It would seem
to me that the guard allocator should catch any errors that would
cause
GMmalloc to fail, and should *not* die in GMmalloc. Is this
observation
correct?
No. If you write to a random memory location which is not protected
by GMalloc, and which happens to contain internal GMalloc state, then
you can corrupt GMalloc's internal state in a way that will cause it
to crash.
Or is a fault in GMmalloc an indication of a memory bug in my
application?
Yes.
If it is a problem in my application dieing in GMmalloc doesn't
help me debug it.
What you can try is figure out from which address GMalloc loads the
invalid pointer value whose dereferencing causes the crash. Chances
are that it'll be the same address each time you run your application
(unless GMallloc includes address randomization logic). If so, you
can try putting a watchpoint on that address to catch all writes to
that address, in order to see where in your program you modify the
value at that address.
Jonas
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/xcode-users/email@hidden
