Eric Botcazou wrote:
Not by GDB.
Then an approach could be to patch the GDB unwinder to make it follow the GCC
conventions.
The correct approach would be to have gcc generate correct DWARF data.
Indeed, I have patch gdb to work around the inaccurate CIE.
You seem to be under the misapprehension that I'm talking about exception
handling in gcc. I'm not.
Well, GCC doesn't generate 2 flavor of CIEs (on SPARC) so considerations that
apply to CIEs in .debug_frame apply to CIEs in .eh_frame as well.
No, they don't. Apparently, EH frame unwinding doesn't need
correct CIE data. Debugging does.
My question was whether there was a good reason for ignoring
the offset in a RA expression.
--
Michael Eager [EMAIL PROTECTED]
1960 Park Blvd., Palo Alto, CA 94306 650-325-8077