backtrace() vs. _Unwind_Backtrace()

Fri Dec 5 21:24:00 GMT 2003

>>>>> "Andrew" == Andrew Haley <aph@redhat.com> writes:

Andrew> The run time is supposed to detect null pointer accesses in CNI code
Andrew> automagically; this works fine with proper signal unwinding, but not
Andrew> otherwise. The change that will enable null pointer checking even in
Andrew> CNI code is not ABI compatible, and is such a disruptive change that I
Andrew> want to postpone it until after 3.4 branches.
I filed PR 13319 for this. It is marked `[ABI]' like all the other
proposed parts of the big ABI change.
Tom