New Fortran regression

[Andrew Pinski]

On Jul 22, 2005, at 4:24 PM, Thomas Koenig wrote:

> Compilation of two Fortran test cases now fails with -O3:
> real_const_1.f and real_const_2.f90.
>
> I have submitted this as PR 22619.
>
> Apparently, running the Fortran testsuite is something not
> everybody does before committing patches :-|

Or it is only happens on some targets and testing on the target
which they are testing on does not have this failure.
This is one such example.

-- Pinski