On Thu, May 5, 2011 at 7:35 AM, Vasily Chekalkin <[email protected]> wrote:
> Hello.
>
>> 1..7
>> # Parsing src/ops/core.ops...
>> src/gc/gc_gms.c:1231: failed assertion '!PObj_on_free_list_TEST(pmc) ||
>> !"Resurrecting of dead objects is not supported"'
>
> This is definitely GC-related failure. Can you try few tips from
> https://github.com/parrot/parrot/blob/master/docs/project/hacking_tips.pod?
> Espcially about "reducing nondeterminism".
>
> Also, something like replacing dinamic selection of gc_threshold in
> gc_gms.c (line 780 or something) with constant will help a lot.
>
> --
> Bacek
> _______________________________________________
> http://lists.parrot.org/mailman/listinfo/parrot-dev
>

James,

I have written and attached two PIR probes to detect the system
parameters upon which hash seed and gc pool size are based (OS X 10.4
does not implement ASLR, so no worries there). Please run these
standalone and under make (by adding a make target that calls them).
Significant differences between make and standalone results might
indicate the root of the issue. Knowing these values on your system
might assist in recreating the failures on other systems.

Thanks,

Peter Lobsinger

Attachment: iv_time_test.pir
Description: Binary data

Attachment: sysmem_test.pir
Description: Binary data

_______________________________________________
http://lists.parrot.org/mailman/listinfo/parrot-dev

Reply via email to