finished changes for solving strict aliasing issue

Barry Smith bsmith at mcs.anl.gov
Tue Oct 2 14:40:54 CDT 2007


   Lisandro,

    You should be able to push into petsc-dev-dalcinl, then anyone will be able
to pull from that.

   Barry


On Tue, 2 Oct 2007, Lisandro Dalcin wrote:

> Not sure if any of you had the time to look at the small example I sended
> to show the strict aliasing issue.
> 
> Anyway, I finished the changes needed to solve this issue, all C and
> Fortran examples pass. If -Wno-strict-aliasing is removed, GCC will
> still complain in many parts, but now we are sure that it will not
> generate bad code when using PetscObject and derived structures. BTW,
> I commented out all thos XXX_Publish() routines. Is this feature used
> in any way? perhaps some external code? If this still have to be
> supported, perhaps I can change PetscHeaderCreate() to pass the
> publish routine.
> 
> Could you create a petsc-dev clone in the hg server in order I can
> push this for your to review?
> 
> Regards,
> 
> PD: Many thanks to Stephan for his 'sed' tip. It was of great help for
> this to be finished.
> 
> 
> 




More information about the petsc-dev mailing list