[Coke] MoarVM release done 00:44
If anyone can verify it looks OK, I'd appreciate it. 00:47
00:47 jjatria left, jjatria joined 01:06 guifa left
[Coke] weird. I have gpg installed, but ake is specifically looking for gpg2 01:11
(but the moarvm release did not)
01:11 guifa joined
[Coke] (I probably thought that was weird a year ago when I first setup the azure box) 02:03
t/spec/S06-multi/type-based.rakudo.moar failing on 6.c-errata 02:05
6.d was fine. 02:12
that test not changed recently on master... 02:16
not ok 56 - native types and where clauses do not cause spurious circularities 02:23
not ok 60 - can distinguish Rat-like string in main
not ok 61 - can distinguish Num-like string in main
not ok 62 - can distinguish Int-like string in main
not ok 63 - can distinguish Complex-like string in main
^^ those 5 tests fail on 6.c errata.
if someone can take a look at that, I'll continue the release in the morning.
Only real hiccup so far.
ugexe arent those marked as todo? 02:44
the "not ok 56 - native types and where clauses do not cause spurious circularities" one isn't but i dont really follow what its trying to do. whatever it is, it relies on the legacy optimizer since it fails if run with --optimize=off 02:53
i'm inclined to say that test isnt valid 02:54
even adding a "say 1" and "say 2" to the two multis also makes it fail since thsoe side effects change what the optimizer can do with it 02:55
i.e.
multi rt107638(int $a) { say 1; 'ok' } #OK not used
multi rt107638(Str $a where 1) { say 2; Nil } #OK not used
my internet is limited this weekend fwiw, but honestly i'd say that test specific test is bad and the other "can distinguish" tests look like they are marked todo 02:57
[Coke] sorry, didn't fudge it. 03:01
when running by hand
checking...
only not ok 56 is a real failure 03:03
ugexe yeah, but thats the one im arguing the test is written incorrectly 03:04
[Coke] m: use v6.c; multi rt107638(int $a) { 'ok' }; multi rt107638(Str $a where 1) { }; rt107638(1)
camelia Cannot resolve caller rt107638(1); none of these signatures matches:
(int $a)
(Str $a where { ... })
in block <unit> at <tmp> line 1
[Coke] bisectable: use v6.c; multi rt107638(int $a) { 'ok' }; multi rt107638(Str $a where 1) { }; rt107638(1)
bisectable6 [Coke], Will bisect the whole range automagically because no endpoints were provided, hang tight
ugexe the test is bad 03:05
you can even see it in the git history of it where it gets switched back and forth between lives ok and throws ok
bisectable6 [Coke], More than 3 changes to bisect, please try a narrower range like old=2026.07 new=HEAD
[Coke], Output on all releases: gist.github.com/0465ce15e1b81ade5b...d891ca95d8
[Coke] bisectable: old=2026.07 new=HEAD use v6.c; multi rt107638(int $a) { 'ok' }; multi rt107638(Str $a where 1) { }; rt107638(1)
bisectable6 [Coke], Bisecting by exit code (old=2026.07 new=af70bff). Old exit code: 0 03:06
[Coke], bisect log: gist.github.com/7659dc62b18b854c39...60543317ab
[Coke], (2026-08-06) github.com/rakudo/rakudo/commit/59...9e0191a366
[Coke] OK. I'll open a ticket to resolve the test and continue with the release. thanks for the eyeballs.
the word "native" doesn't appear in that test on master at all 03:10
github.com/Raku/roast/issues/901 03:11
Geth nqp/main: 45907ede4c | (Will Coleda)++ | tools/templates/MOAR_REVISION
[release] Bump MoarVM revision to 2026.08
03:29
nqp/main: 8cc0a6355b | (Will Coleda)++ | VERSION
[release] Bump VERSION to 2026.08
rakudo/release-2026.08: b3d9119f5a | (Will Coleda)++ | tools/templates/NQP_REVISION
[release] Bump NQP revision to 2026.08
rakudo/release-2026.08: 24e6e5312f | (Will Coleda)++ | VERSION
[release] Bump VERSION to 2026.08