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.
jnthnwrthngtn morning o/ 09:29
nine mrnng 09:31
Nicholas \o 09:32
Geth MoarVM/new-disp: e969abe64c | (Jonathan Worthington)++ | src/core/frame.c
Ensure spesh correlation ID is initialized

When creating a frame for uninlining, we neglected to zero it, which could lead to an uninitialized value being read, and potentially to bogus spesh logging taking place.		 09:38
jnthnwrthngtn That fixes the "not great" valgrind output :) 09:40
dogbert17 nine: did you manage to repro the MoarVM Panic? 14:41
nine dogbert17: not zet 16:27
...until now that is 16:29
dogbert17 nine++, that's excellent news 16:54
nine So far it has eluded rr though 16:56
dogbert17 I thought you had given up on rr
nine Why would I? 17:00
If an error is reproducible in rr, that's certainly preferrable to doing it the old fashioned way. It's just some errors where there doesn't seem to be a chance for rr 17:02
dogbert17 I see, so it's on a bug by bug basis then 17:03
nine This may be one where the good old fashioned way may be the only one 18:15
dogbert17 so it plays elusive 18:27
nine Intruigingly I have yet to see it fail with either MVM_SPESH_BLOCKING=1 or MVM_SPESH_DISABLE=1 18:56
nine But then, the same is true for running it in gdb 20:31
dogbert17 FWIW: the panic always seem to appear after all the tests have been run 20:33
nine true
dogbert17 and it is always trying to find the method !cursor_init 20:34
nine Maybe it's actually about Log::Async 20:36
anyway gotta go to bed now 20:37
dogbert17 sleep well

Powered by Cro and the Raku® Programming Language.

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

Thank you!