This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: -m(no-)align-double compatible libgfortran
On Tue, Feb 20, 2007 at 06:57:08PM +0100, Tobias Schl?ter wrote:
> Steve Kargl wrote:
> >>>This is a "m"achine specific option which
> >>>should not be touched in the front-end at all.
> >>I think most of us Fortran people decided to disagree with you.
> >
> >FWIW, I agree with Andrew.
>
> Sorry, I may have gotten a wrong impression from the previous
> discussion(s). Maybe my belief that interfaces internal to the compiler
> should not be broken by compiler flags skewed my perspective.
>
> Anyway, if there were a patch, would you oppose it?
I will not oppose the majority.
> >>There seems to be only one point where this affects the
> >>gfortran/libgfortran interface, and this can easily be made to work by
> >>adding padding to a structure, independent of whether the option is given
> >>or not.
> >
> >Have you checked libgomp?
>
> No, admittedly I didn't think of it yet. A brief scan of its .h files
> revealed no doubles in structures, and only two global doubles (I don't
> know if these can be broken by -malign-double, and aligning them
> wouldn't be hard), so I'm inclined to believe that libgomp is ready for
> -malign-double.
I think someone will also need to audit SEQUENCE in derived types
and the interaction with ISO C Binding. Personally, I think it
is dangerous slope to try to "fix" -malign-double issues; particularly,
when it is *documented* to cause problems.
--
Steve