[webkit-dev] Potential deadlock in database--still there?

[Eric Uhrhane]

Revision 30233 added a change [and a comment about it] working around
a deadlock described in rdar://problem/5743768, which I can't see.
That change is 2 years old, and I'm wondering if the comment is still
true.  I can't find a deadlock by inspection; could someone post the
text from the bug, or explain the chain of calls that could bring it
about?

I'm trying to make DatabaseTracker thread-safe in preparation for
adding Database support in Workers, and if I can hold a lock where it
says not to, I can remove a potential race condition.

Thanks,

     Eric
     ericu at chromium.org