PATCH: Restore i386 binary compat for EH info

[Jason Merrill]

>>>>> Jason Merrill <jason@cygnus.com> writes:

 > 1) That should not be necessary for binary compatibility.  So long as
 >    the register mapping is backward-compatible, adding new registers should
 >    be fine.  Why was it breaking?

Looking back at the patch, I see what the deal was.  Since
FIRST_PSEUDO_REGISTER changed, the return column changed, since the one is
calculated from the other.  All we need to preserve is the column number, so
the new hard registers could go after the return column--but since they're
call-clobbered anyway we don't need to represent them.

So the patch is OK with me.

Jason