00:18 benjif joined 01:09 reportable6 left, reportable6 joined 01:32 benjif left 02:47 ggoebel left 03:24 benjif joined 04:25 benjif left 04:50 Kaiepi left 04:57 Kaiepi joined 06:20 [TuxCM] joined 06:25 [TuxCM] left 06:30 benjif joined, benjif left 06:31 benjif joined 06:59 benjif left 07:13 squashable6 left 07:16 squashable6 joined 07:59 [TuxCM] joined 08:19 [TuxCM] left 08:32 [TuxCM] joined
nine Ooooh...it's the optimizer! The bug appears at optimization level 2 08:34
tellable6 2019-10-19T22:35:11Z #raku-dev <patrickb> nine wrt the Inline::Perl5 with `make test`. I think I have an idea what's going wrong and what needs to be done to fix it. I'll report back in the next days.
nine And suddenly the inexplicable behaviour makes sense. All the subtle changes in the code prevent static inlining of the original function body. 08:40
Also interesting: the argument to the native trait leads to an additional block getting added to the code. This inner block then prevents inlining. 08:50
We also only inline if all arguments are native types (that's why using Pointer fixes things) 08:51
09:01 sena_kun joined
Geth rakudo: d662912e21 | (Stefan Seifert)++ | 2 files
Fix optimizer interfering with some native subs

When a native sub has only native arguments and is declared as "is native" without args (i.e. already loaded library like libc) and has a very simple function body like {} or { * } but not { ... } or other stub markers, the optimizer could statically inline the original function body, circumventing the call to the native function.
Mark all native subs as soft, to prevent static inlining. Fixes GH #3244		 09:22
rakudo: c1ccff8224 | (Stefan Seifert)++ | src/Perl6/Metamodel/Naming.nqp
Revert "Fix another MOP-related StrLexRef serialization bug"

This reverts commit 9b181a700e51c26802dce8b48ef2d234c303eb66 as that change breaks Inline::Perl5		 09:26
nine Kaiepi: ^^^
releasable6: status 09:33
releasable6 nine, Next release will happen when it's ready. 5 blockers. Unknown changelog format
nine, Details: gist.github.com/8a60c412bc444895d2...1b90d9a331
09:54 [Coke]_ joined, [Coke]_ left, [Coke]_ joined 09:55 pmurias joined, [Coke] left 09:56 ilogger2 left 09:58 ilogger2 joined 10:09 tony-o joined
nine releasable6: status 10:58
releasable6 nine, Next release will happen when it's ready. 4 blockers. Unknown changelog format
nine, Details: gist.github.com/801cdfb7a7e8941feb...2cd105334a
nine releasable6: status 11:15
releasable6 nine, Next release will happen when it's ready. 3 blockers. Unknown changelog format
nine, Details: gist.github.com/ebba78dc916d3ed69e...19fd9e4e23
11:18 pmurias joined 11:30 pmurias left 11:41 pmurias joined
Geth rakudo: 2dd570dee4 | (Stefan Seifert)++ | src/core.c/CompUnit/PrecompilationRepository.pm6
Fix leaking an open file handle when re-precompiling

When checking if another process has precompiled the exact module that we need precompiled while we were waiting for the lock on the precomp repository, we opened the unit but never closed it. This would trip up Windows which refuses to rename over an open file. Fix by ensuring that we close the unit when we're done with our checks.
Hopefully fixes GH #3170		 11:46
nine I'm pretty sure this takes care of another release blocker. Just needs confirmation by a test on Windows. 11:48
[TuxCM] Rakudo version 2019.07.1-422-gc1ccff822 - MoarVM version 2019.07.1-284-gc61236f96
csv-ip5xs0.749 - 0.753
csv-ip5xs-206.329 - 6.640
csv-parser20.919 - 21.135
csv-test-xs-200.422 - 0.425
test6.922 - 7.149
test-t1.740 - 1.751
test-t --race0.788 - 0.791
test-t-2029.968 - 30.156
test-t-20 --race8.776 - 9.177
nine That just leaves the build system issue and that weird telemetry in MoarVM related thing, for both of which there are people more qualified than me. Guess I've earned a visit to the sauna now :) 11:49
lizmat nine++ 11:50
12:44 lucasb joined 12:46 pmurias left
Geth roast: samcv++ created pull request #589:
Add samemark tests for prepend and flags		 13:40
rakudo: samcv++ created pull request #3251:
.samemark: Support emoji, flags and Prepend
14:13 ZzZombo_ joined, ZzZombo_ is now known as ZzZombo 14:25 Kaiepi joined 14:37 Kaiepi left 14:38 Kaiepi joined, softmoth joined 14:39 Kaiepi left, Kaiepi joined 14:40 ggoebel joined 14:52 Altai-man_ joined
Kaiepi argh, my commit broke Inline::Perl5? 14:55
i'm guessing the actual fix for this is something else because i'm pretty sure this used to behave differently: 14:56
bisectable6, BEGIN { my str $name = 'foo'; our Mu:U $subset := Metamodel::SubsetHOW.new_type: :$name; $*W.add_object_if_no_sc: $subset }; say $subset.^name.WHAT 14:58
bisectable6 Kaiepi, On both starting points (old=2015.12 new=2dd570d) the exit code is 1 and the output is identical as well
Kaiepi, gist.github.com/4424d7f1796b9ec357...aa9a4128b5
Kaiepi bisectable6, BEGIN { my str $name = 'foo'; our $subset := Metamodel::SubsetHOW.new_type: :$name; $*W.add_object_if_no_sc: $subset }; say $subset.^name.WHAT
bisectable6 Kaiepi, On both starting points (old=2015.12 new=2dd570d) the exit code is 1 and the output is identical as well
Kaiepi, gist.github.com/6c6def919c696ea967...531a6f4fb0
Kaiepi bisectable6, BEGIN { my str $name = 'foo'; our $subset := Metamodel::SubsetHOW.new_type: :$name, :refinee(Any), :refinement(Str | Int); $*W.add_object_if_no_sc: $subset }; say $subset.^name.WHAT
bisectable6 Kaiepi, On both starting points (old=2015.12 new=2dd570d) the exit code is 1 and the output is identical as well
Kaiepi, Output on both points: «04===SORRY!04=== Error while compiling /tmp/qTPKARmtaO␤Variable '$subset' is not declared␤at /tmp/qTPKARmtaO:1␤------> 03$*W.add_object_if_no_sc: $subset }; say 08⏏04$subset.^name.WHAT␤»
Kaiepi bisectable6, my Mu:U $subset; BEGIN { my str $name = 'foo'; $subset := Metamodel::SubsetHOW.new_type: :$name, :refinee(Any), :refinement(Str | Int); $*W.add_object_if_no_sc: $subset }; say $subset.^name.WHAT 14:59
bisectable6 Kaiepi, On both starting points (old=2015.12 new=2dd570d) the exit code is 0 and the output is identical as well
Kaiepi, Output on both points: «(Str)␤»
Kaiepi hm 15:00
when does StrLexRef pop up? 15:01
in general, i mean 15:12
lizmat when you pass a native str to a sub that accepts the parameter raw ? 15:18
aka: my str $a = "foo"; sub a(\b) { }; a $a 15:19
?
AlexDaniel nine: maybe you can help them: github.com/frithnanth/perl6-Math-F...-544262136 15:23
Kaiepi that turned out to be the issue lizmat 15:30
i was declaring $name as a str rather than a Str:D because i thought i'd need to since the MOP is nqp 15:32
is not being able to use native strings with the MOP in raku modules worth making an issue over? 15:38
timotimo m: sub takes(\a) { say a.WHAT }; my str $a = "hi"; takes($a) 15:40
camelia (Str)
timotimo m: sub takes(\a) { say a.VAR.WHAT }; my str $a = "hi"; takes($a)
camelia (Str)
timotimo m: sub takes(\a is rw) { say a.VAR.WHAT }; my str $a = "hi"; takes($a)
camelia 5===SORRY!5=== Error while compiling <tmp>
Can only use 'is rw' on a scalar ('$' sigil) parameter, not 'a'
at <tmp>:1
timotimo m: sub takes($a is rw) { say $a.VAR.WHAT }; my str $a = "hi"; takes($a) 15:41
camelia (StrLexRef)
timotimo m: sub takes($a is rw) { say $a.VAR.WHAT }; my str $a = "hi"; takes($a<>)
camelia Parameter '$a' expected a writable container, but got Str value
in sub takes at <tmp> line 1
in block <unit> at <tmp> line 1
timotimo m: sub takes($a is raw) { say $a.VAR.WHAT }; my str $a = "hi"; takes($a<>)
camelia (Str)
timotimo m: sub takes($a is raw) { say $a.VAR.WHAT }; my str $a = "hi"; takes($a)
camelia (Str)
timotimo just raw doesn't seem to give you a ref, at least not in perl6 code
Kaiepi the issue i was having only happened during precomp
so you'd need to test with a module
15:43 pmurias joined 15:46 Xliff joined
Xliff \o 15:46
Kaiepi o/
pmurias \o 16:06
Geth nqp/truffle: 8fad04b0c4 | (Paweł Murias)++ | 39 files
[truffle] Use VMArray instances instead of NQPList

To do that we implement nqp::sethllconfig (and some other things) so the type objects we need get installed appropiately.		 16:07
nqp/truffle: 097e7436d3 | (Paweł Murias)++ | 2 files
[truffle] Set hll role on BOOTArray
rakudo/ugexe-patch-2: 91c7b28cb3 | (Nick Logan)++ (committed using GitHub Web editor) | t/04-nativecall/00-misc.t
FIx test count failure on windows
 16:18
rakudo: ugexe++ created pull request #3252:
FIx test count failure on windows
16:38 softmoth left 16:49 Altai-man_ left 17:08 softmoth joined
Geth rakudo: 91c7b28cb3 | (Nick Logan)++ (committed using GitHub Web editor) | t/04-nativecall/00-misc.t
FIx test count failure on windows
 17:46
rakudo: bfaec97417 | niner++ (committed using GitHub Web editor) | t/04-nativecall/00-misc.t
Merge pull request #3252 from rakudo/ugexe-patch-2

Fix test count failure on windows
19:08 Kaiepi left 19:11 Kaiepi joined 19:21 pmurias left 19:23 pmurias joined
Geth nqp: 511266a0cb | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/templates/MOAR_REVISION
[MoarVM Bump] Brings 2 commits

MoarVM bump brought: github.com/MoarVM/MoarVM/compare/2...g0f0e6d3a2 0f0e6d3a2 Work around MSVC not coping well with preprocessor instructions inside an MVMROOT c61236f96 Fix access to freed memory in resolve_using_guards		 19:48
rakudo: bf318db684 | (Aleks-Daniel Jakimenko-Aleksejev)++ | tools/templates/NQP_REVISION
[NQP Bump] 511266a0c [MoarVM Bump] Brings 2 co […]

NQP bump brought: github.com/perl6/nqp/compare/2019....g511266a0c
20:00 Xliff left 20:49 Kaiepi left 20:50 Kaiepi joined
pmurias jnthn: do you recommend just copying over the IO abstraction to nqp-truffle? 20:59
StandardWriteHandle, IIOClosable etc. 21:00
21:03 sena_kun joined 21:13 tellable6 joined
Geth roast/elems-on-seq-cache: 9f07cb4b57 | Altai-man++ | S32-list/seq.t
Test caching by elems call on Seq

Closes github.com/rakudo/rakudo/issues/2976		 21:15
21:16 ggoebel left
Geth roast: Altai-man++ created pull request #590:
Test caching by elems call on Seq		 21:16
21:23 sena_kun left
Geth nqp/truffle: 4 commits pushed by (Paweł Murias)++ 22:21
22:53 pmurias left 22:54 pmurias joined, lucasb left
Geth rakudo: tbrowder++ created pull request #3253:
Improve error msgs: invalid open bracket, declarator blocks		 23:05
rakudo/ugexe-patch-3: feb6d43d3a | (Nick Logan)++ (committed using GitHub Web editor) | appveyor.yml
Add debug output for appveyor
 23:11
23:56 pmurias left 23:59 Kaiepi left, Kaiepi 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!