[PATCH] Fix PR tree-opt/25500 SRA vs struct containing a vector

[Andrew_Pinski@PlayStation.Sony.Com]

----- Message from Diego Novillo <dnovillo@redhat.com> on Wed, 15 Nov 2006 
09:20:39 -0500 -----
> Richard Guenther wrote on 11/15/06 09:01:
> > On 11/15/06, Paolo Bonzini <paolo.bonzini@lu.unisi.ch> wrote:
> >> bool single_field_in_record_p (tree type)
> >> {
> >> [ ... ]
> > single_scalar_field_in_record_p probably.  And just return false 
> on the first
> > aggregate or the second scalar.
> > 
> > Richard.
> 
> I'm fine with both suggestions.  Thanks.

Here is the newest patch with those changes included.
OK? Bootstraped and tested on i686-linux-gnu with no regressions.

Thanks,
Andrew Pinski


ChangeLog:


        * tree-sra.c (single_scalar_field_in_record_p): New function.
        (decide_block_copy): Use it.


        * gcc.dg/tree-ssa/sra-4.c: New testcase.





-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: newpatch.diff.txt
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20061118/c2e38f6f/attachment.txt>