__builtin_longjmp and rs6000 TOC

[Jeffrey A Law]

  In message < 9801300234.AA26552@rios1.watson.ibm.com >you write:
  > 	I thought that nonlocal_goto was itself used for
  > __builtin_longjmp.  Also, while I agree that nonlocal_goto is meant for
  > nested functions, nothing prevents the nesting from crossing module
  > boundaries which does require setjmp/longjmp saving and restoring.  I
  > thought that was the source of the current problems.
Agreed -- in fact the ability to non-local goto from one boundary
to another is one of the things that makes it so bloody difficult!
(and so much like setjmp/longjmp).


  > While longjmp does
  > not guarantee the nesting, unless nonlocal_goto is defined only to be
  > valid for intra-module nested jumps, I do not see how one can avoid the
  > worst case intermodule case.
I don't remember anything restricting nonlocal gotos to these cases.

jeff