On Thu, Nov 03, 2011 at 07:30:52AM +1000, Peter Hutterer wrote: > On Wed, Nov 02, 2011 at 11:23:55AM -0700, Jamey Sharp wrote: > > The best option, of course, is to ensure that no bug can occur. Peter, > > what about making the length argument take 4-byte units, dividing in the > > callers, and multiplying inside _XGetRequest? > > I had that in the first version but we cannot change GetReqExtra and that > takes bytes, not 4-byte units. And having two ways to get a request with a > specific size that take different units seemed more error-prone. > SIZEOF() returns bytes as well, so imo switching to a different unit to help > those potential future developers that don't read the documentation for the > macros isn't overly useful. It's not like the 4-byte alignment requirement > of the protocol is a secred either :)
Well, I'm convinced. As far as I'm concerned, push when you're ready. I believe I already provided r-b tags for the series? Jamey
signature.asc
Description: Digital signature
_______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
