(In reply to comment #57) > Honza: how do you figure on 95% of crashes? This is the #2 topcrash by my > queries.
95% = my guess. For this signature I can see a lot of stacks with call to initWithString. Only few stack are with call to init.
> > Is the assertion here that our first fix was so insufficient that this needs to > block again? That's what's happened.
Seems so. I'll revisit the patch ones again or try to reproduce.
(In reply to comment #57)
> Honza: how do you figure on 95% of crashes? This is the #2 topcrash by my
> queries.
95% = my guess. For this signature I can see a lot of stacks with call to initWithString. Only few stack are with call to init.
>
> Is the assertion here that our first fix was so insufficient that this needs to
> block again? That's what's happened.
Seems so. I'll revisit the patch ones again or try to reproduce.