HP-PA ABI issues - framepointer handling

[Jeffrey A Law]

  In message < 20001019141038.631BD34D8E@nile.gnat.com >you write:
  > Of course if the "lame HP unwinder" could be used, we would not need to
  > be using the relatively inefficient setjmp/longjmp approach for handling
  > exceptions on HP. Indeed our Ada interest in this is both to get an
  > accurate traceback for heterogenous (gcc and non-gcc mixed) code, and
  > also to implement efficient exception handling in such environments.
Go with dwarf2 unwinders.  That's a much more efficient implementation and
it doesn't unduly restrict code generation.  That's the long term plan for
most platforms.  It just needs someone to spend the time to make it work
for the PA.

On a similar note it would be good if someone switched the PA32 port from
using embedded stabs to dwarf2 debug info.

jeff