Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2017-10-09

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

All times shown according to UTC.

Time Nick Message
00:54 benchable6 joined #perl6-dev
01:56 ilbot3 joined #perl6-dev
01:56 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
02:03 committable6 joined #perl6-dev
02:19 bisectable6 joined #perl6-dev
03:19 bisectable6 joined #perl6-dev
03:19 committable6 joined #perl6-dev
04:09 Geth ¦ roast: 919a2ddde6 | (Daniel Green)++ | S02-types/native.t
04:09 Geth ¦ roast: Add test for RT #127973
04:09 Geth ¦ roast:
04:09 Geth ¦ roast: Native strings can be as large as regular strings.
04:09 Geth ¦ roast: review: https://github.com/perl6/roast/commit/919a2ddde6
04:09 synopsebot RT#127973 [open]: https://rt.perl.org/Ticket/Display.html?id=127973 [BUG] native strings not allowed to be as big as regular strings
04:32 buggable New CPAN upload: v0.1.30.tar.gz by UGEXE https://cpan.metacpan.org/authors/id/U/UG/UGEXE/Perl6/v0.1.30.tar.gz
05:04 Zoffix "Stage parse      : 101.480"
05:04 * Zoffix weeps
05:52 Zoffix "Stage parse      :  74.669"
05:52 * Zoffix unweeps :)
06:03 bartolin .
06:03 yoleaux 8 Oct 2017 17:47Z <jnthn> bartolin: https://github.com/perl6/nqp/blob/master/src/vm/moar/QAST/QASTOperationsMAST.nqp#L2895 is how it's defined on Moar, so 1 means evaluate to argument 1 (zero-based)
06:06 bartolin jnthn: thanks! so (if i don't misunderstand) it would be ok to change the op to return $RT_INT (since we call nqp::killprocasync with an Int as the second argument). i'll take another look later today.
06:13 Zoffix m: use Test; my ($a, $b); $b = 'b'; :(:$a) := $b; is $a, 'b', 'can bind to single pair';
06:13 camelia rakudo-moar 12fcec: OUTPUT: «not ok 1 - can bind to single pair␤# Failed test 'can bind to single pair'␤# at <tmp> line 1␤# expected: 'b'␤#      got: (Mu)␤»
06:14 Zoffix This throws X::Cannot::Capture with my patch... It's a TODOed test.
06:16 Zoffix m: use Test; my ($a, $b); $b = 'b'; :(:$a) := \(:a($b)); is $a, 'b', 'can bind to single pair';
06:16 camelia rakudo-moar 12fcec: OUTPUT: «ok 1 - can bind to single pair␤»
06:17 Zoffix I'm calling bullshit on it. Even if Capture would not throw and return the self, it'd throw "too many positionals" error. And it's a TODO 'cause Str.Capture returns an empty Capture, so $a ends up as Mu
06:22 Geth ¦ roast: 4bf399347c | (Zoffix Znet)++ | S06-other/pairs-as-lvalues.t
06:22 Geth ¦ roast: Fix and unfudge incorrect Pair binding test
06:22 Geth ¦ roast:
06:22 Geth ¦ roast: The current version attempts to bind a Str variable to a Signature
06:22 Geth ¦ roast: with one named arg in it. The test is fudged, as :$a gets a Mu
06:22 Geth ¦ roast: value instead of $b's value, and it doesn't outright die only
06:22 Geth ¦ roast: because Str.Capture returns an empty capture.
06:22 Geth ¦ roast:
06:22 Geth ¦ roast: Fix by binding a proper Capture, with $b passed as a named :a arg.
06:22 Geth ¦ roast: review: https://github.com/perl6/roast/commit/4bf399347c
06:23 Geth ¦ roast: 3cbeaea94d | (Zoffix Znet)++ | S06-other/pairs-as-lvalues.t
06:23 Geth ¦ roast: Actually remove the fudge
06:23 Geth ¦ roast:
06:23 Geth ¦ roast: Part of https://github.com/perl6/roast/commit/4bf399347c
06:23 Geth ¦ roast: review: https://github.com/perl6/roast/commit/3cbeaea94d
06:25 Zoffix oh damn, it fails now >_< I thought I did run it
06:26 Zoffix I suck
06:27 Geth ¦ roast: 0a6daaac25 | (Zoffix Znet)++ | S06-other/pairs-as-lvalues.t
06:27 Geth ¦ roast: Fix the bind test
06:27 Geth ¦ roast:
06:27 Geth ¦ roast: Part of fix described in https://github.com/perl6/roast/commit/4bf399347c
06:27 Geth ¦ roast:
06:27 Geth ¦ roast: If actual binding is expected, the $a needs to be is raw.
06:27 Geth ¦ roast: review: https://github.com/perl6/roast/commit/0a6daaac25
06:30 Geth ¦ roast/6.c-errata: 6c6385a162 | (Zoffix Znet)++ | S06-other/pairs-as-lvalues.t
06:30 Geth ¦ roast/6.c-errata: Fix and unfudge incorrect Pair binding test
06:30 Geth ¦ roast/6.c-errata:
06:30 Geth ¦ roast/6.c-errata: The current version attempts to bind a Str variable to a Signature
06:30 Geth ¦ roast/6.c-errata: with one named arg in it. The test is fudged, as :$a gets a Mu
06:30 Geth ¦ roast/6.c-errata: value instead of $b's value, and it doesn't outright die only
06:30 Geth ¦ roast/6.c-errata: because Str.Capture returns an empty capture. In addition, if
06:30 Geth ¦ roast/6.c-errata: actual bind is wanted, the :$a needs to be received raw.
06:30 Geth ¦ roast/6.c-errata:
06:30 Geth ¦ roast/6.c-errata: Fix by binding a proper Capture, with $b passed as a named :a arg.
06:30 Geth ¦ roast/6.c-errata: review: https://github.com/perl6/roast/commit/6c6385a162
06:32 patrickz joined #perl6-dev
06:50 [Tux] This is Rakudo version 2017.09-350-g12fcece49 built on MoarVM version 2017.09.1-576-g64b5dc03
06:50 [Tux] csv-ip5xs        1.247 -  1.352
06:50 [Tux] test            10.373 - 10.969
06:50 [Tux] test-t           3.199 -  3.292
06:50 [Tux] csv-parser       0.944 -  1.006
06:50 Zoffix Gonna add Signature.ACCEPTS: Mu $topic that does self.ACCEPTS: $topic.Capture
06:51 Zoffix Gonna add Signature.ACCEPTS: Mu $topic that does (try self.ACCEPTS: $topic.Capture) || False
06:52 Geth ¦ rakudo/nom: 4ba12ff17c | (Zoffix Znet)++ | src/core/Exception.pm
06:52 Geth ¦ rakudo/nom: Implement X::Cannot::Capture
06:52 Geth ¦ rakudo/nom:
06:52 Geth ¦ rakudo/nom: To be thrown from .Capture methods of certain types
06:52 Geth ¦ rakudo/nom: per https://irclog.perlgeek.de/perl6/2017-03-07#i_14221839
06:52 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/4ba12ff17c
06:52 [Tux] This is Rakudo version 2017.09-350-g12fcece49 built on MoarVM version 2017.09.1-576-g64b5dc03
06:52 [Tux] csv-ip5xs        1.244 -  1.256
06:52 [Tux] test             9.988 - 10.312
06:52 [Tux] test-t           3.161 -  3.173
06:52 [Tux] csv-parser       0.861 -  0.878
06:52 [Tux] (second double run)
06:53 [Tux] looking good
06:53 Geth ¦ rakudo/nom: bad9fefdd7 | (Zoffix Znet)++ | 8 files
06:53 Geth ¦ rakudo/nom: Make .Capture of certain core types throw
06:53 Geth ¦ rakudo/nom:
06:53 Geth ¦ rakudo/nom: To avoid supersticious, performance-reducing parens;
06:53 Geth ¦ rakudo/nom: e.g. -> ($x) {} where -> $x {} was meant.
06:53 Geth ¦ rakudo/nom:
06:53 Geth ¦ rakudo/nom: Discussion: https://irclog.perlgeek.de/perl6/2017-03-07#i_14221839
06:53 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/bad9fefdd7
06:53 Zoffix buggable: speed 100
06:53 buggable Zoffix, ▄▄▅▃▄▄↑▅▅▅▅↑▇↑▆▄▅▅▅▆▄▅▄▄▅▇▆▅↑↑▅▅▅▅▄▄▄▅▄▄▄▄▅▄▄▆▄▄▃▄▄▄▄▄▅▅▄▅▄▃▆↑█▆▃▂▁▂▂▁▂▂▂▂▃▅▂▃▂▃▆▂▁▁▁▂▂▂▂▂▃▂▂▂▃▃▁▃▁▁ dates: 2017-09-05–2017-10-09; range: 3.154s–4.237s; speed: 10% faster
06:53 Zoffix \o/
07:03 Zoffix ZOFFLOP: t/spec/integration/precompiled.t
07:04 Geth ¦ rakudo/nom: 1ea3297b21 | (Zoffix Znet)++ | src/core/Range.pm
07:04 Geth ¦ rakudo/nom: Block out Mu.new from Range
07:04 Geth ¦ rakudo/nom:
07:04 Geth ¦ rakudo/nom: To avoid throwing incorrect error that Range.new takes only
07:04 Geth ¦ rakudo/nom: named args
07:04 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1ea3297b21
07:07 Geth ¦ roast: 6ac1f272f4 | (Zoffix Znet)++ | S02-types/range.t
07:07 Geth ¦ roast: Test Range.new with wrong args
07:07 Geth ¦ roast:
07:07 Geth ¦ roast: Does not claim it takes only named args
07:07 Geth ¦ roast:
07:07 Geth ¦ roast: Rakudo fix: https://github.com/rakudo/rakudo/commit/1ea3297b21
07:07 Geth ¦ roast: review: https://github.com/perl6/roast/commit/6ac1f272f4
07:41 lizmat Files=1227, Tests=75620, 343 wallclock secs (14.66 usr  5.17 sys + 2413.72 cusr 208.00 csys = 2641.55 CPU)
07:41 yoleaux 8 Oct 2017 17:43Z <jnthn> lizmat: iirc NQP uses some codes that aren't used in Rakudo, and Rakudo supported them for that reason. I think you asked me about them a while back 'cus you were considering removing them from Mu.BUILDALL.
07:41 yoleaux 8 Oct 2017 17:44Z <jnthn> lizmat: If you changed the meanings of the integers and didn't align NQP too then it's possible something coulda broken.
07:41 yoleaux 8 Oct 2017 17:48Z <jnthn> lizmat: auto-generating BUILD_LEAST_DERIVED would make mixins cheaper; note it's only worth doing for mixin types, but I think you have that information available to you in method compose, or via some .is_mixin method or some such. Can't remember off the top of my head, but I'm sure it's available. So I'd only generate them in that case, not for all classes where it'd be useless
08:14 jnthn bartolin: Yes, fine to change that. Thanks for working on it. :-)
08:19 robertle joined #perl6-dev
08:25 Zoffix m: say 'x'.match(/./).^mro
08:25 camelia rakudo-moar 1ea329: OUTPUT: «((Match) (Capture) (Cool) (Any) (Mu))␤»
08:25 Zoffix TIL a Match is a Capture :o
08:28 W4RL0RD joined #perl6-dev
08:28 releasable6 joined #perl6-dev
08:37 Zoffix Weird...
08:37 Zoffix m: -> (Str() $a, Str() :$b) { say [$a, $b] }('abc'.match(/(.) $<b>=. ./))
08:37 camelia rakudo-moar 1ea329: OUTPUT: «Cannot find method 'gist': no method cache and no .^find_method␤  in block <unit> at <tmp> line 1␤␤»
08:37 Zoffix m: -> ($a, :$b) { say [$a.Str, $b.Str] }('abc'.match(/(.) $<b>=. ./))
08:37 camelia rakudo-moar 1ea329: OUTPUT: «[a b]␤»
08:37 Zoffix m: -> ($a, :$b) { say [$a.^name, $b.^name] }('abc'.match(/(.) $<b>=. ./))
08:37 camelia rakudo-moar 1ea329: OUTPUT: «[Match Match]␤»
08:38 Zoffix Looks like signature's Str() coercer results in a different value
08:54 Zoffix m: use Test; is-deeply set(42, <70>).Capture, \("42" => True, "70" => True)
08:54 camelia rakudo-moar 1ea329: OUTPUT: «not ok 1 - ␤# Failed test at <tmp> line 1␤# expected: \("42" => Bool::True, "70" => Bool::True)␤#      got: \("42" => Bool::True, "70" => Bool::True)␤»
08:54 Zoffix Expected and Got are the same... wtf
08:57 Zoffix m: dd (\("42" => True, "70" => True)).Capture::hash
08:57 camelia rakudo-moar 1ea329: OUTPUT: «Map.new(())␤»
08:57 Zoffix Hm, these end up as positionals :\
09:00 Zoffix Ah, OK, it's meant to.
09:11 Zoffix ZOFLOP: t/spec/S17-procasync/kill.t
09:11 Zoffix ZOFVM: Files=1277, Tests=152640, 150 wallclock secs (20.83 usr  3.28 sys + 3264.45 cusr 169.26 csys = 3457.82 CPU)
09:18 Geth ¦ roast: ca273503b6 | (Zoffix Znet)++ | S02-types/capture.t
09:18 Geth ¦ roast: Spec .Capture on core types
09:18 Geth ¦ roast:
09:18 Geth ¦ roast: Discussions:
09:18 Geth ¦ roast: https://irclog.perlgeek.de/perl6/2017-03-07#i_14221839
09:18 Geth ¦ roast: https://irclog.perlgeek.de/perl6-dev/2017-10-08#i_15275349
09:18 Geth ¦ roast: https://irclog.perlgeek.de/perl6-dev/2017-10-08#i_15274676
09:18 Geth ¦ roast: https://irclog.perlgeek.de/perl6-dev/2017-05-24#i_14629113
09:18 Geth ¦ roast: review: https://github.com/perl6/roast/commit/ca273503b6
09:19 Geth ¦ rakudo/nom: cd5864cfc1 | (Zoffix Znet)++ | 5 files
09:19 Geth ¦ rakudo/nom: Implement .Capture on types per spec
09:19 Geth ¦ rakudo/nom:
09:19 Geth ¦ rakudo/nom: Spec: https://github.com/perl6/roast/commit/ca273503b6
09:19 Geth ¦ rakudo/nom:
09:19 Geth ¦ rakudo/nom: Discussions:
09:19 Geth ¦ rakudo/nom: https://irclog.perlgeek.de/perl6/2017-03-07#i_14221839
09:19 Geth ¦ rakudo/nom: https://irclog.perlgeek.de/perl6-dev/2017-10-08#i_15275349
09:19 Geth ¦ rakudo/nom: https://irclog.perlgeek.de/perl6-dev/2017-10-08#i_15274676
09:19 Geth ¦ rakudo/nom: https://irclog.perlgeek.de/perl6-dev/2017-05-24#i_14629113
09:19 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/cd5864cfc1
09:40 Zoffix So... caller's language version. A var like $/ to store it in sounds good for it?   This stuff's a bit over my head, but I'd like to get 6.d moving
09:44 Zoffix Actually, I'm not 100% sure why $*PERL.version has to be wrong.
09:49 robertle_ joined #perl6-dev
09:54 Zoffix ZOFFLOP: t/spec/S17-promise/stress.t
10:17 stmuk m: use v6.d.PREVIEW; say $*PERL.version;
10:17 camelia rakudo-moar cd5864: OUTPUT: «v6.d␤»
10:19 Zoffix Looks like t/spec/S17-channel/stress.t hung altogether now.
10:19 Zoffix Oh maybe it finished... Pressing CTRL+C didn't react right away, but when it was done it said all tests pass, with ridiculous runtime Files=1277, Tests=152645, 741 wallclock secs (21.19 usr  3.45 sys + 4085.28 cusr 179.47 csys = 4289.39 CPU)
10:20 stmuk m: use v6.c; { use v6.d.PREVIEW }; say $*PERL.version;
10:20 camelia rakudo-moar cd5864: OUTPUT: «v6.d␤»
10:22 Zoffix That will throw. You can only specify one language version per comp unit
10:22 Zoffix But $*PERL.version still ain't working right when modules and precompilation is involved. https://github.com/perl6/6.d-prep/blob/master/TODO/FEATURES.md#need-a-way-to-know-callers-language
10:24 Geth ¦ rakudo/nom: 27131ed8d1 | (Zoffix Znet)++ | src/core/Signature.pm
10:24 Geth ¦ rakudo/nom: Make Signature.ACCEPTS coerce args to Capture
10:24 Geth ¦ rakudo/nom:
10:24 Geth ¦ rakudo/nom: - Instead of relying to Any.ACCEPTS to always give False
10:24 Geth ¦ rakudo/nom: - We already did do this for Positional and Associative arguments
10:24 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/27131ed8d1
10:25 Geth ¦ roast: 48d5d39e2b | (Zoffix Znet)++ | S03-smartmatch/capture-signature.t
10:25 Geth ¦ roast: Spec Signature.ACCEPTS with non-Signature/Capture args
10:25 Geth ¦ roast:
10:25 Geth ¦ roast: Rakudo impl: https://github.com/rakudo/rakudo/commit/27131ed8d1
10:25 Geth ¦ roast: review: https://github.com/perl6/roast/commit/48d5d39e2b
11:17 pmurias joined #perl6-dev
11:46 Geth ¦ roast: 064707a320 | (Zoffix Znet)++ | S32-num/expmod.t
11:46 Geth ¦ roast: Add .expmod with negative power tests
11:46 Geth ¦ roast:
11:46 Geth ¦ roast: RT#130713: https://rt.perl.org/Ticket/Display.html?id=130713
11:46 Geth ¦ roast: review: https://github.com/perl6/roast/commit/064707a320
11:46 synopsebot RT#130713 [open]: https://rt.perl.org/Ticket/Display.html?id=130713 42.expmod(-1,1) hangs
11:53 * Zoffix observes that writing a test while fixing a bug takes a lot less time than writing one for a fix that happened awhile back. While fixing you know what the cause it, so you can cover that cause right away.
11:53 Zoffix buggable: tag testneeded
11:53 buggable Zoffix, There are 55 tickets tagged with TESTNEEDED; See http://fail.rakudo.party/t/TESTNEEDED for details
11:53 Zoffix I'm surprised so many tickets are left over... I was sure they'd go like hot cakes during squashathon
12:00 Geth ¦ rakudo/nom: 56eef6967b | (Elizabeth Mattijsen)++ | 2 files
12:00 Geth ¦ rakudo/nom: Fix for RT #132249
12:00 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/56eef6967b
12:01 synopsebot RT#132249 [new]: https://rt.perl.org/Ticket/Display.html?id=132249 [LTA] use lib ‘’ spills guts (use lib ‘’)
12:14 ggoebel joined #perl6-dev
12:15 ggoebel zoffix: have you added https://fail.rakudo.party/t/TESTNEEDED to hacktoberfest?
13:17 perlpilot joined #perl6-dev
13:23 MasterDuke_ joined #perl6-dev
13:32 evalable6 joined #perl6-dev
13:51 jnthn Just build HEAD and t/02-rakudo-repl.t is very faily
13:54 jnthn *built
13:55 robertle joined #perl6-dev
13:56 skids joined #perl6-dev
14:48 MasterDuke_ anyone mind merging https://github.com/rakudo/rakudo/pull/1184 ?
14:49 Geth ¦ rakudo/nom: 16c2a15767 | (Daniel Green)++ | src/Perl6/Grammar.nqp
14:49 Geth ¦ rakudo/nom: Use cached line numbers for token label
14:49 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/16c2a15767
14:49 Geth ¦ rakudo/nom: f28875f109 | (Jonathan Worthington)++ (committed using GitHub Web editor) | src/Perl6/Grammar.nqp
14:49 Geth ¦ rakudo/nom: Merge pull request #1184 from MasterDuke17/use_cached_line_numbers_for_token_label
14:49 Geth ¦ rakudo/nom:
14:49 Geth ¦ rakudo/nom: Use cached line numbers for token label
14:49 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f28875f109
14:49 jnthn Looks sensible enough to me :)
14:49 MasterDuke_ jnthn: thanks
14:58 ugexe that test is failing regularly on windows now as well
14:59 ugexe all windows failures started on the full-make-test merge, so maybe that is the cause of it elsewhere too
15:00 ugexe oh those tests probably werent run before at all
15:15 Geth ¦ nqp/master: 4 commits pushed by pmurias++
15:15 Geth ¦ nqp/master: 16fff80593 | [js] Throw when unboxing an bigint that doesn't fit in 64bits
15:15 Geth ¦ nqp/master: b94eb1e15a | Test that converting huge bignums to ints throws exceptions
15:15 Geth ¦ nqp/master: 69261edf2d | [js] Fix nqp::setdimensions return value
15:15 Geth ¦ nqp/master: 6b4207bd4f | Test nqp::setdimensions return value
15:15 Geth ¦ nqp/master: review: https://github.com/perl6/nqp/compare/943f7f757c...6b4207bd4f
15:31 travis-ci joined #perl6-dev
15:31 travis-ci NQP build failed. pmurias 'Test nqp::setdimensions return value'
15:31 travis-ci https://travis-ci.org/perl6/nqp/builds/285636737 https://github.com/perl6/nqp/compare/943f7f757c9b...6b4207bd4fc4
15:31 travis-ci left #perl6-dev
15:39 Zoffix jnthn: "faily" means what? You always see it fail or does it flop?
15:40 jnthn Zoffix: Fails some of its tests, reliably.
15:40 jnthn Seems like the same ones
15:41 Zoffix jnthn: can you pastebin the output?
15:42 jnthn Zoffix: https://gist.github.com/jnthn/f0bab3cce7efa467a9de8eb7ee4fdeed
15:42 Zoffix Thanks.
15:42 jnthn I guess if you're asking, you ain't seeing them?
15:43 Zoffix jnthn: nope. And on my box, the stuff REPL is fed with isn't part of the tested output, while on your box it is.
15:44 jnthn Odd
15:44 Zoffix jnthn: what OS is that?
15:44 jnthn Linux. Ubuntu 16.04
15:44 jnthn So, about as boring as possible. :)
15:45 Zoffix I'm Bodhi Linux, based off Ubuntu 14.04
15:45 jnthn fwiw, the repl seems to function correctly if I enter it
15:45 Zoffix jnthn: do you Readline/Lineoise installed?
15:45 jnthn Hmm
15:46 jnthn `use Readline` works
15:46 jnthn So seems I have that one
15:46 jnthn `use Linenoise` doesn't
15:46 * Zoffix tries installing it
15:46 jnthn Though I should check if that's what the module is called
15:47 jnthn Yeah, it is. So, just Readline
15:47 jnthn And I don't get the text suggesting that I install Readline
15:47 jnthn Which I guess suggests it's being used
15:55 gfldex zoffix++ # for closing my tickets
15:56 ugexe multi sub is-run-repl isn't closing handles fwiw... maybe thats why windows is failing that test
15:56 ugexe can't test atm
16:01 Geth ¦ rakudo/nom: 1ce3a36df4 | (Zoffix Znet)++ | t/packages/Test/Helpers.pm6
16:01 Geth ¦ rakudo/nom: Implement :$line-editor option
16:01 Geth ¦ rakudo/nom:
16:01 Geth ¦ rakudo/nom: - Sets REPL_LINE_EDITOR for duration of the test
16:01 Geth ¦ rakudo/nom: - Default to `none`, which prevents spurious failures for
16:01 Geth ¦ rakudo/nom: users who have one of the recommended line editors installed.
16:01 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/1ce3a36df4
16:01 Geth ¦ rakudo/nom: 973f88dc65 | (Zoffix Znet)++ | t/packages/Test/Helpers.pm6
16:01 Geth ¦ rakudo/nom: Document is-run-repl sets ERROR_COLOR option
16:01 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/973f88dc65
16:04 Geth ¦ rakudo/nom: 484f987259 | (Zoffix Znet)++ | t/packages/Test/Helpers.pm6
16:04 Geth ¦ rakudo/nom: Close pipes in is-run-repl; ugexe++
16:04 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/484f987259
16:08 * Zoffix backlogs #perl6 a bit
16:08 jnthn Zoffix: That looks to have helped, nice :)
16:08 Zoffix lizmat: dam. looks like all the Cool method that take and pass a Capture are uber slower
16:08 Zoffix m: .contains: 'meow' for ^10000_00; say now - INIT now
16:08 camelia rakudo-moar f28875: OUTPUT: «11.0881977␤»
16:08 Zoffix m: .Str.contains: 'meow' for ^10000_00; say now - INIT now
16:08 camelia rakudo-moar f28875: OUTPUT: «0.6811970␤»
16:09 Zoffix And there's lots of them
16:09 lizmat yeah, if you look at the profile
16:09 Zoffix m: say 11/.68
16:09 camelia rakudo-moar f28875: OUTPUT: «16.176471␤»
16:09 lizmat you'll see that it cannot cache the method lookup
16:09 Zoffix :(
16:11 lizmat Routine.find_best_dispatchee being called all the time
16:11 lizmat taking 30% of time
16:12 lizmat which also implies that the proto for Str.contains gets hammered, which takes another 30%
16:13 lizmat which then selects Cool.contains, taking another 15% of time
16:13 AlexDaniel_ joined #perl6-dev
16:13 lizmat to the actual workhorse, which takes 7% of time (and which somehow isn't jitted)
16:14 jnthn Yeah, flattening things is not yet well optimized.
16:14 jnthn I've been pondering how we might handle those a bit better down in the VM, so the benefit of both multi cache and spesh
16:14 jnthn *for the benefit
16:15 lizmat jnthn: can it be that a multi-candidate doesn't get jitted because its proto doesn't get jitted ?
16:16 jdv79 yup.  600M isnt enough mem to build.  OOMed
16:16 jnthn lizmat: No, they'd be independent, and often protos never get jitted
16:16 lizmat ah ok
16:16 jnthn 'cus they get bypassed entirely by a cache lookup
16:16 Zoffix jdv79: you need ~1.2GB
16:16 jdv79 hopefully 1.5G will be enough
16:16 jnthn So the call is directly into the candidate
16:17 jdv79 ok.  itll just take all day since 1G of that is on a SD card
16:17 jnthn But the cache doesn't handle flattening
16:18 lizmat another weird thing: in that example I also see "add_to_cache" being called 1M times
16:18 lizmat you would think that once is enough ?
16:19 jnthn It'd be better named try_add_to_cache :)
16:19 jnthn If there's flattening then it won't actually add it
16:21 jnthn fwiw, I'm doing a bit of work on hyper/race this week. Locally for now, will hopefully land something in the repo by the end of the week
16:22 MasterDuke_ very cool
16:25 gfldex jnthn: that is very welcome because I already got a blog post title and the title tends to be the hard part.
16:25 jnthn :P
16:26 jnthn Trick will be not getting distracted debugging something else
16:26 jnthn :)
16:26 jnthn Anyways, home/dinner time now, I think :)
16:34 Zoffix At least some of the failures of t/02-rakudo/repl.t on windows is 'cause $*OUT.tell always returns 0
16:39 dogbert2 joined #perl6-dev
16:41 eater joined #perl6-dev
16:43 robertle joined #perl6-dev
17:02 buggable New CPAN upload: App-Mi6-0.0.6.tar.gz by SKAJI https://cpan.metacpan.org/authors/id/S/SK/SKAJI/Perl6/App-Mi6-0.0.6.tar.gz
17:23 jdv79 Stage parse      : 3731.644 :)
17:26 Zoffix :)
17:29 MasterDuke_ that handily beats my ~2400 with a debugging, unoptimized moarvm in valgrind
17:30 timotimo i have received a rather confusing comment on the weekly blog today o_O
17:31 timotimo in the moderation queue
17:32 MasterDuke oh?
17:33 Zoffix What was it?
17:34 timotimo about keys for hashes
17:34 timotimo i'll let it go up on the site, though i can't help but feel it's the wrong place for the question?
17:34 timotimo i think it went onto the weekly blog's comment section because it linked to the thread "Tip: assign a value to a hash using a variable as a key" on the users list
17:34 timotimo (because the comment explicitly references it)
17:34 MasterDuke forward to the mailing list?
17:35 Zoffix Well... I was wondering how long it'd take: "No more Texas operators? Political correctness infected Perl 6, too? Sad!"
17:35 timotimo i can't take any person seriously who ends their comments in "Sad!"
17:38 gfldex i'm tempted to feed the troll :->
17:38 lizmat joined #perl6-dev
17:39 * Zoffix wouldn't
17:40 timotimo i'd prefer no troll feeding on the weekly blog comments section
17:40 Zoffix Let them feel sad that no one read their comment :)
17:42 buggable New CPAN upload: App-Mi6-0.0.7.tar.gz by SKAJI https://cpan.metacpan.org/authors/id/S/SK/SKAJI/Perl6/App-Mi6-0.0.7.tar.gz
17:43 gfldex timotimo: you are feeding that troll by proving him with a platform. Esp. as he got nothing else to say.
17:44 timotimo well, it was let through and now it'd seem rather petty to delete it after-the-fact
17:44 Zoffix toss it :)
17:45 [Coke] it's your site, you can be a little petty
17:45 gfldex being able to change your mind when presented with more arguments is no sign of weekness
17:46 gfldex m: sub strong(**@arguments){ say ‚I'm Strong!‘ }; strong(|(1..∞));
17:47 gfldex ok, I was wrong :D
17:47 camelia rakudo-moar 484f98: OUTPUT: «(timeout)»
17:48 Zoffix m: sub strong(**@arguments){ say ‚I'm Strong!‘ }; strong(1..∞);
17:48 camelia rakudo-moar 484f98: OUTPUT: «I'm Strong!␤»
17:48 gfldex that's just one argument, even though it is quite a big one
17:51 gfldex m: sub strong(*@arguments){ say ‚I'm Strong!‘ }; strong(|(1..∞));
17:51 gfldex that leaves the question if a slurpy should explode a Slip
17:51 camelia rakudo-moar 484f98: OUTPUT: «(timeout)»
17:52 timotimo [Coke]: fwiw, i'd say it's liz' site by now :)
17:52 Zoffix It likely hangs before it gets to any slurpies
17:52 Zoffix m: (1..∞).FLATTENABLE_LIST
17:52 camelia rakudo-moar 484f98: OUTPUT: «(timeout)»
18:51 lizmat_ joined #perl6-dev
18:51 Geth ¦ nqp: 0b6fa59726 | pmurias++ | src/vm/js/nqp-runtime/reprs.js
18:51 Geth ¦ nqp: [js] Treat a non primitive type to multidim arrays as an object
18:51 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/0b6fa59726
18:51 Geth ¦ nqp: 01429c3c15 | pmurias++ | src/vm/js/nqp-runtime/reprs.js
18:51 Geth ¦ nqp: [js] Make nqp::atposnd return null when there is a missing element
18:51 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/01429c3c15
18:52 Geth ¦ nqp: ce60c2bf38 | pmurias++ | t/nqp/102-multidim.t
18:52 Geth ¦ nqp: Fix test count
18:52 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/ce60c2bf38
18:52 Geth ¦ nqp: faa67169a3 | pmurias++ | t/nqp/094-clone.t
18:52 Geth ¦ nqp: More multidim tests
18:52 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/faa67169a3
18:55 AlexDaniel_ .tell jnthn you told me to ping you later, so here it is. Plz give Zoffix or me more privileges for Rakudo and MoarVM repos so that we can manage webhooks. Alternatively, go and remove webhooks to 94.23.219.181
18:55 yoleaux AlexDaniel_: I'll pass your message to jnthn.
18:56 jnthn Ah, something my tired brane might actually usefull be able to do tonight...
18:56 yoleaux 18:55Z <AlexDaniel_> jnthn: you told me to ping you later, so here it is. Plz give Zoffix or me more privileges for Rakudo and MoarVM repos so that we can manage webhooks. Alternatively, go and remove webhooks to 94.23.219.181
19:05 travis-ci joined #perl6-dev
19:05 travis-ci NQP build failed. pmurias '[js] Make nqp::atposnd return null when there is a missing element'
19:05 travis-ci https://travis-ci.org/perl6/nqp/builds/285733487 https://github.com/perl6/nqp/compare/6b4207bd4fc4...01429c3c1566
19:05 travis-ci left #perl6-dev
19:06 jnthn AlexDaniel_: I've tried it for the MoarVM repo (so you and zoffix should have invites for that); let me know if you can get to the webhooks page
19:06 jnthn (Also, I removed the one you meneionted since I was there anyway)
19:06 jnthn Provided I got it right I'll do the same for the Rakudo one :)
19:06 Zoffix Joined in, but no webhook access
19:07 AlexDaniel_ same
19:07 jnthn d'oh
19:07 jnthn try again now?
19:08 AlexDaniel_ that works
19:08 jnthn \o/
19:09 jnthn And should have it for Rakudo too :)
19:09 Zoffix jnthn++
19:09 Zoffix jnthn: that also gave me commit bit to MoarVM. What's the policy for merging/committing stuff?
19:10 jnthn Yeah, you'll both have quite a few extra perms as a result of this...
19:10 jnthn For MoarVM - typo fixes and other stuff nobody could argue with, just push/merge
19:11 AlexDaniel_ mhm github allows more refined control of privileges for access tokens, but that's per user
19:13 lizmat joined #perl6-dev
19:13 jnthn For more serious changes, I'd start out with PR-ing things. It's C, with a GC running around, and a missed MVMROOT for MVM_ASSIGN_REF can means hours of frustrating bug hunt later...
19:13 jnthn s/for/or/
19:13 Zoffix OK. I'll then merge https://github.com/MoarVM/MoarVM/pull/723 once Travis/Appveyor clear it
19:14 92AAB6NK0 joined #perl6-dev
19:14 92AAB6NK0 NQP build failed. pmurias 'More multidim tests'
19:14 92AAB6NK0 https://travis-ci.org/perl6/nqp/builds/285733646 https://github.com/perl6/nqp/compare/01429c3c1566...faa67169a349
19:14 92AAB6NK0 left #perl6-dev
19:14 AlexDaniel_ uh… that was a very interesting nickname
19:17 jnthn Zoffix: Yeah, that looks fine
19:20 * jnthn bbiab
19:25 Geth ¦ roast: 63950b4b52 | (Zoffix Znet)++ | S32-io/tell.t
19:25 Geth ¦ roast: Test .tell on TTY handles gives faked out values
19:25 Geth ¦ roast:
19:25 Geth ¦ roast: RT#132254: https://rt.perl.org/Ticket/Display.html?id=132254
19:25 Geth ¦ roast: MoarVM fix: https://github.com/MoarVM/MoarVM/pull/723
19:25 Geth ¦ roast: review: https://github.com/perl6/roast/commit/63950b4b52
19:25 synopsebot RT#132254 [new]: https://rt.perl.org/Ticket/Display.html?id=132254 [WIN32] $*OUT.tell always gives 0
20:46 AlexDaniel_ unicodable6: hug
20:46 unicodable6 AlexDaniel_, U+0F19 TIBETAN ASTROLOGICAL SIGN SDONG TSHUGS [Mn] ( ༙)
20:46 unicodable6 AlexDaniel_, U+D6C5 Hangul Syllable HUG [Lo] (훅)
20:46 unicodable6 AlexDaniel_, 5 characters in total: https://gist.github.com/e3393fdfded96991d9b330b9e697d4b4
20:48 lizmat and another Perl 6 Weekly hits the Net: https://p6weekly.wordpress.com/2017/10/09/2017-41-the-case-for-empathy/
20:49 Zoffix woooo
20:52 AlexDaniel_ lizmat: have you seen https://irclog.perlgeek.de/perl6/2017-10-08#i_15275712 ?
20:54 Zoffix lizmat++ good weekly
20:55 bartolin yes, very much lizmat++
20:57 AlexDaniel_ cool weekly. lizmat++
20:58 AlexDaniel_ lizmat: “More than 250 commits” did you mean 150?
20:59 AlexDaniel_ also, maybe it's worth not mentioning it given that Zoffix++ doubled that number unintentionally
20:59 teatime haha
21:00 Zoffix (if you're going in to mend the weekly, there's an unclosed paren after `(such as @a[3..*]`  )
21:01 lizmat AlexDaniel_ : oops, my math skills need some tuning  :-)
21:02 lizmat AlexDaniel_ Zoffix fixed and fixed  :-)
21:03 AlexDaniel_ lizmat: the actual number is 110
21:04 AlexDaniel_ so over 100 is fair to say I guess :)
21:04 lizmat well, I just did a git describe on the last commit of last Monday, and the one now
21:04 AlexDaniel_ lizmat: (it's because of commits like this: https://github.com/rakudo/rakudo/commit/1e747e01f880ed9626b417d166361bf763759972 )
21:04 lizmat and the difference is more than 150
21:05 AlexDaniel_ yeah but 40 commits show up twice… once normal and once rebased :S
21:05 lizmat AlexDaniel_: lalala I can't hear you  :)
21:05 AlexDaniel_ :D
21:06 lizmat (fixed  :-)
21:06 * lizmat steps away from the keyboard for a bit
21:06 bartolin hmm, we have SIGHUP as the default for method kill: https://github.com/rakudo/rakudo/blob/484f987259bd8b1bbb356780f1a0315ae9d327e3/src/core/Proc/Async.pm#L409
21:07 bartolin wouldn't it make more sense to use SIGTERM there?
21:08 AlexDaniel_ bartolin: have you tried sending SIGTERM to a process group?
21:09 bartolin AlexDaniel_: no, i didn't. i didn't think about a process group until now :-)
21:09 AlexDaniel_ bartolin: what happens sometimes is that the PID you're sending it too exits, but everything else stays alive
21:10 AlexDaniel_ so in theory we'd want to have something like, errr, -SIGTERM
21:10 AlexDaniel_ or better some named argument to do the same
21:11 AlexDaniel_ bartolin: but so far I don't even think that we have a ticket for that :S
21:11 * bartolin is trying to implement nqp::killprocasync for the jvm backend and the provided way to kill a started process there is destroy (what calls SIGTERM under the hood). https://docs.oracle.com/javase/8/docs/api/java/lang/Process.html#destroy--
21:23 bartolin i googled a bit (no experiments, yet), but didn't see specific advices against sending SIGTERM to a process group. however, SIGHUP looks a bit surprising to me
21:24 AlexDaniel_ it's not that it's bad or anything, it's just that in some cases it won't kill much
21:27 bartolin yes, and also both the command and the shell 'kill' builtin default to SIGTERM afaik
21:28 bartolin at least there are not tests in roast wrt SIGHUP
22:03 skids joined #perl6-dev
22:13 ugexe i wonder if its related to windows not generating sigterm
22:15 teatime does windows do signals at all
22:15 ugexe http://docs.libuv.org/en/v1.x/signal.html
22:16 ugexe the first part is "here is how windows is handled"
22:25 jnthn lizmat++ # weekly :)
22:30 ggoebel joined #perl6-dev
22:38 BenGoldberg joined #perl6-dev
22:59 Geth ¦ nqp: a65dbe0c77 | (Zoffix Znet)++ | tools/build/MOAR_REVISION
22:59 Geth ¦ nqp: Bump MoarVM
22:59 Geth ¦ nqp: review: https://github.com/perl6/nqp/commit/a65dbe0c77
22:59 Geth ¦ nqp: version bump brought these changes: https://github.com/MoarVM/MoarVM/compare/2017.09.1-575-gd4e230a...2017.09.1-580-g631b3bf
22:59 Geth ¦ rakudo/nom: e11f4ea266 | (Zoffix Znet)++ | tools/build/NQP_REVISION
22:59 Geth ¦ rakudo/nom: Bump NQP
22:59 Geth ¦ rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e11f4ea266
22:59 Geth ¦ rakudo/nom: version bump brought these changes: https://github.com/perl6/nqp/compare/2017.09-66-g943f7f7...2017.09-75-ga65dbe0
23:11 Zoffix ZOFVM: Files=1277, Tests=152647, 150 wallclock secs (21.11 usr  3.34 sys + 3248.13 cusr 169.64 csys = 3442.22 CPU)
23:14 travis-ci joined #perl6-dev
23:14 travis-ci NQP build failed. Zoffix Znet 'Bump MoarVM'
23:14 travis-ci https://travis-ci.org/perl6/nqp/builds/285823905 https://github.com/perl6/nqp/compare/faa67169a349...a65dbe0c7775
23:14 travis-ci left #perl6-dev
23:52 timotimo jnthn: i think cro-http would be a good candidate to get a HTTP tag in its meta info

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