Proposed targets to deprecate for 3.4

[Hans-Peter Nilsson]

On Wed, 1 Oct 2003, Zack Weinberg wrote:
> Anyway, the
> suggestion to use prologue_epilogue_contains seems like the Right
> Thing to me.

Aha, you mean:

#define FINAL_PRESCAN_INSN(INSN, OPVEC, NOPERANDS) \
 (cfun->machine->in_prologue = prologue_epilogue_contains (INSN))

Yeah, that seems like it.  I should also rename in_prologue.
(Huh, bug there?)

Also, prologue_epilogue_contains is undocumented but already
used in rs6000.c.

brgds, H-P