00:42 sena_kun left 00:56 sena_kun joined 02:42 sena_kun left 02:56 sena_kun joined 04:42 Kaiepi left, sena_kun left 04:57 sena_kun joined 05:01 Kaiepi joined 05:14 Kaiepi left 05:22 Kaiepi joined 06:41 sena_kun left 06:56 sena_kun joined 07:34 domidumont joined 07:42 domidumont left 08:21 go|dfish left 08:23 go|dfish joined 08:41 sena_kun left 08:55 sena_kun joined 10:42 sena_kun left 10:57 sena_kun joined
lizmat Files=1297, Tests=109772, 216 wallclock secs (28.02 usr 8.15 sys + 2962.37 cusr 275.04 csys = 3273.58 CPU) 11:09
that's 6 wallclock and 30 CPU seconds more than yesterday 11:10
wonder if 4366980681015624bb2c302b0 is responsible for that
sena_kun bumps 11:12
bisectable6: enum Foo <bar baz>; say bar ~~ Foo; 11:31
bisectable6 sena_kun, On both starting points (old=2015.12 new=5d6c0b6) the exit code is 0 and the output is identical as well
sena_kun, Output on both points: «True␤»
sena_kun bisectable6: start=2019.11 enum Foo <bar baz>; say bar ~~ Foo;
bisectable6 sena_kun, Bisecting by output (old=2015.12 new=5d6c0b6) because on both starting points the exit code is 1
sena_kun, bisect log: gist.github.com/46de5bbdd8f4b9ee9e...15f28d296f 11:32
sena_kun, (2017-10-07) github.com/rakudo/rakudo/commit/2d...5f79466491
sena_kun sigh
Geth nqp: 7f7efd65dd | Altai-man++ | tools/templates/MOAR_REVISION
[MoarVM Bump] Brings 14 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...g1fa9764bf 1fa9764bf Merge pull request #1224 from MasterDuke17/fix_compiler_warnings 24159f759 Merge pull request #1227 from ZhongnianTao/master 61ac671ea Update gb18030.c e46fd6f3e Merge pull request #1 from MoarVM/master ... (10 more lines)
11:35
rakudo: 06db4c2499 | Altai-man++ | tools/templates/NQP_REVISION
[NQP Bump] 7f7efd65d [MoarVM Bump] Brings 14 c […]

NQP bump brought: github.com/perl6/nqp/compare/2019....g7f7efd65d
MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...g1fa9764bf 1fa9764bf Merge pull request #1224 from MasterDuke17/fix_compiler_warnings 24159f759 Merge pull request #1227 from ZhongnianTao/master ... (12 more lines)
lizmat bisectable6: my %h; %h<foo><bar> := my $b; dd %h 12:36
bisectable6 lizmat, On both starting points (old=2015.12 new=06db4c2) the exit code is 0 and the output is identical as well
lizmat, Output on both points: «Hash %h = {:foo(${:bar(Any)})}␤»
12:42 sena_kun left 12:57 sena_kun joined
Geth roast: 2510c04114 | (Elizabeth Mattijsen)++ | 13 files
Batch #21 of roast RT -> GH ticket migration
13:00
lizmat only 439 files to go :-) 13:01
13:03 pmurias joined 13:10 lucasb joined
lizmat AlexDaniel: why did github.com/Raku/whateverable/issues/372 get closed ? 14:40
ah, it's not about Proc::Feed
AlexDaniel yeah, cuz I fixed it :) 14:41
lizmat so Proc::Feed is just broken? because it breaks for me reliably
14:41 sena_kun left
AlexDaniel lizmat: I don't know, in this case it was trying to kill it because of the timeout 14:45
14:56 sena_kun joined
Geth rakudo: 59f6a031e8 | (Elizabeth Mattijsen)++ | src/core.c/Supply-factories.pm6
Implement Supply.decode($encoding)
15:02
roast: 28bf1fd09c | (Elizabeth Mattijsen)++ | 3 files
Add and run tests for Supply.decode

Also some cleanup on the Supply.encode tests
15:03
15:25 pmurias left 15:26 pmurias joined
Geth rakudo: ace9853146 | (Elizabeth Mattijsen)++ | src/core.c/Supply-factories.pm6
Implement Supply.first

As a frontend of Supply.grep.
15:55
roast: b7e2c2ca95 | (Elizabeth Mattijsen)++ | 2 files
Add and run Supply.first tests
16:19 pmurias left
Geth roast: ecd28b5478 | (Elizabeth Mattijsen)++ | 2 files
Add and run Supply.collate tests
16:31
rakudo: 4079702846 | (Elizabeth Mattijsen)++ | 2 files
Implement Supply.collate

Also make it a multi
16:40 pmurias joined 16:41 sena_kun left 16:56 sena_kun joined
MasterDuke lizmat: any thoughts about the expmod tests? 17:46
lizmat MasterDuke: please remind me which ones :-) 17:47
MasterDuke github.com/perl6/roast/blob/master....t#L25-L37 and github.com/perl6/nqp/blob/master/t...#L206-L207 17:48
github.com/Raku/problem-solving/issues/150 17:49
lizmat the first one is a JVM backend issue, no?
so you're saying the default value should be something like NaN or Nil ? 17:50
MasterDuke no, it's that those values now cause an error 17:53
in the libtommath function
they used to put the libtommath function into an infinite loop, but now it reports "value out of range". but since we don't check for errors, we just give a/the default value as the result 17:55
17:56 Guest78373 left
MasterDuke pmurias: in github.com/perl6/nqp/commit/cfddd0...29ed1b24f, can i assume you don't care what the actual values are, just testing that some negative values for exponents work? 18:00
lizmat MasterDuke: hmmmm 18:07
so returning NaN in such a case would be appropriate ?
MasterDuke could be 18:08
in python, `pow(4, -4, 2)` gives `ValueError: base is not invertible for the given modulus` 18:09
lizmat hmmm... since it did put it into an infinite loop before, there is no compatibility issue if we would make that an execution error, right ? 18:13
die "base is not invertible for the given modulus" # feels like a good one solution 18:14
hopefully with the offending values being part of the error message
MasterDuke akf for dinner, back later 18:18
Geth rakudo: 7925bfd485 | (Elizabeth Mattijsen)++ | src/core.c/Supply-coercers.pm6
Make Supply.squish a multi

With some optimizations, specifically for the default &with
18:25
roast: dcb3af9508 | (Elizabeth Mattijsen)++ | S17-supply/squish.t
You *can* run .squish on a Supply type object
rakudo: ff08697ace | (Elizabeth Mattijsen)++ | src/core.c/Supply-coercers.pm6
Implement Supply.repeated
18:40
roast: 9bf0ab0d31 | (Elizabeth Mattijsen)++ | 2 files
Add and run tests for Supply.repeated
18:41 Guest78373 joined, sena_kun left 18:58 sena_kun joined
MasterDuke lizmat: nqp and moarvm just pass values to libtommath function and all checks at the raku level? so typed exceptions can be thrown? 19:19
lizmat I'll settle for an X::AdHoc, which I think you can throw from MoarVM 19:20
MasterDuke k, i'll see what i can add to my PR. but we still need to change the nqp and roast tests 19:21
afk for a bit again 19:22
pmurias MasterDuke: I'm not sure if I undestand your question 19:30
MasterDuke: I don't remember 42 being of some particular importance 19:31
MasterDuke: but what do you want to return other than 0? 19:32
m: say 7.expmod(-2, 5); 19:35
camelia 4
pmurias ^^ why is it 4?
lizmat -2 + 5 = 3
pmurias m: say ((7 ** -2) mod 5)
camelia Cannot resolve caller infix:<div>(Rat:D, Int:D); none of these signatures match:
(Int:D \a, Int:D \b --> Int:D)
(int $a, int $b --> int)
in block <unit> at <tmp> line 1
pmurias lizmat: I don't understand? 19:37
lizmat no, *I* didn't understand, I have no idea :-)
pmurias ahh, so we both don't understand, seem like we have a consensus ;) 19:38
the explanation why it's 4 in the docs is bogus: docs.perl6.org/routine/expmod 19:39
Geth rakudo: 13c5d04622 | (Nick Logan)++ (committed using GitHub Web editor) | src/core.c/CompUnit/RepositoryRegistry.pm6
Fix repository name lookup on windows

When setting up the repositories it will key a hash based (mostly) on normalized absolute file paths such as `inst#/home/perl6/share/site`. It appears RAKUDO_PREFIX or RAKUDO_HOME dont normalize their path separator, and thus `$site` was left with a value of e.g. `inst#/home/perl6\share/site` (which would not match the hash lookup due to the backslash). This would break the following construct inside a module during installation on windows:
  `BEGIN my $v = $?DISTRIBUTION.meta<version>; use MyDist::Bar:ver($v)`
This normalizes `$prefix` by replacing back slashes with forward slashes on windows such that the hash keys match the format used during lookup by e.g. `$site`.
20:12
20:14 pmurias left 20:17 pmurias joined 20:42 sena_kun left 20:56 sena_kun joined
MasterDuke pmurias: python gives 4 for pow(7, -2, 5) 21:10
pmurias, lizmat: sounds like i can change the nqp test values no problem right away, and then change the roast tests after some minor tweaking to my pr? 21:13
lizmat MasterDuke: sounds like a plan 21:14
Geth nqp: c7f4dfe9e1 | (Daniel Green)++ | t/nqp/060-bigint.t
Use better negatives values for testing expmod

The original ones (randomly chosen) happen to throw an exception in MoarVM when we add error checking of the underlying libtommath function return value.
21:22
MasterDuke hm, i can't restart the builds in travis-ci.org/MoarVM/MoarVM/builds...ification, can anybody else? 21:28
21:38 travis-ci joined
travis-ci NQP build failed. Daniel Green 'Use better negatives values for testing expmod 21:38
travis-ci.org/perl6/nqp/builds/638937157 github.com/perl6/nqp/compare/7f7ef...f4dfe9e13d
21:38 travis-ci left
MasterDuke ha. jvm error. let me take a look... 21:39
21:41 pmurias left
Geth ¦ problem-solving: vrurg assigned to jnthn Issue Visibility of dynamic symbols in async blocks github.com/Raku/problem-solving/issues/151 21:57
22:16 pmurias joined 22:42 sena_kun left 22:55 sena_kun joined
Geth rakudo: 8291337799 | (Elizabeth Mattijsen)++ | 2 files
Add BUILDPLAN helper library

This library allows one to quickly see the BUILDPLAN of a class or a role. It is intended to be used for debugging work on the object creation logic of Rakudo. Some example output:
   use BUILDPLAN;
... (12 more lines)
22:56
lizmat ok, that should be helpful implementing `has $.a is built(:bind)` 23:03
23:19 lucasb left
pmurias MasterDuke: why is expmod(7, -2, 5) supposed to be 4? 23:51
MasterDuke: ahh, it seems that python docs have an explanation: docs.python.org/3/library/functions.html#pow 23:53
MasterDuke: Raku docs are wrong for negative values tho 23:54
docs.perl6.org/routine/expmod