> I think the discussion should begin with reevaluating whether or not
> the memory savings from recycling SSA_NAMEs is still worth the headache.
That's a separate project that I'd rather not bundle with strength
reduction, because the two are unrelated conceptually.
My opinion is that instead, the following question should be the
starting point: Some code in GCC has a demonstrable need to iterate
over the SSA_NAMEs in current use by the program being compiled. Do we
want to provide a reasonable way to do that iteration? Currently we
don't provide it. That seems wrong to me.
I don't much care about the details of *how* the facility is provided,
although if implementing it doesn't divert me too much I'm happy to do
it. Plenty of people who know more than I do will have opinions about
how it should be done, and that's fine. I don't want to decide how it
should be done, necessarily.
I would simply like to begin with agreement that it should be
done. Or, if the consensus is that it should not be done in spite of
the need, some reasonable advice about what to do instead (I don't see
a good alternative).
-- Robert
