From: Adam J. <aj...@nw...> - 2005-06-29 16:41:26
|
On Wednesday 29 June 2005 11:33, Ian Romanick wrote: > Keith Whitwell wrote: > > I think there's pressure building for various backwards-compatibility > > breakages - it may make sense to try and do as much as possible all > > together - not just libGL/client driver, but open things up to all > > possibilities. > > Did you have anything specific in mind? The only issue is time. > Various freeze dates for 7.0 are rapidly approaching (or have past). > Anything that's going to be done, needs to be done soon if it wants to > get in. I really only have two goals in this area for 7.0: > > 1. Get some form of this patch in. > > 2. Make '-DDRI_NEW_INTERFACE_ONLY' standard and remove all of the > resulting dead code / structure elements. > > Anything beyond that would be icing on the cake for me. It's a little softer here than elsewhere, in terms of when the freeze needs= to=20 happen. Assuming we get a Mesa release done before 7.0, these changes just= =20 need to happen in time to stabilize for that release. I don't know what th= e=20 Mesa release schedule looks like just yet though. =46or reference, the 7.0 schedule looks like: Jul 4: Feature freeze Jul 15: Release Candidate 0 Aug 19: Release Candidate 1 Sep 9: Release Candidate 2 Sep 23: Final Release Candidate Sep 30: Release I'm willing to do the first Mesa 6.3 import anywhere up to about July 30, a= nd=20 potentially a stable branch import after that. Obviously sooner is better,= =20 but it's really Brian's call as to when the Mesa releases happen. The only major change I want to see happen in Mesa before then is the abili= ty=20 to build libGLcore from Mesa. That's pretty much mandatory for 7.0, and I= =20 have it pretty close to working. The other fixes I want are little things= =20 like 'install' targets for make. I'd consider the second set of changes (new interface only) to be sufficien= tly=20 low-impact that I'd take them later in the process, since they're basically= =20 just dropping dead code. =2D ajax |