[PATCH 1/2] dix: make screens structures run-time adjustable

[Tiago Vignatti]

Jamey Sharp wrote:
> 
> I notice that every call to MAXSCREENSALLOC and friends is passed a
> size of screenInfo.numScreens. That suggests these macros are
> misnamed, because they're really generic xalloc wrappers that do a
> multiplication for you. (They don't even check for overflow, which
> probably doesn't matter here, but it's the only reason I'd want to
> have an API for doing multiplication...)
> 
> Do we really need to introduce a new pile of magic macros? If I were
> going to introduce a macro here, it would be an assert-like macro that
> calls FatalError instead of abort; then the _FATAL cases turn into a
> simple xalloc plus a simple assert.
> 

...

> The CALLOC variants aren't used anywhere in the patch, which is good,
> because I think they're wrong. :-) As far as I can tell, xalloc
> doesn't zero the allocated memory, which the standard C library calloc
> function does.
> 
> The PLUSONE variants are also unused, and seem like API I'd want to
> discourage anyone from ever using...
> 

All right. All you talked here are just clean up and we can address 
easily after the RFC be properly discussed and the patch set be built. 
Anyway, thanks for reviewing Jamey!


Cheers,

            Tiago