Re: [tree-ssa ] Misc small improvements

[law at redhat dot com]

In message <wvlbrwe2zad.fsf@prospero.boston.redhat.com>, Jason Merrill writes:
 >> I'm not sure what you're referring to.  Are you referring to deferred pops
 >> or something else?
 >
 >I'm referring to the emit_stack_restore stuff in expand_goto_internal.
Ah.  Got it.  That was the other possibility I came to overnight.  It 
doesn't look terribly difficult to handle.  The only question is how
to test the code.  I'm open to suggestions on that issue.

 >>  > It handles more cases than yours,
 >> Do you have an example handy?
 >
 >My patch also handles COND_EXPRs with two interesting arms, at least one of
 >which is a GOTO_EXPR.
Ah.  I'll cover that as well.  I hadn't bothered to see if that was an
interesting case or not (all the ones I had seen in my analysis had only
one interesting arm..)

jeff