🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | Logs available at irclogs.raku.org/raku-dev/live.html | For MoarVM see #moarvm
Set by lizmat on 8 June 2022.
00:17 finanalyst left 06:32 sena_kun joined 06:34 sena_kun left 07:19 [Tux] left 07:25 [Tux] joined 08:08 sena_kun joined 09:00 lizmat_ left, lizmat joined 13:09 Altai-man joined 13:13 sena_kun left
jdv how come i get the same zef error on head now? 13:39
nevermind, it seems to be working now 13:40
no, there's something actually wrong 13:47
gist.github.com/jdv/50922d3642b8be...5cd1ea1019 13:49
lizmat: ^ 13:50
ugexe: ^
release on hold i guess
and now spectests are failing. wonder if its on my side... 13:51
lizmat I could revert 4a66ddd6cf5f8f71e886 and see if that helps?
linkable6 (2024-06-24) github.com/rakudo/rakudo/commit/4a66ddd6cf Improve `.flat` logic (2nd attempt)
jdv do spectests pass for you?
lizmat checks 13:52
jdv i'll try that again
lizmat (first need to build a clean rakudo again, it had local changes) 13:53
make test and make spectest are clean for me 13:59
jdv the output from 2 spectests looks the same to me - a lot of fail 14:00
hmm
lizmat if you want me, I could nuke install and rebuild from scratch to make sure
I'll do that anyway
jdv i'll do that myself first i guess just to confirm
oh ok
lizmat rebuilding... 14:01
testing... 14:05
jdv now my build is hanging on fetching moarvm's submodules. wut? 14:08
lizmat github hickup?
fetch and build ok for me 14:09
had one flapper: t/spec/S17-promise/start.t
jdv idk, 3 times now 14:10
lizmat 2nd time ok: ok 58 - No crash when doing parallel parsing of grammars on first time
perhaps nuke ~/.raku/precomp as well 14:11
zef installs itself ok for me
ake installation fails because of deprecation messages messing up some tests that test STDERR output 14:18
installs ok with RAKUDO_NO_DEPRECATIONS=1
jdv my box is slower than yours:)
i think that's why i install from a checkout
how can we get that eco'd? 14:19
lizmat how can we get what eco'd ?
jdv released
head is not out there but the fix is in git
lizmat I've lost you: are we talking about ake ? 14:20
jdv yes. the last commit on the ake github repo has the fix for the deprecation that's breaks tests on install.
but those changes are on on zef of p6c or cpan... 14:21
*fez
not on
i can't type. new keyboard.
lizmat ack... I'll do a release of ake momentarily
jdv gotta do an errand. bb in a bit. 14:26
lizmat release of ake done, you may need to do a "zef update" to have zef know about it 14:37
jdv i'm still unable to get a clean spectest run 15:05
i guess i'll try a build without that flat commit
lizmat fwiw, are you on Intel ? 15:08
jdv amd so "yes" 15:09
Linux jdv-lap2 6.1.0-21-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.90-1 (2024-05-03) x86_64 GNU/Linux
lizmat ok, lemme try a clean install on an Intel box 15:10
this will take a bit longer :-) 15:11
jdv head seems unstable
lizmat is that with the flat commit removed? 15:14
jdv not yet. just investigating the individual failures from the whole run 15:15
rebuilding with that commit reverted now
so with that reverted make test passes first time and spectest passed mostly first time with the exception of S32-list/flat.t 15:31
before i had to run make test multiple times to get it to pass and spectest never passed and had a lot of errors 15:32
can you repro something like that? 15:33
if not maybe we should skip the release until this is understood better?
lizmat S32-lisr/flat fails because of the reverted commit 15:37
jdv yeah but unreverted gist.github.com/jdv/c237639c2bfb03...f616e1315e is a bit concerning that you don't see that 15:39
and it seems reproducable here
not sure if its the exact same failures
Geth rakudo/main: 739559906a | (Elizabeth Mattijsen)++ | 3 files
Revert "Improve `.flat` logic (2nd attempt)"

This reverts commit 4a66ddd6cf5f8f71e8868bebff3a0745666fde9b.
This appears to have ununderstood consequences
roast: 08ff778649 | (Elizabeth Mattijsen)++ | S32-list/flat.t
Revert "Revert "Revert "Add tests for .flat($level, :hammer)"""

This reverts commit 7c5440aa74e26bde4f0f1ea07bd87f12f2a36ef4.		 15:40
lizmat well, I'm seeing spectest failures on Intel without the revert as well 15:41
so... I'm thinking the .flat commit is tickling the JIT (as the M1 doesn't have any JIT) 15:42
jdv ok, so we're ok with doing the release with that stuff reverted in rakudo and roast? 15:43
lizmat I'm ok with it
jdv ah, right - forgot about that jit tidbit
lizmat now that I know bout the different behaviour on Intel, I can check it out
and come back with a better version :-) 15:44
jdv cool. i'll restart the release then.
lizmat *after* the release
jdv good luck!
lizmat jdv: running make spectest on Intel without the revert with MVM_JIT_DISABLE=1 is clean 15:56
so it *is* definitely JIT related
I'll investigate more *after* the release
jdv ah, nice 15:57
thanks
lizmat afk& 16:03
Geth rakudo/release-2024.06: 73ad19e2e1 | (Justin DeVuyst)++ | 3 files
Update changelog + announcement

Deliberately not logged:
  [33d0feaf][549842bd][f20bd324][9502bcbb][f85de735][d7425592]
  [55348374][8518cc49][2d2fe081][b3d5c5d7][9ca6046e][4a66ddd6]
  [73955990]		 17:10
jdv for some reason git submodule calls are failing for me today 17:49
the release might take forever to get through:)
[Coke] :| 18:02
18:03 finanalyst joined
jdv somehow seems like half of hte submodule updates end in timeout oh gh... but seems to be ok for the moment. 18:07
Geth nqp/main: 4a60152250 | (Justin DeVuyst)++ | tools/templates/MOAR_REVISION
[release] Bump MoarVM revision to 2024.06
 19:02
nqp/main: 7bba3a97c6 | (Justin DeVuyst)++ | VERSION
[release] Bump VERSION to 2024.06
rakudo/release-2024.06: 0a05f9df30 | (Justin DeVuyst)++ | tools/templates/NQP_REVISION
[release] Bump NQP revision to 2024.06
rakudo/release-2024.06: 9a8da6f2e3 | (Justin DeVuyst)++ | VERSION
[release] Bump VERSION to 2024.06
rakudo: jdv++ created pull request #5600:
Release 2024.06		 19:05
jdv the 2024.06 release is done. commit away! 19:21
patrickb: 2024.06 release happened 19:22
19:26 Geth joined
lizmat jdv++ 19:29
logger Geth and IRC logs site should be all operational again
jdv nice, thanks 19:30
afk for a bit cause its somehow nice out &
lizmat builds 2024.06
jdv: thanks again!
japhb jdv++ 19:42
finanalyst I'm having a problem using `zef install .` on my RakuAST::RakuDoc::Render distribution. The central module RakuDoc::Render requires use experimental :rakuast, but other modules in the distribution do not. However, one of those modules uses OO::Monitors, which RakuAST does not like. I can get the sub-modules to compile with 'ordinary' raku, then compile RakuDoc::Render with use experimental. How do I do this with zef 21:43
install?
23:03 Altai-man left 23:22 finanalyst 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!