[tree-ssa] Out of SSA status and issues

[law@redhat.com]

In message <1052755028.2743.368.camel@p4>, Andrew MacLeod writes:
 >On another note, shouldn't the into-ssa-pass's copy propagation only be
 >turned on when -ftree-copyprop is specified?  I think we do it all the
 >time right now... thus my problem in libstdc++ even without
 >-ftree-copyprop...
Probably...


FWIW, I dug up my changes to identify the problem variables (those
occurring in abnormal PHIs) and cobbled together the two lines of
code necessary to use that information to avoid copy propagating
those variables during SSA renaming.

Interested in playing with it to see if it resolves your problems?


jeff