https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61397
Jeffrey A. Law <law at redhat dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Priority|P2 |P4
Status|UNCONFIRMED |NEW
Last reconfirmed| |2015-02-24
CC| |law at redhat dot com
Ever confirmed|0 |1
--- Comment #9 from Jeffrey A. Law <law at redhat dot com> ---
It seems to me Andreas has a valid point in that the generated for that test
could be improved for the set of options in c#0. He also has a point that
apparently there was a period of time when we did generate better code.
And while the code exhibits undefined behaviour for -m32, unless someone shows
that undefined behaviour is a factor in the difference in quality of code
regression Andreas saw I would suggest we simply ignore the undefinedness of
that part of the testcase.
So it seems to me that a regression marker is appropriate. However, it's hard
to see it as worthy of a P2 status. So I'm going to downgrade to P4.
