A plea from Rakudo-land
francois.perrad at gadz.org
Sat Jul 4 16:54:10 UTC 2009
2009/7/4 Andrew Whitworth <wknight8111 at gmail.com>:
> On Sat, Jul 4, 2009 at 10:48 AM, Andrew Whitworth<wknight8111 at gmail.com> wrote:
>> This is quite interesting because it suggests that somewhere in the
>> Object cloning logic a PMC is not being properly anchored early enough
>> before a GC run is kicked off. It's also interesting that the stack
>> marking code is indeed included here.
> Okay, in r39879 I fixed one bug in the Object cloning logic that was
> causing the particular error that we were seeing. However, now that
> same test segfaults in a different place because a different PMC is
> being prematurely collected. I'll dig into that one this morning.
use mem_allocate_zeroed_typed instead of mem_allocate_typed + memset
> --Andrew Whitworth
More information about the parrot-dev