new ICE in 3.3.4 on alpha

[Gabriel Dos Reis]

Joe Buck <Joe.Buck@synopsys.COM> writes:

| See PR 15718.  There is an ICE in g77 in 3.3.4-pre that causes 10 new test
| failures (two distinct input programs, failing at multiple optimization
| levels) on the Alpha.  Seems it's a subreg problem affecting Fortran
| programs that use COMPLEX.  This is a regression from 3.3.3.

Yes, it seems to me a latent bug that has been uncovered.  I think
we've been through similar problems in the past in V3-land which had
caused us to have things like

#ifdef _GLIBCPP_BUGGY_COMPLEX
    complex(const complex& __z) : _M_value(__z._M_value) { }
#endif

We thought it was fixed...

| Given the high standards for a point release, I would consider this a
| showstopper.  We can't tolerate new ICEs in our own regression test suite.

-- Gaby