Welcome to the main channel on the development of MoarVM, a virtual machine for NQP and Rakudo (moarvm.org). This channel is being logged for historical purposes.
Set by lizmat on 24 May 2021.
lizmat Running this code with --profile starts to reliably segfault around 22000 for me 09:50
for ^22000 { next $_ }
lower values generally create a profile ok
note that this uses the new "next with a value" feature 09:51
looking at other similar code, it appears to be dependent on the number of times "next $_" is called 09:53
additional note: the profile generated with the values that won't segfault, is massive (10MB+), which feels to me the problem is in logging "next" itself? 09:58
hmmm... "for ^220000 { next }" generates a 4MB profile, which also feels meh 09:59
so this maybe a more general issue with profiling control exceptions ? 10:00
dogbert17 lizmat: did you try with MVM_SPESH_INLINE_DISABLE=1 16:34
according to ASAN it's a stack overflow 16:40
it looks like this: gist.github.com/dogbert17/fd9f22e7...2c0dbe745d 16:47
timo it's possible the profiler is accidentally making deeper and deeper stacks 17:48
timo i think the slack app on my phone is in a state where exactly one slack just doesn't work properly at all. i don't see new messages, every channel i open just shows the exact same stale messages wtf 18:20
brrt \o 18:59
dogbert17 Hello brrt 19:32
brrt hello dogbert17 19:36
timo greetingses 19:40
brrt ohai timo

Powered by Cro and the Raku® Programming Language.

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

Thank you!