This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: MIPS: o32 ABI problem with structs that require padding
On Mar 10, 2002, Mark Mitchell <mark@codesourcery.com> wrote:
>> OTOH, if 3.1 is to be an ABI freeze for real, installing a patch that
>> fixes bugs in the implementation of the MIPS ABIs at this point would
>> be a win.
> We might be mixing apples and oranges. We hope to freeze the C++ ABI.
The C++ ABI is a superset of the C ABI. If you change the way you
pass small structs, that are definitely present in C++, you'are
breaking the C++ ABI. So, either we put the fix in now and freeze the
ABI for real, or we'll not be holding our promise of not modifying the
ABI after GCC 3.1.
Unless you're using a much narrower definition of C++ ABI than what
I (and probably many others) would expect, in that you're only talking
about name mangling, machine-independent type layout rules and dynamic
dispatching, rtti and EH data structures, leaving calling conventions
out for their being machine-specific. If that's the case, this point
had better be made very clear whenever you say we're freezing the
C++ ABI.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{cygnus.com, redhat.com}
CS PhD student at IC-Unicamp oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist Professional serial bug killer