[PATCH] Bug in Xextproto

[Bill Crawford]

On Monday 30 March 2009 11:40:33 Simon Thum wrote:
> Bill Crawford wrote:
> > On Saturday 28 March 2009 17:42:54 Simon Thum wrote:
> >> Rémi Cardona wrote:
> >
> > ...
> >
> >>> Can't we work something out with Qt folks?
> >>
> >> Good idea, but I think the above sounds like adopting a
> >> what-breaks-gets-fixed policy is the most realistic option. Adam
> >> probably has to say something enlightening.
> >
> > Weeeeell ... ISTR X existed before Qt ;o) so probably it should be they
> > who don't decide to conflict. But I'm just one man, one opinion ...
>
> I believe the problem is there was never a strict policy what header is
> defining what symbols, including by proxy. That changed, breakage.
>
> I like the idea behind CARDn & friends, but (u)intN_t is C99, and Xmd.h
> doesn't seem to be so great either.

Yeah, ...

IMO we should be defining {CARD,INT}nn in terms of the C99 types where available 
(which is pretty much everywhere now) anyway. Qt ought to be fixing this though 
as X has used these names for about twenty years :o)

> Let's hope X is C99 by the time memory models are overhauled again. Or
> whatever is the master plan here.

By which time we can probably just have u?intNN_t everywhere anyway, I guess.