[Bug ipa/70646] [4.9/5/6/7 Regression] Corrupt truncated function

[jamborm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70646

--- Comment #28 from Martin Jambor <jamborm at gcc dot gnu.org> ---
(In reply to rguenther@suse.de from comment #27)
> I was arguing that if IPA proves a condition to true/false then it
> should adjust it that way in modification phase.

The thing is that it does not prove it correctly.  If we modify the
testcase so that only the byte at offset zero is constant and others
are not, inlining would still think the whole eight-byte access is to
a known constant.

Redirecting to some builtin_true or builtin_false would have made
debugging easier, I agree.