nwc10 good *, * 07:07
jnthn: mutex abort: Device or resource busy 08:03
er, wrong order. that's from ./perl6-m -Ilib t/spec/S17-supply/syntax.t
ok 51 - Can only be in one whatever block at a time 1..3
mutex abort: Device or resource busy
oh, paste got a bit mungd 08:04
nwc10 that's with ASAN. And also under valgrind. 08:52
No earlier problems reported
it flaps
lizmat sanity check: it's MVM_JIT_DISABLE to disable the jit, right ? 14:13
psch yes 14:14
nwc10 my shell history agrees
lizmat psch nwc10 thanks 14:19
brrt aye, that is it 14:33
lizmat: anything you suspect to be jit-broken?
lizmat I was looking at #126842 14:39
the problem goes away if you do an nqp::force_gc after each run
consistently fails in the 77th or the 70th (with MVM_SPESH_DISABLE) run 14:40
added it to the ticket
brrt it fails even with MVM_SPESH_DISABLE? then it is certainly a GC issue 14:44
aw.. :-(
you know what would be evil? randomized GC
timotimo as a tool for shaking out more failures? 14:47
force into the open, so to say
i think i want an api for bytecode annotations that lets me more or less go through all annotations in a range of pc values 14:48
would probably be enough to have a second function that gives the next pc with an annotation that comes after a given pc 14:49
alternatively always put the next annotated pc into the result struct
but making a struct like that bigger has an ABI implication
timotimo appointment & 15:01
timotimo parents are taking me out to dinner <3 17:54
konobi jnthn: unlikely... but still around? 23:41
lizmat jnthn is enjoying some R&R :-) 23:48
konobi ah... can someone have a look at the last comment on github.com/MoarVM/MoarVM/issues/306 ? 23:49
timotimo ooooh

Powered by Cro and the Raku® Programming Language.

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

Thank you!