On Tue, 6 May 2008 09:37:50 -0600, Jerry wrote:
On Tue, May 6, 2008 at 5:28 AM, Adam Sjøgren
> I have tried to do this now, this is what happened:
> Does that help?
Yes, it certainly does. The watchpoint was never triggered. Since
that is the only place where XEmacs had the pointer in question, that
means that XEmacs did not touch the pointer, which makes the
possibility of the bug being in XEmacs code even more remote.
It doesn't prove that the bug is not in XEmacs, but it makes it
less likely. We could still have some kind of random heap corruption
bug that touches the memory in question without going through the
pointer. But that would tend to trigger segfaults, which you aren't
Did you notice that when I just open the png with C-x C-f face.png RET,
I _do_ get a segfault (as reported in <87prrzhdtq.fsf(a)topper.koldfront.dk>),
or does that not affect the conclusion?
"The unavoidable price of reliability is simplicity" Adam Sjøgren
XEmacs-Beta mailing list