timotimo i seem to get quite a few SEGV originating from the decont op in interp.c 00:51
arnsholt jnthn: I think doing positional access on an uncomposed P6opaque will deref a null pointer. Not a big deal, or something that should be fixed? 10:48
nwc10 well, being able to SEGV the VM is a bad thing 10:49
it feels like it ought to fire off a suitable VM-level exception 10:50
is it easy & cheap to do that? Or is it complex and painful?
arnsholt if(!repr_data) MVM_exception_throw_ad_hoc(tc, "Don't do that!"); // More or less 10:54
nwc10 I think that it's worth it, with a descriptive message 10:55
ie define "that" :-)
(I am not jnthn) 10:56
arnsholt Yeah, other code paths in P6opaque already do that check 10:57
[Coke] gist.github.com/awwaiid/ef3f0abcfa96e34977b4 - this showed up in my perl6 news net. 13:18
arnsholt can haz HashAttrStore with positional and associative delegation
[Coke] ... www
arnsholt \o/
jnthn: Do you want me to add some tests to NQP to excercise the delegation stuff too? 13:20
[Coke] in general, tests is good. 13:43
arnsholt Hmm. Making HashAttrStore delegate boxing is a bit trickier than positional and associatives 14:41
jnthn arnsholt: Please do, though be careful where you put them unless you're going to impl this on JVM also :) 15:30
japhb [Coke]: That awwaiid gist makes me unnaturally happy. 16:32
japhb glares at the mountains of Python libraries at $day-job
[Coke] I got someone who just uses perl to get their job done interested in perl6 by showing them that gist. 16:34
japhb [Coke]: I bet. It's a pretty powerful dozen lines or so. 16:37
timotimo yay, i has a fedora 22 18:24
flussence gist.github.com/flussence/1f7180a162f84337188d - here's a(nother) GC crash from trying to do parallel threaded stuff. I think this one's been around a few months now... 19:38

Powered by Cro and the Raku® Programming Language.

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

Thank you!