[Webkit-unassigned] [Bug 29026] CRASH: fastRealloc crashes on realloc(ptr, 0)

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=29026





--- Comment #8 from Mike Belshe <mike at belshe.com>  2009-09-11 13:38:41 PDT ---
(In reply to comment #7)
> I agree that fastRealloc should have the same semantics as realloc, or be
> renamed. However, I'm not sure which specification you're citing, as the
> behavior is different per Mac OS X man page:
> 
> "If ptr is NULL, realloc() is identical to a call to malloc() for size bytes.
> If size is zero and ptr is not NULL, a new, minimum sized object is allocated
> and the original object is freed."
> 
> As a result, the behavior of realloc(0, 0) is implementation defined, but
> realloc(nonNull, 0) doesn't return 0 unless minimum sized object allocation
> fails.

Good digging, Alexey! I was using this reference: 
http://www.opengroup.org/onlinepubs/009695399/functions/realloc.html

So it appears that BSD (and Mac OSX) already diverges from the C spec.

In light of this, the bug gets deeper!  fastRealloc() is implemented using
realloc(), and realloc on different platforms behaves differently in this case.
 We should document that, and normalize the fastRealloc() to always behave like
BSD realloc (I assume that would be apple's preferred API?)

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.