Perl 6 language and compiler development | Logs at colabti.org/irclogger/irclogger_logs/perl6-dev | For toolchain/installation stuff see #perl6-toolchain | For MoarVM see #moarvm
Set by AlexDaniel on 12 June 2018.
MasterDuke benchable6: 2017.01,2017.06,2018.01,2018.06,HEAD my $i; while $i++ < 5_000_000 {} 00:00
benchable6 MasterDuke, starting to benchmark the 5 given commits
MasterDuke, benchmarked the given commits and found a performance difference > 10%, now trying to bisect
MasterDuke hm, it should have said the initial times it found... 00:01
timotimo was it too long? 00:02
benchable6 MasterDuke, gist.github.com/d47662062c49c2d20e...1cb7f440fb
MasterDuke oh, right, might only do that if there are less than 5 commits given initially
benchable6: 2017.01,2017.06,2018.01,2018.06,HEAD my $i; while ++$i < 5_000_000 {} 00:03
benchable6 MasterDuke, starting to benchmark the 5 given commits
MasterDuke, benchmarked the given commits and found a performance difference > 10%, now trying to bisect
MasterDuke, gist.github.com/1f5f9e4568f25276eb...35fcf4db60 00:04
MasterDuke huh, very different looking graph 00:05
Geth rakudo: jmaslak++ created pull request #2123:
Add contributor information
timotimo wow.
MasterDuke back in a bit &
timotimo i wonder if my patch would bring ++$i back down to what it used to be 00:06
same for $i++, but there it isn't as dramatic
Geth rakudo: 4ee71ece1d | (Joelle Maslak)++ | CREDITS
Add contributor information
 00:08
rakudo: 9d1e624ae3 | (Zoffix Znet)++ (committed using GitHub Web editor) | CREDITS
Merge pull request #2123 from jmaslak/joelle-contributor

Add contributor information
00:24 x[LGWs4x4i]uG2N0 left, reppie joined 00:30 lizmat left 00:33 fake_space_whale joined 02:21 reppie is now known as x[LGWs4x4i]uG2N0
MasterDuke i'm seeing a consistent fail on t/04-nativecall/02-simple-args.t when MoarVM is built with clang 02:29
not ok 11 - passed uint16 0xFFFE # Failed test 'passed uint16 0xFFFE' # at t/04-nativecall/02-simple-args.t line 72 # expected: '11' # got: '0'
02:53 benjikun left
zostay AlexDaniel: re: Hash::MultiValue testing .perl, I'd kind of like .perl to actually output the right sort of thing, how would you recommend testing that instead? 03:09
the reason the test is failing is because .perl being called on Pair is outputting `$(:x(42))` when it previously output just `:x(42)` 03:10
03:16 benjikun joined 04:09 undersightable6 joined 04:23 Zoffix joined
Zoffix zostay: A lot of core .perl isn't specced to return a particular string. The only requirement is it returns a string that you can .EVAL to the equivalent object that you .perl'ed. So the right way to test is `is-deeply $obj.perl.EVAL, $obj` 04:24
zostay: and the reason it was putting those as $(:42foo) is because those pairs are containerized and before I believe it was impossible to containerize them or something. In either case, the commit that brought that change was reverted: github.com/rakudo/rakudo/commit/b5a7439838 04:27
zostay: but you still shouldn't test exact string of .perl
.tell zostay colabti.org/irclogger/irclogger_log...-07-24#l40 04:32
yoleaux Zoffix: I'll pass your message to zostay.
04:33 Zoffix left 04:45 Kaiepi left 06:15 fake_space_whale left 06:18 robertle joined 06:25 ufobat joined
[Tux] Rakudo version 2018.06-245-g9d1e624ae - MoarVM version 2018.06-313-gebf0464e0
csv-ip5xs1.071 - 1.162
csv-ip5xs-208.330 - 8.493
csv-parser24.607 - 24.960
csv-test-xs-200.431 - 0.451
test9.902 - 10.519
test-t2.288 - 2.307
test-t --race0.969 - 1.010
test-t-2039.800 - 39.984
test-t-20 --race12.801 - 13.009
 07:02
08:35 lizmat joined
lizmat Files=1246, Tests=76415, 380 wallclock secs (16.40 usr 5.67 sys + 2676.40 cusr 255.91 csys = 2954.38 CPU) 08:36
08:51 diakopter left 08:57 a3f_ left 09:39 robertle left
Geth nqp/truffle: 987a27004b | (Paweł Murias)++ | 2 files
[truffle] Move deserialization code into the class
 09:41
09:47 camelia left, lizmat left 09:48 camelia joined 09:50 ChanServ sets mode: +v camelia 10:03 Kaiepi joined 10:04 a3f_ joined 10:05 squashable6 left, squashable6 joined 10:09 pmurias joined
Geth nqp/truffle: b64648d873 | (Paweł Murias)++ | src/vm/jvm/runtime/org/perl6/nqp/truffle/nodes/NQPToBooleanNode.java
[truffle] Use @Specialization when boolifing stuff
 10:20
10:30 squashable6 left 11:20 benjikun left 11:21 benjikun joined
Geth nqp/truffle: 0817412c16 | (Paweł Murias)++ | src/vm/jvm/runtime/org/perl6/nqp/truffle/nodes/NQPNodeWithBoolification.java
[truffle] Missing part of moving boolification to specialization
 11:24
nqp/truffle: fef920ced7 | (Paweł Murias)++ | src/vm/jvm/runtime/org/perl6/nqp/truffle/nodes/expression/NQPSmartStringifyNode.java
[truffle] Use specialization when smart stringifiying
AlexDaniel zostay: Hash::MultiValue should be fine on HEAD because lizmat++ reverted the commit: github.com/rakudo/rakudo/commit/b5...ca4345f858 11:27
yoleaux 03:58Z <zostay> AlexDaniel: Something weird is going on in Test::Path::Router... it has a `use Test` at the top, but if I insert a `dd Test::output()` inside a sub, e.g., `sub mapping-is`, I get `Could not find symbol '&output'`... that makes no sense to me
04:00Z <zostay> AlexDaniel: that test package does some unusual things to cope with the fact that there's no good Test2::API/Test::Builder/thing to build tests that cooperate with each other yet (something I'm looking into along with a bajillion other things)
AlexDaniel zostay: that being said, yes, you shouldn't be testing .perl :) 11:29
zostay: you also test .gist, and I think that's fragile too :) 11:30
Geth nqp: 3a4cbe0423 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/MOAR_REVISION
[MoarVM Bump] Brings 4 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...g4199a9564 4199a9564 Enable gc_free of MVMContext ebf0464e0 be extra careful about prev 389a3a10f Make sure output of spesh resolve gets optimized, too f4662285d Don't use 32-bit unaligned reads in SipHash implementation		 11:36
¦ nqp: version bump brought these changes: github.com/MoarVM/MoarVM/compare/2...g4199a9564
rakudo: 9e752718b9 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/build/NQP_REVISION
[NQP Bump] Brings 2 commits

NQP bump brought: github.com/perl6/nqp/compare/2018....g3a4cbe042 3a4cbe042 [MoarVM Bump] Brings 4 commits 15c977471 unfudge test, co #274 (#485)
synopsebot RAKUDO#274 [closed]: github.com/rakudo/rakudo/pull/274 Eliminate a been_tapped race.
rakudo: version bump brought these changes: github.com/perl6/nqp/compare/2018....g3a4cbe042
a71e370bb7 | (Elizabeth Mattijsen)++ | src/core/ShapedNArray.pm6

This basically moves the dim2role lookup from compile time to run time. For some reason, having this lookup at compile time, caused Array:: to be somehow special wrt serialization contexts.
This removes the error:
  ===SORRY!===
Object does not exist in serialization context
13:05 MasterDuke left 13:47 skids joined 13:48 robertle joined 13:59 diakopter joined
zostay AlexDaniel: it is fragile, but I still have to have some way of guaranteeing that .perl/.gist output something that makes sense... clearly $(:a(2)), $(:b(3)), ... doesn't make sense and if something changes in Perl such that that becomes the normal outcome of .map(*.perl) on a list of Pair, then I would want to fix that because it is a real bug. 14:18
yoleaux 04:32Z <Zoffix> zostay: colabti.org/irclogger/irclogger_log...-07-24#l40
zostay So, I suppose the test I should have is .perl.EVAL and see if the result is the same as the input. 14:20
14:37 fake_space_whale joined 15:19 samcv left 15:34 samcv joined 15:44 lizmat joined 15:56 ufobat left 15:59 brrt joined 16:12 Ven`` joined 16:25 robertle left 16:27 fake_space_whale left 16:38 samcv left 16:40 samcv joined 16:52 brrt left 17:27 Ven`` left 18:09 dogbert17 left 18:29 dct joined
lizmat hmmm... apparently you cannot make a module Array::Foo and precompile it :-( 18:31
===SORRY!=== 18:33
Object does not exist in serialization context
apparently this is some clash with Array:: 18:34
m: dd Array::
camelia {:dim2role((Array::ShapedArray, Array::Shaped1Array, Array::Shaped2Array, Array::Shaped3Array))}
19:12 ufobat joined 20:56 dct left 21:06 pmurias left 21:19 skids left 22:52 MasterDuke joined 23:14 dct joined 23:58 j3nnn1 left

Powered by Cro and the Raku® Programming Language.

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

Thank you!