[Gambas-user] Signal #11 when overriding Variant[] - Bug revived?

[tobi]

Hi Benoît,

this afternoon, we discovered that with a newer revision the program I wrote to demonstrate sorting
of multi-dimensioned Variant[] broke. I again get Segfault when exporting an overridden Variant[]
class from my project (wasn't there a discussion about this recently?)
When I developed this program, I reported a bug which results in a Segfault that was fixed in
rev#4715. This seemed to be the result of

[main/gbx/gbx_class.c]
--8<---------------
[...]
1345         CLASS *array_type = (CLASS *)class->array_type;
[...]
--8<---------------

where array_type gets 0xc (GB_T_VARIANT or something like that) and is used as a pointer without
further checking when overriding the Variant[] class (other native array classes may suffer from
similar problems). I don't know how you made this working for me (just tried to get back to that
revision to see, if something changed in that function but found nothing).
Can you remember? - the subject of the mails was "Segfault when overriding Variant[]" (1st May 2012).
Can you tell me something about the state of affairs?

Regards,
Tobi