On 9/22/07, Matthew Knepley <[email protected]> wrote:

What do you think about 2.4 in October? I really want to get the new install out there.'

I'm experimenting at home about a way to compile petsc with '-fstrict-aliasing', in order to follow C99 strict aliasing rules. However, If I am right, this requiered a new layout for PetscObect's structure (I basically followed the approach being used for Python 3.0). As a side effect, some very commonly used struct members (like 'comm', 'type_name', and 'prefix') have to be accessed by casting derived objecs as PetscObject, i.e.: Vec v; ((PetscObject)v)->comm. I have also some problems when calling PetscObjectQueryFunction, as the compiler complains about strict aliasing rules, and I need to define an additional function pointer, get the function, and next cast to the a actual func. pointer to do the call. Not suure yet if this can be improved, perhaps a macro can help. Is there any interest in this be completed? It's a nonsense? In case you want me to continue this work, I will need a bit of help in a few places related to low-level calls related to sockets and sys calls. -- Lisandro Dalc�n --------------- Centro Internacional de M�todos Computacionales en Ingenier�a (CIMEC) Instituto de Desarrollo Tecnol�gico para la Industria Qu�mica (INTEC) Consejo Nacional de Investigaciones Cient�ficas y T�cnicas (CONICET) PTLC - G�emes 3450, (3000) Santa Fe, Argentina Tel/Fax: +54-(0)342-451.1594