Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-dev, 2016-11-07

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

All times shown according to UTC.

Time Nick Message
00:07 travis-ci joined #perl6-dev
00:07 travis-ci Rakudo build errored. Elizabeth Mattijsen 'Give 1dimmed arrays its own .antipairs
00:07 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173746785 https://github.com/rakudo/rakudo/compare/c5c600112561...1c425f972644
00:07 travis-ci left #perl6-dev
00:07 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
00:50 mst I ... don't think this is right
00:50 mst m: 1...10 >>*>> 2
00:50 camelia rakudo-moar 1c425f: ( no output )
00:50 mst m: say 1...10 >>*>> 2
00:50 camelia rakudo-moar 1c425f: OUTPUT«(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20)␤»
00:50 mst m: say 1,2...10 >>*>> 2
00:50 camelia rakudo-moar 1c425f: OUTPUT«(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20)␤»
00:50 mst m: say 1,2,4...10 >>*>> 2
00:50 camelia rakudo-moar 1c425f: OUTPUT«(1 2 4 8 16)␤»
00:51 mst the compile time expansion of the arithmetic sequence is fucking up precedence, I guess?
01:21 viki ... is lower precedence
01:25 viki m: say 2 + 2 «*» 10
01:25 camelia rakudo-moar 1c425f: OUTPUT«22␤»
01:25 viki m: say 2 * 2 «+» 10
01:25 camelia rakudo-moar 1c425f: OUTPUT«14␤»
01:25 viki m: say (1,2,4…10) »*» 2
01:25 camelia rakudo-moar 1c425f: OUTPUT«(2 4 8 16)␤»
02:49 ilbot3 joined #perl6-dev
02:49 Topic for #perl6-dev is now Perl 6 language and compiler development 2.0 | Logs at http://irclog.perlgeek.de/perl6-dev/today
03:36 MasterDuke jnthn: i'm not following something. just to see what's going on i added a print in lineof (https://github.com/perl6/nqp/blob/master/src/HLL/Compiler.nqp#L664). i see it when building nqp, but it's not getting triggered when building rakudo
05:37 TimToady joined #perl6-dev
06:49 RabidGravy joined #perl6-dev
07:19 [Tux] This is Rakudo version 2016.10-196-g1c425f9 built on MoarVM version 2016.10-39-g4b9e3a8
07:19 [Tux] csv-ip5xs        3.060
07:19 [Tux] test            14.601
07:19 [Tux] test-t           6.931
07:19 [Tux] csv-parser      16.426
07:50 dalek roast: 779d02e | moritz++ | S12-construction/TWEAK.t:
07:50 dalek roast: Initial tests for submethod TWEAK
07:50 dalek roast: review: https://github.com/perl6/roast/commit/779d02eee3
07:50 moritz timotimo: time to merge your branch ^^
08:40 dalek rakudo/POSTBUILD_method: 6b8d481 | moritz++ | t/spectest.data:
08:40 dalek rakudo/POSTBUILD_method: Run TWEAK tests
08:40 dalek rakudo/POSTBUILD_method: review: https://github.com/rakudo/rakudo/commit/6b8d481bea
08:41 tadzik so, TWEAK, eh?
08:41 tadzik it's a nice name :)
08:41 tadzik timotimo++ for putting this in motion :)
09:18 lizmat Files=1151, Tests=53712, 214 wallclock secs (12.77 usr  3.66 sys + 1301.86 cusr 122.57 csys = 1440.86 CPU)
11:03 lizmat m: my @a[2;2] = <a b>,<c d>; say @a[0;0]:!exists  # buglet
11:03 camelia rakudo-moar 1c425f: OUTPUT«Partially dimensioned views of arrays not yet implemented. Sorry. ␤  in block <unit> at <tmp> line 1␤␤»
11:03 lizmat m: my @a[2;2] = <a b>,<c d>; say @a[0;0]:exists
11:03 camelia rakudo-moar 1c425f: OUTPUT«True␤»
11:07 dalek rakudo/js: b48c448 | (Pawel Murias)++ | src/vm/js/perl6-runtime/runtime.js:
11:07 dalek rakudo/js: [js] Call $$decont directly instead of using the no longer present nqp.op.decont.
11:07 dalek rakudo/js: review: https://github.com/rakudo/rakudo/commit/b48c4481bc
11:07 dalek rakudo/js: 91dd06e | (Pawel Murias)++ | src/vm/js/ (2 files):
11:07 dalek rakudo/js: [js] Fix some bugs after null -> Null transition.
11:07 dalek rakudo/js: review: https://github.com/rakudo/rakudo/commit/91dd06ea9a
11:16 lizmat re: TWEAK, has anybody looked at the influence that has on general object building ?
11:16 lizmat performance wise, I mean ?
11:22 dalek rakudo/nom: 0b97362 | lizmat++ | src/core/array_slice.pm:
11:22 dalek rakudo/nom: postcircumfix[;] 9x faster 2/3 dimmed arrays
11:22 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/0b973628ee
11:30 pmurias joined #perl6-dev
11:38 dalek rakudo/nom: ce85ba3 | lizmat++ | src/core/array_slice.pm:
11:38 dalek rakudo/nom: assign to [;] for 2/3 dimmed arrays 10x faster
11:38 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/ce85ba378d
11:53 dalek nqp: 34828da | (Pawel Murias)++ | src/vm/js/bin/run_tests_bootstrapped.pl:
11:53 dalek nqp: [js] Update our nqp-js-on-js test run to take in account changes where the compiled code is stored.
11:53 dalek nqp: review: https://github.com/perl6/nqp/commit/34828da19d
11:53 dalek nqp: a620574 | (Pawel Murias)++ | src/vm/js/Operations.nqp:
11:53 dalek nqp: [js] Remove trailing space.
11:53 dalek nqp: review: https://github.com/perl6/nqp/commit/a6205747fa
11:53 dalek nqp: 3c6892a | (Pawel Murias)++ | src/vm/js/RegexCompiler.nqp:
11:53 dalek nqp: [js] $¢ is declared by the surrounding Block no need to do weird stuff with it.
11:53 dalek nqp: review: https://github.com/perl6/nqp/commit/3c6892a558
11:56 dalek rakudo/nom: bb662ff | lizmat++ | / (4 files):
11:56 dalek rakudo/nom: Split off postcircumfix[;] candidates
11:56 dalek rakudo/nom:
11:56 dalek rakudo/nom: into a separate multidim_slice.pm file, as there will be plenty more
11:56 dalek rakudo/nom: candidates coming and most of it will be rewritten by me anyway.
11:56 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/bb662ff6cf
12:09 travis-ci joined #perl6-dev
12:09 travis-ci Rakudo build errored. Elizabeth Mattijsen 'postcircumfix[;] 9x faster 2/3 dimmed arrays'
12:09 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173864832 https://github.com/rakudo/rakudo/compare/1c425f972644...0b973628ee20
12:09 travis-ci left #perl6-dev
12:09 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
12:45 travis-ci joined #perl6-dev
12:45 travis-ci Rakudo build failed. Elizabeth Mattijsen 'assign to [;] for 2/3 dimmed arrays 10x faster'
12:45 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173868125 https://github.com/rakudo/rakudo/compare/0b973628ee20...ce85ba378d29
12:45 travis-ci left #perl6-dev
12:45 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
12:59 dalek rakudo/nom: e3e3fef | lizmat++ | src/core/multidim_slice.pm:
12:59 dalek rakudo/nom: Make [;]:exists 7x faster for 2/3 dimmed arrays
12:59 dalek rakudo/nom:
12:59 dalek rakudo/nom: Also fix :!exists, it would break before
12:59 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e3e3fefd34
13:14 TimToady lizmat: I'd say that the existence of a submethod TWEAK can be determined at composition time when we put together the build plan, so shouldn't be a problem in theory
13:15 lizmat ah, indeed  :-)
13:15 ggoebel joined #perl6-dev
13:17 dalek rakudo/nom: 3b9c4c9 | lizmat++ | src/core/multidim_slice.pm:
13:17 dalek rakudo/nom: Make [;]:delete 10 faster for 2/3 dimmed arrays
13:17 dalek rakudo/nom:
13:17 dalek rakudo/nom: Also fix :!delete, it would break before
13:17 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/3b9c4c95b9
13:17 lizmat oops, 10x
13:18 travis-ci joined #perl6-dev
13:18 travis-ci Rakudo build errored. Elizabeth Mattijsen 'Split off postcircumfix[;] candidates
13:18 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173871978 https://github.com/rakudo/rakudo/compare/ce85ba378d29...bb662ff6cfff
13:18 travis-ci left #perl6-dev
13:18 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
13:20 dalek rakudo/nom: b71a8bc | ab5tract++ | src/core/set_operators.pm:
13:20 dalek rakudo/nom: Two-arg version of symmetric difference should DWIM
13:20 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b71a8bc800
13:20 dalek rakudo/nom: 6ab0c48 | ab5tract++ | src/core/set_operators.pm:
13:20 dalek rakudo/nom: Make chained (^) work as expected for bags/mixes
13:20 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/6ab0c48eba
13:20 dalek rakudo/nom: f9fec1c | ab5tract++ | src/core/set_operators.pm:
13:20 dalek rakudo/nom: Unicode op version should not be constrained to returning Setty
13:20 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/f9fec1c3bd
13:20 dalek rakudo/nom: 40704b8 | (Zoffix Znet)++ | src/core/set_operators.pm:
13:20 dalek rakudo/nom: Merge pull request #911 from ab5tract/ab5tract/sym-diff-bag-mix
13:20 dalek rakudo/nom:
13:20 timotimo lizmat: hold on, that's 10 fast 20 me!
13:22 dalek roast: 21fc586 | ab5tract++ | S03-operators/ (3 files):
13:22 dalek roast: Test updated for proper bag/mix symmetric difference (#179)
13:22 dalek roast: review: https://github.com/perl6/roast/commit/21fc586d71
13:27 [Coke] we should probably decomission perl6-release; hasn't been used for the past few releases, only send there in months was a political spammer.
13:27 nine +1
13:33 dalek roast: 6b20963 | (Zoffix Znet)++ | S03-operators/mix.t:
13:33 dalek roast: Adjust plan
13:33 dalek roast:
13:33 dalek roast: Per discussion[^1], the adjustment is needed due to PR#179 removing
13:33 dalek roast: some redundant tests, while forgetting to adjust the plan.
13:33 dalek roast:
13:33 dalek roast: https://github.com/perl6/roast/pull/179#discussion_r86182898
13:33 dalek roast: review: https://github.com/perl6/roast/commit/6b20963c31
13:38 dalek roast: c667421 | (Zoffix Znet)++ | S03-operators/bag.t:
13:38 dalek roast: Unfudge now-passing tests
13:38 dalek roast:
13:38 dalek roast: Tests made passing by https://github.com/rakudo/rakudo/pull/911
13:38 dalek roast:
13:38 dalek roast: RT#124529: https://rt.perl.org/Public/Bug/Display.html?id=124529
13:38 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=124529
13:38 dalek roast: review: https://github.com/perl6/roast/commit/c667421fa3
13:38 mst viki: yes, but why does the precedence of ... *change* ?
13:38 * timotimo misses that context
13:38 viki mst: it doesn't
13:39 viki timotimo: https://irclog.perlgeek.de/perl6-dev/2016-11-07#i_13525128
13:39 viki mst: ** is higher precedence than ...
13:39 viki Or *
13:39 mst viki: ...
13:39 mst viki: please read the last two examples I sent to the bot again, more carefully.
13:40 viki m: say (1,2,4...Inf)[30]
13:40 camelia rakudo-moar 40704b: OUTPUT«1073741824␤»
13:40 viki m: .say for (1,2,4...Inf)[30]
13:40 camelia rakudo-moar 40704b: OUTPUT«1073741824␤»
13:40 viki oh
13:41 viki m: say (1,2,4...Inf)[^30]
13:41 camelia rakudo-moar 40704b: OUTPUT«(1 2 4 8 16 32 64 128 256 512 1024 2048 4096 8192 16384 32768 65536 131072 262144 524288 1048576 2097152 4194304 8388608 16777216 33554432 67108864 134217728 268435456 536870912)␤»
13:41 masak m: say (1,2,4...10) >>*>> 2
13:41 camelia rakudo-moar 40704b: OUTPUT«(2 4 8 16)␤»
13:41 masak mst: the precedence of `...` is very loose. usually needs parentheses when interacting with something as tight as `>>*>>`
13:41 viki mst: the penultimate example stops at 20, because it's one of the generated numbers, the other stops at 16, because 32 is higher than 20
13:42 viki mst: the two examples are basically m: say 1,2,4...20 and say 1,2...20, and they stop at different points because of different generated sequences.
13:42 mst ah BALLS I misread the output of the geometric one
13:42 viki \o/
13:42 masak mst: the one thing that always bites me is a (similar) interaction between `...` and (something like) `.map`
13:43 masak or `map`
13:43 masak very tricksy
13:43 masak I'm not one to superstitiously put in parentheses -- and Perl 5/6 generally allow you to avoid parentheses because the prec levels make sense
13:43 masak but `...` is one of the exceptions, where I tend to put them in, just to be safe :)
13:44 masak also gives the sequence a little bit of a visual-pill-ness
13:46 dalek rakudo/nom: eaf4132 | lizmat++ | src/core/multidim_slice.pm:
13:46 dalek rakudo/nom: Make [;] := foo 10 faster for 2/3 dimmed arrays
13:46 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/eaf4132f46
13:47 lizmat argh...  10x   :-)
13:47 lizmat cutnpastefail
13:48 mst viki: yeah, sorry, thing is
13:48 mst m: 1,2,4...10 >>*>> 2
13:48 camelia rakudo-moar 40704b: ( no output )
13:48 mst m: say 1,2,4...10 >>*>> 2
13:48 camelia rakudo-moar 40704b: OUTPUT«(1 2 4 8 16)␤»
13:48 mst m: say (1,2,4...10) >>*>> 2
13:48 camelia rakudo-moar 40704b: OUTPUT«(2 4 8 16)␤»
13:48 mst viki: ^^ notice how similar those are if you're mst and not paying attention properly
13:48 viki Oh yeah :)
13:48 masak m: say 1,2,4...(10 >>*>> 2)
13:48 camelia rakudo-moar 40704b: OUTPUT«(1 2 4 8 16)␤»
13:49 masak much clearer :)
13:49 mst masak: yeah, I understand fine now
13:49 masak thought so
13:49 mst just explaining precisely how I confused myself
13:49 mst since I prefer to be laughed at accurately
13:49 viki I guess ... takes 2 nums of the LHS and say prints the "1" as is.
13:49 * masak .oO( How to be Properly Confused, a mini-course )
13:49 viki m: .say for 1,2,4...10 >>*>> 2
13:49 camelia rakudo-moar 40704b: OUTPUT«1␤2␤4␤8␤16␤»
13:50 masak viki: guess again :)
13:50 viki masak: what is it?
13:50 masak see my eval above
13:50 mst masak: there is actually an entire section in the lesswrong sequences about that.
13:50 viki masak: your eval has the same result?
13:50 viki as without any parens
13:50 masak because that's how natural precedence makes it
13:51 masak `...` is very loose
13:51 masak `say` is a listop here, so that's already out of consideration
13:51 viki Oh, OK. Now I confused myself :D
13:51 masak `...` ends up getting `1,2,4` as lhs and `10 >>*>> 2` as rhs
13:51 viki Yeah
13:56 travis-ci joined #perl6-dev
13:56 travis-ci Rakudo build passed. Elizabeth Mattijsen 'Make [;]:exists 7x faster for 2/3 dimmed arrays
13:56 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173887279 https://github.com/rakudo/rakudo/compare/bb662ff6cfff...e3e3fefd34ce
13:56 travis-ci left #perl6-dev
13:59 dalek roast: 3c6277f | (Zoffix Znet)++ | S03-operators/mix.t:
13:59 dalek roast: Fix broken test
13:59 dalek roast:
13:59 dalek roast: The test was previously fudged and rework in
13:59 dalek roast: https://github.com/perl6/roast/pull/179 left some old code in it, causing
13:59 dalek roast: it to remain failing after unfudging.
13:59 dalek roast: review: https://github.com/perl6/roast/commit/3c6277f59e
14:02 viki ZOFVM: Files=1200, Tests=130105, 142 wallclock secs (20.38 usr  3.20 sys + 2462.70 cusr 227.48 csys = 2713.76 CPU)
14:07 viki awww abstr5ct's PR did break 6.c-errata >_<
14:07 viki So much for my detailed review :/
14:07 viki m: say <a network> (^) <b network> (^) <a network>
14:07 camelia rakudo-moar eaf413: OUTPUT«set(b, network)␤»
14:15 viki Well, can't think of anything better than to methodically revert all of the commits I pushed today in reverse order.... :/
14:16 viki Totally missed that the very last file had changes to non-fudged tests: https://github.com/perl6/roast/pull/179/files#diff-7ee4515bc24fab05803a3028a0a42ac1
14:16 viki And those are in 6.c-errata too
14:17 moritz you can squash the revert commits into one
14:17 moritz or do a git diff -R start..end | git apply
14:17 moritz or something like
14:25 * viki pokes dalek
14:25 viki error: Commit 40704b82dd6126c60e7f9a99a2ebd0aaab445003 is a merge but no -m option was given.
14:25 viki hmmm
14:26 viki now I have to learn what a parent is.
14:30 viki This sounds scary... talking about trees and never wanting things :/
14:31 moritz a merge brings two branches together, so it has two "previous" commits, aka parents
14:32 moritz if you want to revert a merge commit, you have to specify which changes to revert: the changes between the commit and its first parent, or between the commit and its second parent
14:32 viki Right, I now know how to revert it, I think, but it comes with "but it's very important you realize that doing so "declares that you will never want the tree changes brought in by the merge. As a result, later merges will only bring in tree changes introduced by commits that are not ancestors of the previously reverted merge. This may or may not be what you want"
14:32 viki And I've no idea what trees I want or not want :)
14:32 travis-ci joined #perl6-dev
14:32 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Make [;]:delete 10 faster for 2/3 dimmed arrays
14:32 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173891521 https://github.com/rakudo/rakudo/compare/e3e3fefd34ce...3b9c4c95b94e
14:32 travis-ci left #perl6-dev
14:33 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
14:34 viki Oh, I can do the git diff -R thing again... \o/
14:34 lizmat viki: is it ok if I push again, or would you like me to refrain from that while you're doing the unwinding ?
14:35 viki OK. I think I'm done.
14:35 viki and dalek is AWOL
14:35 dalek rakudo/nom: 45236f1 | (Zoffix Znet)++ | src/core/set_operators.pm:
14:35 dalek rakudo/nom: Revert merge of #911 as it breaks 6.c-errata
14:35 dalek rakudo/nom:
14:35 dalek rakudo/nom: PR: https://github.com/rakudo/rakudo/pull/911
14:35 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/45236f10b5
14:36 viki lizmat: go ahead. I'm finished.
14:36 lizmat viki: okidoki
14:41 brrt joined #perl6-dev
15:09 travis-ci joined #perl6-dev
15:09 travis-ci Rakudo build errored. Zoffix Znet 'Merge pull request #911 from ab5tract/ab5tract/sym-diff-bag-mix
15:09 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173892370 https://github.com/rakudo/rakudo/compare/3b9c4c95b94e...40704b82dd61
15:09 travis-ci left #perl6-dev
15:09 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
15:22 dalek rakudo/nom: bcf1d9a | lizmat++ | src/core/Rakudo/Internals.pm:
15:22 dalek rakudo/nom: Re-imagine R:I:NOT_ALL_(DEFINED|)TYPE
15:22 dalek rakudo/nom:
15:22 dalek rakudo/nom: - rewrite using nqp ops
15:22 dalek rakudo/nom: - don't bother coercing to bool as this is an internal method anyway
15:22 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/bcf1d9a6bb
15:22 dalek rakudo/nom: 93ca094 | lizmat++ | src/core/ (4 files):
15:22 dalek rakudo/nom: Refactor R:I:NOT_ALL_(DEFINED)_TYPE removing NOT
15:22 dalek rakudo/nom:
15:22 dalek rakudo/nom: Turned out it would be very useful if these internal methods worked
15:22 dalek rakudo/nom: the other way around and return number of elems if all of the given
15:22 dalek rakudo/nom: type.
15:32 dalek rakudo/nom: fc7c480 | lizmat++ | src/core/multidim_slice.pm:
15:32 dalek rakudo/nom: Use R:I:ALL_DEFINED_TYPE in [;] handling
15:32 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/fc7c480df6
15:39 FROGGS joined #perl6-dev
15:39 FROGGS o/
15:40 lizmat FROGGS  o\
15:42 dalek rakudo/nom: 29a3faa | lizmat++ | src/core/multidim_slice.pm:
15:42 dalek rakudo/nom: Revert "Use R:I:ALL_DEFINED_TYPE in [;] handling"
15:42 dalek rakudo/nom:
15:42 dalek rakudo/nom: This reverts commit fc7c480df6c952c0a7eedfef1aa729e4863ad775.
15:42 dalek rakudo/nom:
15:42 dalek rakudo/nom: I should have checked the performance effects on this before pushing.
15:42 dalek rakudo/nom: Making it 4x slower is not an option for such hot code paths   :-(
15:42 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/29a3faa8df
15:45 hoelzro joined #perl6-dev
15:51 zostay joined #perl6-dev
15:54 ugexe joined #perl6-dev
15:55 travis-ci joined #perl6-dev
15:55 travis-ci Rakudo build errored. Elizabeth Mattijsen 'Make [;] := foo 10 faster for 2/3 dimmed arrays'
15:55 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173899058 https://github.com/rakudo/rakudo/compare/40704b82dd61...eaf4132f4682
15:55 travis-ci left #perl6-dev
15:55 geekosaur joined #perl6-dev
15:55 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
16:07 CQ joined #perl6-dev
16:12 dalek rakudo/nom: 764cfcd | lizmat++ | src/core/multidim_slice.pm:
16:12 dalek rakudo/nom: Implement :kv,:p,:k,:v on 1 element multidim [;]
16:12 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/764cfcd882
16:36 lizmat So I have some LHF/MHF for someone more versed in Grammar/Actions/World
16:37 lizmat if you do:  for ^10000 { my @a }
16:38 lizmat what in fact happens, is that at compile time, an Array object is created, and then install_lexical_symbol adds the code to clone the created object
16:38 lizmat during execution (with an nqp::clone)
16:38 lizmat what happens if you do for ^10000 { my @a[10] }, is that the object is completely created at runtime every time, instead of a cloned one
16:39 lizmat that in itself wouldn't be so bad, but for some reason, each time we do a @a[10], we get a global deopt
16:40 lizmat I think this is the one showstopper between normal and shaped arrays that we need to fix to have performance between normal and shaped arrays on par
16:40 lizmat so, any pointers / patches would be greatly appreciated
16:40 * lizmat now starts working on the P6W
16:41 travis-ci joined #perl6-dev
16:41 travis-ci Rakudo build failed. Zoffix Znet 'Revert merge of #911 as it breaks 6.c-errata
16:41 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173913947 https://github.com/rakudo/rakudo/compare/eaf4132f4682...45236f10b54a
16:41 travis-ci left #perl6-dev
16:41 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
16:43 nine lizmat: do you think it's a good idea to post a notice in the weekly about lexical_module_loading? That people should prepare for the change by making sure they add use statements for all modules they use and not rely on indirect loading?
16:44 lizmat nine: excellent point, will do!
16:52 CQ left #perl6-dev
17:12 dalek rakudo/nom: e10f767 | (Zoffix Znet)++ | src/core/set_operators.pm:
17:12 dalek rakudo/nom: Handle negative Mixy weights in ∪ operator
17:12 dalek rakudo/nom:
17:12 dalek rakudo/nom: Lookup of a non-existant key in a Mixy gives a zero, which when `max`ed against
17:12 dalek rakudo/nom: a key with a negative weight would lead to that key being removed from the final union,
17:12 dalek rakudo/nom: making the operator trim all the negative weights for keys present in just one Mixy.
17:12 dalek rakudo/nom:
17:12 dalek rakudo/nom: Add a check for whether a key exists: if it does, take a `max` value, if it doesn't,
17:12 dalek rakudo/nom: just use the value of the key from other Mixy as is.
17:12 dalek rakudo/nom:
17:12 dalek rakudo/nom: ab5tract++ for submitting code that revealed this bug.
17:12 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e10f76758b
17:22 dalek roast: c28b010 | (Zoffix Znet)++ | S02-types/mix (2 files):
17:22 dalek roast: Test negative weights survive ∪ operator
17:22 dalek roast:
17:22 dalek roast: Fixed in Rakudo in
17:22 dalek roast: https://github.com/rakudo/rakudo/commit/e10f76758b09f30b1c17f7103bd7f2ecd4f82927
17:22 dalek roast: review: https://github.com/perl6/roast/commit/c28b0108a5
17:27 travis-ci joined #perl6-dev
17:27 travis-ci Rakudo build errored. Elizabeth Mattijsen 'Refactor R:I:NOT_ALL_(DEFINED)_TYPE removing NOT
17:27 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173928761 https://github.com/rakudo/rakudo/compare/45236f10b54a...93ca094c9f76
17:27 travis-ci left #perl6-dev
17:27 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
17:31 viki .seen psch
17:31 yoleaux2 I saw psch 25 Oct 2016 19:12Z in #perl6-dev: <psch> don't we do that for closures as substituting argument?
17:31 dogbert17 lizmat++, shaped array benchmark: real 1m53.602s user 1m52.960s, This is Rakudo version 2016.10-213-ge10f767 built on MoarVM version 2016.10-37-gf769569
17:31 dogbert17 compare with: Real 4m38.345s, user 4m37.488s, This is Rakudo version 2016.10-176-gaa97f86 built on MoarVM version 2016.10-37-gf769569
17:32 dogbert17 and real 6m17.384  user 6m16.172s with Rakudo version 2016.09-105-g4abc28c built on MoarVM version 2016.09-13-g34c375a
17:32 lizmat m: say 377 / 113
17:32 camelia rakudo-moar e10f76: OUTPUT«3.336283␤»
17:33 lizmat not too shabby  :-)
17:33 dogbert17 very impressive
17:41 lizmat dogbert17: oddly enough, I'm on 2016.10-199 ?  so you're 14 commits ahead ?
17:41 dogbert17 lizmat: interesting
17:42 * viki is on 2016.10-213-ge10f767
17:42 dogbert17 oO
17:43 * lizmat pulls and rebuilds
17:44 lizmat hmmm.. I pulled a commit from viki, then rebuilt: still at 2016.10-199-gbb662ff
17:45 mst git clean dxf ?
17:45 japhb .oO( git clean --aliens-style )
17:45 * viki shrugs
17:46 mst git clean --gangnam-style
17:46 viki bb66 is 2 commits prior to my merge-revert fiasco, so at least it's not that :)
17:46 * ilmari watched Aliens Live yesterday, it was great
17:46 ilmari in Royal Albert Hall with the Royal Philharmonic Concert Orchetra playing the score live
17:47 japhb That's ... really damn cool, actually.
17:48 viki lizmat: hm. Interesting. The 213 is what I got from git describe... the actual perl6 -v is This is Rakudo version 2016.10-207-g45236f1
17:49 lizmat ah, I know what it is
17:49 lizmat git describe tells me 2016.10-213-ge10f767
17:49 viki And the reason for that is likely 'cause my build was made by running make t/spec.... instead of full thing
17:49 lizmat could very well be that -199 was the last time I did a Configure.pl
17:50 viki yeah, prolly
17:50 nine ilmari: sounds pretty epic
17:50 lizmat do we have some docs on TWEAK ?
17:50 * japhb is rather jealous ... that had to be an awesome experience (Aliens Live)
17:51 viki yeah, I now got 213 in perl6 -v after running perl Configure.pl --gen-moar --gen-nqp --backends=moar; make; make test; make install
17:51 * lizmat remembers seeing Koyaanisqatsi live
17:52 lizmat viki: yup, same here
18:05 lizmat nine: do you have a writeup somewhere I can refer people to re lexical_module_loading ?
18:16 nine lizmat: sorry, no :(
18:17 travis-ci joined #perl6-dev
18:17 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Use R:I:ALL_DEFINED_TYPE in [;] handling'
18:17 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173931704 https://github.com/rakudo/rakudo/compare/93ca094c9f76...fc7c480df6c9
18:17 travis-ci left #perl6-dev
18:18 buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
18:19 viki t/04-nativecall/20-concurrent.t
18:19 viki And reprd on 4th run: ===SORRY!=== Cannot invoke this object (REPR: P6opaque; Scalar)
18:23 AlexDaniel joined #perl6-dev
18:47 viki Slightly more info on the 20-concurrent.t crash: https://gist.github.com/zoffixznet/26cd8786d5c6133129047804b53948e0
18:56 dogbert17 viki: could it be here https://github.com/rakudo/rakudo/blob/nom/src/core/Promise.pm#L175
18:56 lizmat fwiw, that's where I've seen this fail as well, with "code" being null all of a sudden
19:00 timotimo when you get that kind of error, we have some kind of oversharing going on, or memory corruption or something like that
19:02 dogbert17 the second row in viki's gist seems to point to https://github.com/rakudo/rakudo/blob/nom/src/core/ThreadPoolScheduler.pm#L163
19:07 travis-ci joined #perl6-dev
19:07 travis-ci Rakudo build errored. Elizabeth Mattijsen 'Revert "Use R:I:ALL_DEFINED_TYPE in [;] handling"
19:07 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173935016 https://github.com/rakudo/rakudo/compare/fc7c480df6c9...29a3faa8df65
19:07 travis-ci left #perl6-dev
19:07 buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
19:45 viki dogbert17: yeah, that's the line referenced by exception backtrace.
19:45 viki It's all above my head :'(
19:46 viki Well, I can follow jnthn++'s advice: golf it down
19:47 masak always golf it down
19:47 viki True.
19:47 arnsholt Golf it down should be the first step in almost any debugging routine, IMO
20:06 masak it's just Science: http://strangelyconsistent.org/blog/im-doing-science-and-im-still-alive
20:08 viki Interestingly, that test just hangs on my 32-bit 2-core box.
20:10 travis-ci joined #perl6-dev
20:10 travis-ci Rakudo build failed. Elizabeth Mattijsen 'Implement :kv,:p,:k,:v on 1 element multidim [;]'
20:10 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173944530 https://github.com/rakudo/rakudo/compare/29a3faa8df65...764cfcd88248
20:10 travis-ci left #perl6-dev
20:10 buggable [travis build above] ☠ Did not recognize some failures. Check results manually.
20:11 viki OK. Well, I'm gonna give masak's broken window theory a go.
20:18 masak in the long run, I'd like for us to give it a structured consistent go in the form of some more or less formalized process :)
20:22 dalek rakudo/nom: c68e85b | (Zoffix Znet)++ | t/04-nativecall/20-concurrent.t:
20:22 dalek rakudo/nom: Temporarily skip flopping test
20:22 dalek rakudo/nom:
20:22 dalek rakudo/nom: To quite down Travis IRC bot. Ticket opened as RT#130042:
20:22 dalek rakudo/nom: https://rt.perl.org/Ticket/Display.html?id=130042
20:22 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130042
20:22 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c68e85b4fa
20:23 dalek rakudo/nom: 62005f1 | (Zoffix Znet)++ | .travis.yml:
20:23 dalek rakudo/nom: Quiet JVM Travis Failures
20:23 dalek rakudo/nom:
20:23 dalek rakudo/nom: Frequent failures with java.lang.OutOfMemoryError. Opened ticket as RT#130043:
20:23 dalek rakudo/nom: https://rt.perl.org/Ticket/Display.html?id=130043
20:23 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=130043
20:23 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/62005f13b5
20:25 viki .tell psch I disabled JVM reporting again on Travis. It frequently gives OutOfMemoryError; dunno if it's just Travis? RT: https://rt.perl.org/Ticket/Display.html?id=130043
20:25 yoleaux2 viki: I'll pass your message to psch.
20:28 viki masak: yeah, maybe have buggable detect frequent failures and cry about it
20:28 viki .oO( hm, but then we're just making TWO bots assault us )
20:49 dalek rakudo/nom: bf51eca | Xliff++ | src/core/DateTime.pm:
20:49 dalek rakudo/nom: Suggested shortcut method similar to the Date methods in Datish.pm
20:49 dalek rakudo/nom:
20:49 dalek rakudo/nom: If yyyy-dd-mm is a thing, then this at least adds consistency. ;)
20:49 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/bf51ecaa49
20:49 dalek rakudo/nom: 25ef3d9 | (Zoffix Znet)++ | src/core/DateTime.pm:
20:49 dalek rakudo/nom: Merge pull request #913 from Xliff/patch-3
20:49 dalek rakudo/nom:
20:49 dalek rakudo/nom: Suggested shortcut method similar to the Date methods in Datish.pm
20:49 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/25ef3d9f85
20:52 masak viki: maybe the same bot that reports the failure could also have some heuristic about tests flapping?
20:52 masak I dunno, just an idea
20:52 viki Yeah, probably.
21:05 lizmat and another Perl 6 Weekly hits the Net: https://p6weekly.wordpress.com/2016/11/07/2016-45-shapes-get-a-boost/
21:10 lizmat fwiw, I cannot build JVM backend on OS X anymore either
21:10 lizmat Stage classfile  : java.lang.OutOfMemoryError: PermGen space
21:10 lizmat in classfile (gen/jvm/stage2/NQPHLL.nqp:93)
21:12 travis-ci joined #perl6-dev
21:12 travis-ci Rakudo build errored. Zoffix Znet 'Handle negative Mixy weights in ∪ operator
21:12 travis-ci https://travis-ci.org/rakudo/rakudo/builds/173962369 https://github.com/rakudo/rakudo/compare/764cfcd88248...e10f76758b09
21:12 travis-ci left #perl6-dev
21:12 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
21:15 timotimo ctrl-f TWEAK, no result :(
21:15 timotimo was it actually merged?
21:19 lizmat timotimo: no. afaik
21:20 timotimo ah, that explains it
21:26 jnthn lizmat: "He also fixed an issue with Lock.protect that would not release the lock if an exception occurred in the code being protected." - that'll confuse most people in so far as what they think of as exceptions worked totally fine :)
21:26 jnthn lizmat: It's control exceptions that were the issue
21:27 lizmat ah, like "next" and "last" and "return" ?
21:27 lizmat jnthn: ?
21:27 jnthn Yes, return was the particular case where I stubbed my toe on it :)
21:29 jnthn lizmat++ for the weekly :)
21:31 lizmat updated
21:31 lizmat m: say await start die
21:31 camelia rakudo-moar 25ef3d: OUTPUT«===SORRY!===␤Died␤»
21:32 lizmat hmmm...
21:32 lizmat m: say await start { die }
21:32 camelia rakudo-moar 25ef3d: OUTPUT«===SORRY!===␤Died␤»
21:34 lizmat m: say await start { LEAVE say "bye"; 42 }
21:34 camelia rakudo-moar 25ef3d: OUTPUT«bye␤42␤»
21:34 lizmat m: say await start { 42; LEAVE say "bye" }  # this is wrong, I think
21:34 camelia rakudo-moar 25ef3d: OUTPUT«WARNINGS for <tmp>:␤Useless use of constant integer 42 in sink context (line 1)␤bye␤Nil␤»
21:34 lizmat both for the warning, and the value being returned.
21:36 jnthn m: try await start { die }; say $!.WHAT
21:36 camelia rakudo-moar 25ef3d: OUTPUT«(AdHoc+{Broken})␤»
21:36 jnthn Huh, it's getting that bit right
21:37 jnthn m: try await start { die }; say $!
21:37 camelia rakudo-moar 25ef3d: OUTPUT«Tried to get the result of a broken Promise␤  in block <unit> at <tmp> line 1␤␤Original exception:␤    Died␤      in block  at <tmp> line 1␤␤»
21:37 jnthn And that
21:37 jnthn So why on earth does it think it's a compile-time error?
21:38 geekosaur it's worse than that
21:38 lizmat m: say await start { die }; INIT say "foo"   # the error is in the messaging
21:38 camelia rakudo-moar 25ef3d: OUTPUT«foo␤===SORRY!===␤Died␤»
21:38 geekosaur m: say 'hi'; say await start die;
21:38 camelia rakudo-moar 25ef3d: OUTPUT«hi␤===SORRY!===␤Died␤»
21:40 geekosaur so either it's using a SORRY preamble on a runtime message, or it's EVAL-ing something
21:42 jnthn Oh, I'm pretty sure it's in the "is this a compile time error" detection
21:42 jnthn afk for a bit
21:42 lizmat yeah, feels like that
22:14 travis-ci joined #perl6-dev
22:14 travis-ci Rakudo build failed. Zoffix Znet 'Temporarily skip flopping test
22:14 travis-ci https://travis-ci.org/rakudo/rakudo/builds/174014267 https://github.com/rakudo/rakudo/compare/e10f76758b09...c68e85b4faf0
22:14 travis-ci left #perl6-dev
22:14 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
22:26 lizmat m: my @a = ^10; say @a.EXISTS-POS(-1)
22:26 camelia rakudo-moar 25ef3d: OUTPUT«False␤»
22:27 lizmat m: my @a[10] = ^10; say @a.EXISTS-POS(-1)  # so is this the correct behaviour, or the previous one?
22:27 camelia rakudo-moar 25ef3d: OUTPUT«Index -1 for dimension 1 out of range (must be 0..9)␤  in block <unit> at <tmp> line 1␤␤»
22:28 lizmat ah, I see negative subscripts are caught by postcircumfix:<[ ]> code
22:29 dogbert17 lizmat: " In the lexical_module_loading branch, Stefan Seifert *as* fixed this."
22:29 lizmat dogbert17++   removed  :-)
22:31 dogbert17 must be cached somehow, looks the same to me :)
22:35 lizmat forced reload fixed it for me
22:35 dogbert17 yeah it's gone now. lizmat++ Weekly
22:43 lizmat m: my str @a[3]; dd @a   # jnthn: could it be we forgot native shaped str arrays ?
22:43 camelia rakudo-moar 25ef3d: OUTPUT«MultiDimArray: atpos expected string register␤  in block <unit> at <tmp> line 1␤␤»
22:46 lizmat I guess there's no VM support for it...
22:59 dalek rakudo/nom: c000614 | lizmat++ | src/core/native_array.pm:
22:59 dalek rakudo/nom: Mark native shaped str arrays as NYI
22:59 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/c00061432d
22:59 lizmat and with I wish #perl6-dev a good night!
23:00 lizmat *that
23:23 travis-ci joined #perl6-dev
23:23 travis-ci Rakudo build errored. Zoffix Znet 'Quiet JVM Travis Failures
23:23 travis-ci https://travis-ci.org/rakudo/rakudo/builds/174014787 https://github.com/rakudo/rakudo/compare/c68e85b4faf0...62005f13b535
23:23 travis-ci left #perl6-dev
23:23 buggable [travis build above] ☠ Did not recognize some failures. Check results manually. All failures are on JVM only.
23:37 seatek joined #perl6-dev
23:38 AlexDaniel hm, I've seen some run-time SORRY! errors before
23:38 AlexDaniel not sure if there are any open tickets for that
23:39 AlexDaniel m: say 42; *...‘WAT’; say 24
23:39 camelia rakudo-moar c00061: OUTPUT«42␤No such method 'succ' for invocant of type 'Whatever'␤  in block <unit> at <tmp> line 1␤␤»
23:39 AlexDaniel that's new
23:39 AlexDaniel m: say 0, * ... "what"
23:39 camelia rakudo-moar c00061: OUTPUT«No such method 'pred' for invocant of type 'Whatever'␤  in block <unit> at <tmp> line 1␤␤»
23:40 AlexDaniel #128161
23:40 synopsebot6 Link:  https://rt.perl.org/rt3//Public/Bug/Display.html?id=128161
23:40 AlexDaniel bisect: *...‘WAT’
23:40 bisectable6 AlexDaniel, Bisecting by output (old=2015.12 new=c000614) because on both starting points the exit code is 1
23:40 bisectable6 AlexDaniel, bisect log: https://gist.github.com/816d9f3ea80621a67988b98997e7ecc6
23:40 bisectable6 AlexDaniel, (2016-08-19) https://github.com/rakudo/rakudo/commit/25e9fd76e85fabda20e263b6f87e27b0673f26e2
23:41 AlexDaniel commit: 25e9fd76e85f^,25e9fd76e85f say 0, * ... "what"
23:41 committable6 AlexDaniel, ¦«25e9fd76e85f^»: ===SORRY!===␤Method 'pred' not found for invocant of class 'Whatever' «exit code = 1»␤¦«25e9fd7»: Method 'pred' not found for invocant of class 'Whatever'␤  in block <unit> at /tmp/UJmaNsyS3r line 1␤ «exit code = 1»
23:42 AlexDaniel yea, so it's no longer sorry after 25e9fd7
23:47 geekosaur joined #perl6-dev

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