dalek arVM: 695fdf9 | jnthn++ | src/ (5 files):
Improve the VM exit strategy.

It now doesn't try to do all the cleanup work by default, instead leaving it to the OS, which can no doubt do it faster. There is now a --full-cleanup flag for requesting full cleanup, however. In either case, special effort to gracefully join foreground threads is made, so "Thread.start({ sleep 1; say 'bye' }); say 'hi'" in Perl 6 will now output hi\nbye\n reliably (and passing :app_lifetime will not show the bye\n, and seems to exit cleanly too provided --full-cleanup isn't passed).		 11:21
dalek arVM: 9a61eb6 | jnthn++ | src/core/interp.c:
Unbreak CGOTO compilation.

It can't cope with ops in oplist but having no stub in interp.c.		 14:46
dalek arVM: c7c18ec | jnthn++ | src/ (4 files):
Correctly handle temp roots stack in callbacks.
 17:38
retupmoca jnthn: how do I see if an MVMObject is a perl6 role? 22:29
(I...think that's what I need?)
I'm looking for a TypedArray role I think 22:30
jnthn With difficulty in C land... 22:31
You can tell if it's a type object at least.
(See flags) 22:32
jnthn Can also see the REPR name (will likely be P6opaque) 22:32
retupmoca yeah, I found the REPR name 22:33
P6opaque
jnthn One other trick is taking the "static" off dump_backtrace and then inserting a call to it in some interesting place to see where you are code-wise. 22:34
retupmoca jnthn: how is repossession of something like TypedArray[Str] supposed to work? The typed role doesn't exist until something uses TypedArray 23:08
and I think that's why it's breaking
man, perl6 at the C level is confusing 23:10

Powered by Cro and the Raku® Programming Language.

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

Thank you!