: I was thinking something like this with Build.pm: github.com/perl6/Perl6-MIME-Base64...t/Build.pm 15:46
but that won't work if star won't use it
jnthn: MIME::Base64 master is doing an EVAL of the PIR code, but FROGGS was saying the EVAL wasn't working 15:55
FROGGS[mobile] exactly 15:57
the eval of the first pir:: explodes 15:58
jnthn Explodes how? 15:59
16:01 guru joined, guru is now known as Guest51982, Guest51982 is now known as ajr_
retupmoca ok, the buildpm-test branch works for panda on parrot - LWP::Simple compiles and runs tests 16:03
16:11 arnsholtAtHome joined
arnsholtAtHome o/ 16:11
jnthn
.oO( A massively distributed search system for arnsholts... )
 16:14
16:14 zakharyas left
arnsholtAtHome Yeah 16:17
Other arnsholt is at work, and forgot to put it in a screen session 16:18
16:18 zakharyas joined 16:20 dwarring left
retupmoca jnthn, FROGGS[mobile]: github.com/retupmoca/star/compare/...r?expand=1 + MIME::Base64 branch buildpm-test should work for all backends 16:20
jnthn retupmoca++ 16:21
retupmoca If that's what we want, I will PR star and merge buildpm-test to master
star won't use M::B::PIR until it's taught Build.pm
it'll just use the pure-perl, but it'll work 16:22
panda (and star once it uses panda / Build.pm) will use ::PIR on parrot, and everything else will use ::Perl 16:23
jnthn Works for me. 16:24
retupmoca MIME::Base64 branch merged; PR: github.com/rakudo/star/pull/37 16:26
(someone should probably double-check all this to make sure I'm not making things worse :P)
16:28 hoverboard left
jnthn If FROGGS don't beat me to it, will have a look after doing walk/shopping. 16:31
arnsholtAtHome jnthn: Not as many problems with nqp/jastcompiler on my laptop 16:33
gist.github.com/arnsholt/9857621 has the spectest summary of j-spectest on my laptop
nqp/jastcompiler with latest nqp/master merged in and rakudo/nom at HEAD 16:34
16:35 raiph left
arnsholtAtHome Rebuilding nqp/master to compare with now 16:35
16:36 denisboyun left, denisboyun joined 16:41 SamuraiJack joined
FROGGS[mobile] jnthn: it exploded with "No STable at index 1" 16:45
I am at home in an hour, and will then continue testing (including the PR) after dinner 16:46
16:50 pecastro_ joined, pecastro left
jnthn arnsholtAtHome: If isDEPRECATED and error-reporting are happy in nqp/master, most probably the issue is something around line number annotations. 17:05
arnsholtAtHome Good to know! 17:10
(And it is happy on master) 17:11
On S32-io now, and the only one that still fails is the mixhash one in S02 17:14
So there's still some breakage =)
S32-io/other.t is REALLY slow though 17:15
17:15 hummeleB1 left 17:16 kaare_ joined 17:18 xinming_ left
jnthn Yes, that one is...on any backend. And JVM especially. 17:18
17:19 xinming_ joined
arnsholtAtHome Why is that, OOC? 17:19
jnthn 'cus it runs Rakudo 60 times, I think loking for some decoding boundary error that we once had 17:23
arnsholtAtHome Ow, ow, ow. Yeah, that'll take a while
Ye olde Parrot bug where Unicode chars spanning a read block boundary borked, possibly? 17:24
Gist updated with master spectest run, BTW
Looks like we have some passing TODOs, but probably something broken with line number annotations 17:25
jnthn arnsholtAtHome: Yeah, that one 17:26
arnsholtAtHome The broken callframe.t test is checking line numbers as well, so you're probably right
jnthn ah, yeah
it fits
arnsholtAtHome isDEPRECATED.t fails with "Cannot look up attributes in a type object" on jastcompiler, not sure what's up with that, but I'll look into it 17:27
Tuesday, most likely 17:28
Oh, derp 17:29
Just looked at the code that handles JAST::Annotation 17:30
It gets the line number from the annotation, and proceeds to do... absolutely nothing with it
*facepalm*
Got to go any moment now, but hopeful fixing that'll get me somewhere =) 17:34
jnthn :)
arnsholtAtHome++ 17:35
arnsholtAtHome And this is why we have a massive testsuite
jnthn aye :) 17:36
17:39 ssutch joined, ssutch_ left 17:40 Psyche^ joined 17:41 ssutch left, ssutch joined 17:44 Psyche^_ left 17:45 MrMeek left, MrMeek joined
arnsholtAtHome isDEPRECATED passes now at least 17:48
arnsholtAtHome is hopeful
arnsholtAtHome & 17:53
timotimo \o/ 17:54
FROGGS[mobile] arnsholt++ 17:56
17:57 denisboyun left 17:58 denisboyun joined 18:01 araujo joined 18:09 arnsholtAtHome left 18:11 arnsholtAtHome joined
jnthn Doing a patch locally that might help the "Can't deserialize_stub a NativeCall object." issue in the *-j 18:14
FROGGS jnthn++ 18:38
dalek ar/tristar: 90eaf28 | (Andrew Egeler)++ | tools/build/module-install.pl:
Require file extension to be at end of filename
 18:40
ar/tristar: f2236b4 | (Tobias Leich)++ | tools/build/module-install.pl:
Merge pull request #37 from retupmoca/tristar

Require file extension to be at end of filename
moritz retupmoca: I've given you a commit bit for rakudo/star 18:43
retupmoca: welcome to the fun!
I mean, welcome even more to even more
fun
FROGGS jnthn: that patch will land in NativeCall itself? 18:48
18:49 Adriaaan joined 18:55 lizmat joined
jnthn FROGGS: No, NQP... 18:55
FROGGS jnthn: hmmm... so we do a 2013.03.1 at some point? 18:56
jnthn Yeah 18:57
FROGGS k
I am rebuilding star for all backends now, and will update my problems gist
moritz if that turns out to be not the only issue, we might want to wait with tristar for 2014.04 18:58
and maybe do an old-fashioned r-p star release for March
FROGGS moritz: sure, but let's try to fix as much as we can :o)
moritz FROGGS: aye 18:59
jnthn moritz: Well, my idea for it was to make sure r-p is in good shape, and if r-j/r-m aren't perfect we just say they're "experimental" or something.
FROGGS however we do it: we should write in underlined and bold that this is the first jvm*/moar* 19:00
jnthn Right.
FROGGS even when all passes I mean 19:01
because the tests usually do not mean much :o)
jnthn ugh
FROGGS hmm?
jnthn I fixed that one, but seems there's still an issue
FROGGS :/
jnthn It looks really familiar 19:02
Oh...
It's the thing I fixed up in r-m for NativeCall and pre-comp
19:03 zakharyas left
FROGGS every "# XXX ..." comes back to you at some point :P 19:03
dalek p: 532788a | jnthn++ | src/vm/jvm/runtime/org/perl6/nqp/sixmodel/reprs/NativeCall.java:
Fix a serialization issue with NativeCall on JVM.
 19:04
19:04 zakharyas joined 19:07 darutoko left
simula67 .tell moritz What do you think ? github.com/perl6/roast/pull/37 19:14
yoleaux simula67: I'll pass your message to moritz.
moritz simula67: I'm no expert when it comes to socket tests 19:15
yoleaux 19:14Z <simula67> moritz: What do you think ? github.com/perl6/roast/pull/37
moritz simula67: does it work?
simula67 yes. it is slightly better. I mean the test does not fail that often :) 19:16
moritz simula67: then go ahead an merge it, please :-)
jnthn moritz, simula67: Seems reasonable to me. 19:17
moritz simula67: I've given you a commit bit
19:18 kaare_ left
dalek ast: a55d4b2 | (Joji Antony)++ | S32-io/IO-Socket-INET.pl:
Use the server ready flag file for discard test also
 19:18
ast: 5bf4b14 | (Joji Antony)++ | S32-io/IO-Socket-INET.pl:
Merge pull request #37 from simula67/master

Use the server ready flag file for discard test also
simula67 thanks 19:19
moritz simula67: thank you for your patch! 19:20
dalek ast: b56f999 | (Elizabeth Mattijsen)++ | S17-concurrency/thread.t:
Differentiate reasons for skipping tests
 19:27
19:35 woolfy joined 19:36 SamuraiJack left, colomon joined 19:41 dwarring joined 19:45 Adriaaan left, colomon left 19:59 zakharyas left 20:05 xenoterracide left
FROGGS updated: gist.github.com/FROGGS/0ea5537eb675588baaa2 20:06
jnthn Working on those NullPointerException things, or what I think is the cause of them. But having a hard time of it. 20:10
oh... 20:12
I mighta got it.
yeah, gets further. Just another piece of NYI serialization.
[Coke] bugadmins: there is a switch on rts to indicate the backend, parrot/moarvm/jvm 20:13
moritz where is it? 20:16
ah, "Basics" 20:17
FROGGS in the general tab, on the lower right I thinl
think*
basics, yeah
moritz [Coke]: gaah, and it lets me select only one value :(
[Coke] moritz: I'll open a ticket with the bugadminadmins 20:18
moritz RT #121530 is a perfect example for needing two bacakend values
synopsebot Link: rt.perl.org/rt3//Public/Bug/Displa...?id=121530
[Coke] moritz++ # sent 20:19
dalek kudo-star-daily: d51c043 | coke++ | log/ (5 files):
today (automated commit)
 20:21
20:26 LLamaRider joined 20:27 LLamalessRider joined
FROGGS r: sub infix:«<<<»(\x, \n) { } 20:32
camelia ( no output )
FROGGS r: sub infix:<⊞>(\x, \y) { (x + y) % 2**32 }
[Coke] n: say 'abc'.split('')[0];
camelia ( no output )
niecza v24-109-g48a8de3: OUTPUT«␤»
lue FROGGS: what hash are you working on? :) 20:34
FROGGS lue: I try to make a pure perl Digest::MD5 20:35
I took the example from rosettacode, but the two snippets I pasted segfault when being precompiled
lue FROGGS: first thing I think of as a possible cause is the \x business. 20:36
(instead of $x or whatever)
[Coke] anyone else get a write error building rm all the time?
FROGGS no, it is either the unicode thingy, or a custom infix itself
[Coke] I make, it fails, I make again, we're all good. 20:37
20:37 telex left
jnthn [Coke]: Never seen that, though I@m typically building on Windows. 20:37
FROGGS [Coke]: I get that sometimes, yes
lue It can't be unicode if <<< fails, can it?
FROGGS lue: true
lue (unless the fault would lie with «»)
20:39 telex joined
lue FROGGS: ooc, does defining an existing operator work (e.g. overloading infix:<+>) ? Would defining an infix:<@> work (not already existing, not unicode, no angles to /just maybe/ screw things up)? 20:39
FROGGS hold on 20:40
Rounin Hm... One could implement APL with that functionality :P
FROGGS lue++
lue whadidido?
FROGGS <+> works, <@> does not, <ll> does not
lue so when you precompile, you enter C++ land? Yeesh. 20:41
FROGGS I do?
lue FROGGS: let me rephrase: "you follow the C++ school of operator definitions"
vendethiel (infix:<@> won't work ?)
lue: is there an operator"" in perl6? 20:42
FROGGS vendethiel: it would if there was not that bug I am seeing right now
dalek kudo/nom: e270a77 | jnthn++ | src/ (4 files):
Fix role outers bug on JVM.

Was fixed on MoarVM already; this brings the same fix to JVM. Gets NativeCall pre-comp closer to working.
lue vendethiel: not unless you define one :)
jnthn FROGGS: got a bt for the segv?
vendethiel lue: I mean, can you define your own "string parsers" like qq ? I think I've read about that in the specs
jnthn e270a77 was tricky... 20:43
vendethiel kinda like scala
FROGGS jnthn: not yet...
jnthn FROGGS: k
lue FROGGS: what I meant is that precompilation kicks into a state like C++, where you can't define your own operators. That's all :)
jnthn FROGGS: Above should help the various nullpointerexception failures for r-j
lue vendethiel: you can define your own string quoters, though that requires macros (because it's a change to how Perl 6 is parsed.)
vendethiel noted 20:44
FROGGS Error while reading from file: Malformed UTF-8
at gen/moar/stage2/NQPHLL.nqp:1279 (/home/froggs/dev/star/work-m/rakudo-star-2014.03/install/languages/nqp/lib/NQPHLL.moarvm::65)
jnthn: ^^
lue vendethiel: perlcabal.org/syn/S02.html#Adverbs_on_quotes
FROGGS this is for sub infix:<@>(\x, \n) { (x +< n) % 2**32 +| (x +> (32-n)) } in a class
jnthn: btw, I have to build star against HEAD, I will do that in a bit 20:45
jnthn FROGGS: Uh, how'd you get that one?
FROGGS: Whenever I've seen that it was correct and I was feeding it latin-1... 20:46
FROGGS jnthn: I'll give you a snippet in a sec
20:46 Sqirrel joined
FROGGS perl6 --target=mbc --output=foo.moarvm -e "sub infix:<@>(\x, \n) { (x +< n) % 2**32 +| (x +> (32-n)) }" 20:47
perl6 -I. -Mfoo -e1
jnthn: that is all ^ 20:48
jnthn: ohh, all is fine on perl6-m HEAD :o) 20:49
dwarring moritz: ok to revert moarvm commit c5f3b5378, "check for invalid code points (jvm port)"?
surrogate range check is also questionable: ord >= 0xd800 && ord <= 0xdfff 20:50
I'll look at nqp Ops.java afterwards 20:51
lue dwarring: actually, I don't think surrogates are valid characters on their own. 20:52
20:52 iolympian joined 20:53 treehug88 joined
jnthn FROGGS: So...nothing for me to look at? :) 20:53
FROGGS jnthn: looks like :o)
I will now test against HEAD
20:54 ajr_ left 20:56 iolympian left
dwarring lue: alright I'll keep the surrogate check in 20:57
20:57 Adriaaan joined 20:59 kurahaupo left
lue yeah, AFAICT unpaired surrogate values (they aren't even characters) are errors. 20:59
lue afk
21:04 iolympian1 joined, iolympian1 left
timotimo "closer to working" :| 21:12
moritz dwarring: I've already revereted it 21:15
21:24 hugme joined, ChanServ sets mode: +v hugme, masak joined, iolympian joined, iolympian left, masak is now known as Guest95000 21:25 cosimo joined 21:31 Adriaaan left 21:33 SHODAN left 21:34 SHODAN joined 21:35 telex left
jnthn FROGGS: I just got a pre-compiled module doing a native call to work out on the JVM. 21:36
FROGGS cool!
21:36 telex joined
FROGGS r-j* installs right now, so I an post test results in a few minutes 21:38
jnthn OK...
It'll go better with this patch, though...
timotimo \o/ 21:39
21:39 ssutch left 21:40 ssutch joined
jnthn Just making sure I didn't but the Rakudo build 21:40
bust
FROGGS it is easy for me to pull and rebuild now
dalek p: dba4289 | jnthn++ | src/vm/jvm/runtime/org/perl6/nqp/sixmodel/reprs/P6Opaque (2 files):
Make sure inlined obj REPRs survive serialization.
 21:43
21:45 telex left
dalek ast: 80e2368 | (Elizabeth Mattijsen)++ | S17-concurrency/supply.t:
Add "todo" for merging taps not working
 21:46
ast: 3fa0240 | (Elizabeth Mattijsen)++ | S17-concurrency/promise.t:
Add "skip" for sleepsort not working
ast: 0a763be | (Elizabeth Mattijsen)++ | S17-concurrency/ (3 files):
Add skips for :in/:at/:every not yet implemented
kudo/nom: fea8aa8 | (Elizabeth Mattijsen)++ | src/core/Supply.pm:
Fix Supply.list and make code more readable
 21:47
kudo/nom: e501475 | (Elizabeth Mattijsen)++ | t/spectest.data:
Make sure all S17 tests are run on moar
21:47 telex joined
lizmat that should add quite a few passing tests for moar :-) 21:47
FROGGS lizmat++ 21:48
jnthn lizmat: ooh, nice :) 21:49
lizmat two things that struck me: sleep sort test not working and merging taps
will look at them in more detail now
jnthn lizmat: All the timer stuff is missing still on Moar 21:50
lizmat but the sleep sort works with "sleep"
jnthn Oh :)
lizmat is that also time dependent?
jnthn OK, then it could be something else :)
lizmat *timer
jnthn No
lizmat: In what way did it not work, ooc? 21:51
21:51 Adriaaan joined
lizmat checks again 21:53
Unhandled exception in code scheduled on thread 4493213696 21:54
Cannot find method 'Mu'
in block at src/gen/m-CORE.setting:17660
in block at src/gen/m-CORE.setting:17657
timotimo huh, yikes
lizmat seems like there is nothing in the queue ? 21:55
aka, $task := Mu
is "Thread.start(:app_lifetime" NYI ? 21:56
jnthn Queue *should* block... 21:57
:app_lifetime is passed down to Moar and stashed, but I don't ink it acts upon it. 21:58
*think
lizmat anyway, I thought I'd fix some low hanging fruit in the concurrency stuff before working on my slides again 22:01
22:03 lustlife left
jnthn :) 22:03
22:07 denisboyun left 22:08 denis_boyun joined 22:10 Rotwang left
FROGGS lib/Perl6/TypeGraph.pm 22:12
===SORRY!===
p6captureouters target must be a CodeRef
22:13 PerlJam left, PerlJam joined 22:14 sergot left, sergot joined
jnthn FROGGS: ah, that's the one that hung on Moar? 22:14
FROGGS jnthn: yes
jnthn I'll bet those two failures are related. 22:15
FROGGS I will not bet against that :o)
TimToady: btw, I propose we allow: use Foo::Bar:ver<v1.2> <&baz> if $*VM<name> eq 'parrot'; 22:17
lizmat FROGGS: a conditional use ? 22:18
FROGGS yes
lizmat what does the <&baz> mean?
FROGGS import that sub
lizmat export ?
ok
timotimo that's kind of scary, though 22:19
FROGGS I just did that to show the ordering
timotimo: see the backlog... evaling pir code that is put into a heredoc is even worse IMO
lizmat I'm not sure how that would work with serialization
timotimo aye 22:20
FROGGS lizmat: if the condition is not true, it is like you never put that use there IMO
22:21 rindolf left
lizmat but if you serialized that module with/without (wrote to disc) and then later tried to run with a different VM ? 22:22
timotimo er, would that ever work?
lizmat I'm not sure 22:23
but allowing this type of thing
would make it impossible for sure :-)
FROGGS lizmat: when we precompile things, then only for a single vm
and for only the current compiler installation
that stuff is not portable anyway
lizmat but can different VM's share the same installed modules? 22:24
FROGGS only the source
lizmat so at least compiled modules need to have some distinguishing extension to allow them to co-exist in the same installed modules tree
jnthn They can't share the compiled output.
lizmat: They already do: .pir, .jar and .moarvm. 22:25
Can't expect MoarVM to run JVM bytecode, or vice versa :)
lizmat ok, then I have no qualms about allowing conditional use :-)
FROGGS I guess TimToady will have some complaints, and I hope he knows an answer to our current problem of backend specific modules like MIME::Base64 22:26
(except than "fix require and precomp") 22:27
22:27 yakudza joined 22:28 nebuchadnezzar left
TimToady can't think about anything till he gives his talk in a few hours 22:29
22:29 nebuchadnezzar joined
FROGGS yeah, I know that feeling :o) 22:30
TimToady: where is he?
22:31 REPLeffect left
araujo TimToady, talk about?.. will be streamed? :) 22:36
lizmat sleeping a short noght 22:44
night rather& 22:45
FROGGS gnight lizmat
rj* and mj* are updated here: gist.github.com/FROGGS/0ea5537eb675588baaa2 22:59
23:05 denis_boyun left 23:07 spider-mario left
jnthn FROGGS: Checking latest. 23:15
Down to 1 JVM build problem :)
timotimo \o/ 23:16
FROGGS yeah, I left out the Digest::MD5 I fixed locally, but I need to PR that tomorrow
jnthn Lots of JVM test failures still, though. Hm. 23:17
FROGGS yes :/
dalek volaj: 91ca678 | (Tobias Leich)++ | t/CompileTestLib.pm:
fixed typo: libraires -> libraries
 23:18
volaj: 05ba4c3 | jonathan++ | t/CompileTestLib.pm:
Merge pull request #34 from FROGGS/patch-2

fixed typo: libraires -> libraries
volaj: 4330e29 | (Tobias Leich)++ | t/02-simple-args.c:
fix for: warning: too many arguments for format [-Wformat-extra-args]
volaj: 1662d4a | jonathan++ | t/02-simple-args.c:
Merge pull request #33 from FROGGS/patch-1

fix for: warning: too many arguments for format [-Wformat-extra-args]
jnthn I thought the uri issue was gonna be resolved with the role outers fix I did. Evidently not... 23:20
dalek Heuristic branch merge: pushed 29 commits to nqp/jastcompiler by arnsholt 23:21
arnsholtAtHome jnthn: Line annotations was exactly it. jastcompiler and master now get same spectest behaviour 23:22
timotimo oooooh! :D
jnthn arnsholtAtHome: Great! :)
japhb Time to merge?
arnsholtAtHome Should be mergeable, yeah 23:23
The old JASTToJVMBytecode stuff is still there, and should be removed, though
23:24 robinsmidsrod joined
jnthn I'm fine with merging. We're weeks away from the monthly. 23:24
japhb: btw, the moar-conc merge yesterday incidentally should have improved performance a bit with regard to concurrent things on JVM.
japhb jnthn: Oh? That's excellent news. What changed? 23:25
*sigh* r-m from yesterday segfaults after a few seconds of running my spawn-a-pile-of-subprocesses tool, and the children are able to print to the terminal still ... 23:27
jnthn japhb: There's some nqp:: ops and REPRs for stuff in place now, so we don't do it all through the JVM's interop layer.
japhb (well, at all)
excellent.
jnthn tries to figure what the "(well, at all)" attaches to. 23:28
japhb: There'll still most certainly be bugs in that area.
japhb: It was merged out of "can get feedback" rather than "as ready as the JVM" 23:29
japhb jnthn: Normally the children's autoput would be captured by a qx. But with that process dead ... they talk to the terminal. I should think they should get a SIGPIPE instead, which indicates something isn't set up right during spawn. 23:30
jnthn: Oh sure, I recognize that. I just was hoping to see it in action, but my favorite test didn't work, so now I have to try something simpler.
jnthn japhb: Does it explicitly use threads, or is it working with promises? 23:31
japhb Promises.
23:33 dmol left
arnsholtAtHome I'll try to get the merge and other stuff done tomorrow. Sleep now 23:33
G'nite &
jnthn 'night
japhb My uses of thread explicitly: Checking ::('Thread') to see if it's running on a concurrent backend; setting the scheduler to ThreadPoolScheduler.new(:max_threads(17 * $jobs)); printing $*THREAD.id in some debugging code. 23:34
g'night arnsholtAtHome
jnthn ok 23:35
I need to set myself up some stress tests on this stuff.
japhb Worth rebuilding today (meaning, have you done stability fixes to r-m since the merge)? 23:36
23:36 BenGoldberg joined
jnthn No 23:38
Been working on JVM-related fixes to get towards rj-*, and some other Moar stuff
23:40 Adriaaan left, dmol joined 23:41 colomon joined 23:44 dmol left
japhb Hmmm, under r-j my threaded program is hanging on exit now (it would normally print its computed output and exit the program). 23:47
jnthn Does it print the output? 23:48
FROGGS personally I don't like hangs very much
jnthn lizmat reported a tst that seemed to hang at exit. Unfortunately it didn't show up when running with evalserver. 23:49
*test
So I missed it.
It's on my todo list to look at it, but tomorrow. :)
FROGGS gnight 23:50
jnthn 'night, #perl6
japhb jnthn: Yeah, it did print the output.
Good night, sleep well.
jnthn japhb: OK, then hopefully it's same issue
timotimo gnite jnthn :)
jnthn o/
timotimo i was wondering
are we pessimizing every for without a -> $foo because we have <-> $_ every time and may not be using it at all? 23:51
as in: we have "is rw" on that every time but we may not be writing to $_
23:55 fuad joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!