Segmentation Fault in xaccFreeSplit

[Thomas Bushnell BSG]

Derek Atkins <warlord at MIT.EDU> writes:

> I see nothing wrong with this code.  It should be perfectly legal to
> set an invalid pointer like this.  Unless you are specifically doing
> bounds-checking on pointer-sets I can't see how setting a pointer
> would cause a SEGV.  You're not ACCESSING the memory then, you're only
> setting the pointer.  It should be no more special than setting the
> pointer to NULL or to some other arbitrary value.

There are processors which have special address registers, and which
do fault as soon as you store certain kinds of illegal addresses in
them rather than waiting for an access.  How important these
processors are, and whether you want to worry about them, is a
different question.

Thomas