RFC/RFA: patch for PR 22156: improve SRA for bit-fields

Andrew Pinski pinskia@gmail.com
Sun Mar 25 08:03:00 GMT 2007


On 3/24/07, Alexandre Oliva <aoliva@redhat.com> wrote:
> (*) Full disclusure: I've actually got very odd, irreproducible
> failures in gfortran's secnds.f testcase on the i686-pc-linux-gnu
> build, running on Fedora development/x86_64.  Since they're not
> repeatable, and the testcase doesn't appear to be in any way related
> with the changed code paths, I decided it's not a regression
> introduced by the patch.

secnds.f testcase is too sensitive to slight variations of timming.
It is trying to test a clock function to return the correct result but
it fails because you are hitting the function at points where the
function's value does not change.  Anyways this is not really a bug
with your code so don't worry about it.

-- Pinski



More information about the Gcc-patches mailing list