This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch] Fix wrong type merging in LTO compiler
On Wed, May 26, 2010 at 6:16 PM, Eric Botcazou <ebotcazou@adacore.com> wrote:
>> Hmm, we already test alignment for scalars at the start of the
>> function. ?So can you instead try to move the alignment check
>> from the INTEGRAL_TYPE_P () ... case to a check right
>> before that covering all types?
>
> Not sure if TYPE_ALIGN really applies to all types (METHOD_TYPE, OFFSET_TYPE).
> At least we're sure it does to RECORD_OR_UNION_TYPE_P as per stor-layout.c.
Can you still move the test before the hash computation and comparison?
I think it definitely also applies to ARRAY_TYPE and POINTER_TYPE_P.
Ok with that change.
Thanks,
Richard.