This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [ping] [4.5/m32c] libstdc++ build crashes emitting debug_info


On Sat, Apr 17, 2010 at 12:00 AM, DJ Delorie <dj@redhat.com> wrote:
>
> Then what about foo[40000] ?
>
> If we arbitrarily choose signed or unsigned, that cuts the maximum
> size of objects on m32c in half, from 64k to 32k.

Yes, we do.  Note that sizetype is _always_ unsigned, so if you keep
your patch as-is you'd always zero-extend.  And it's certainly worse
to miscompile for foo[-4] than to cut maximum object size in half
(which btw is already half of the address-space for all targets).

So - please follow the "sizetypes are sign-extended" rule as GCC
does everywhere else.

Richard.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]