Hi,

On Thu, Oct 22, 2009 at 06:49:26PM +0900, Keith Packard wrote:
> Excerpts from Daniel Stone's message of Thu Oct 22 18:23:42 +0900 2009:
> > Right.  Why don't we just ditch CARD32 and move to the native inttypes.h
> > types everywhere but Xlib?
> 
> Do you want to do this for all XID types? Also, I assume you're
> proposing that we drop Cray support from the protocol headers at the
> same time, right?

I mean just for CARDxx and INTxx, i.e. they really only have the number
of bits implied by the type.  XID/Atom/etc, where they're just used as
unsigned long, will have to stay that way.  Shrug.

(And I have no objection to dropping Cray support.)

> My only question is whether we want to retain the existing abstract
> types to make the interfaces a bit more 'self documenting'.

In what sense is CARD32/INT32 more self-documenting than
uint32_t/int32_t? The only times I've seen people pick up CARD32,
they've gone 'oh, this is a 32-bit value we hand to the video card'.
Fail.

Cheers,
Daniel

Attachment: pgpUq0c38gr5h.pgp
Description: PGP signature

_______________________________________________
xorg-devel mailing list
[email protected]
http://lists.x.org/mailman/listinfo/xorg-devel

Reply via email to