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: altivec: even more boring altivec.h patches


Am Mit, 2002-04-24 um 13.12 schrieb Aldy Hernandez:

> more patches from the 
> "so-boring-i-can't-believe-i'm-doing-this-i-hate-my-job-somebody-kill-me"
> department.

Haha...
 
> it turns out a bit of unclear wording in the altivec specs mention that
> we must have overloaded accessor functions for the individual instructions.

Care to share where this unclear wording might be?

> ie.  we must have vec_vaddubm(), vec_vadduhm(), vec_adduwm(), ad nauseum 
> functions in addition to vec_add().  don't ask me why.  i didn't write
> the specs.

This sucks, but....

> --- config/rs6000/altivec.h	24 Apr 2002 10:54:32 -0000
> *************** Boston, MA 02111-1307, USA.  */
> *** 44,50 ****
>   /* You are allowed to undef this for C++ compatability.  */
>   #define vector __vector
>   
> ! #define bool unsigned
>   #define pixel short
>   #define __pixel short
>   
> --- 44,50 ----
>   /* You are allowed to undef this for C++ compatability.  */
>   #define vector __vector
>   
> ! #define bool signed
>   #define pixel short
>   #define __pixel short

What is this about and why isn't it mentioned in the ChangeLog?

-- 
Servus,
       Daniel


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