[Bug middle-end/61577] [4.9.0 Regression] can't compile on hp-ux v3 ia64

[cameron.heide at betasystems dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61577

--- Comment #79 from C. Heide <cameron.heide at betasystems dot com> ---
(In reply to The Written Word from comment #75)
> 
> I think a local patch might be doing this. Rebuild without it.

I did have some other patches applied from other PRs, from previous desperate
attempts to get anything working, including PR90714, PR85412, PR87338, and tips
from PR64919.

I'm doing some more testing with combinations of them, and so far I've found
that if I start from a fresh 8.3 and apply only the essential ones, the fix
here, the operand64 removal, the nftw workaround, and -O1 to work around the
stage 2 ICE, I don't get the duplicate symbols but the compiler segfaults in
libgcc configure tests in stage 2.

If I add the '#undef MAKE_DECL_ONE_ONLY' fix from PR64919 (is it even still
relevant? It does seem to be for the stage 2 libgcc segfault symptom I see.), I
start getting the duplicate symbols in stage 1 libstdc++.