José Fonseca wrote: > > I've been trying to figure out what is needed for releasing DRI binary > snapshots. There are two different trees with different particularities > that I'm interested for now: the trunk and the mach64-0-0-2-branch. > > Common: > > In both cases there is the need for the sources of the DRM to be built on > each particular user machine/kernel combination. For these I plan to > release a tarball with the sources from > xc/programs/Xserver/hw/xfree86/os-support/linux/drm/kernel/ as Alan made > in http://www.xfree86.org/~alanh/ . > > Trunk: > > For the drivers at the trunk I'll assume that the user will have XFree > 4.2.0 so it will only be necessary to pack: > - the DDX driver (/usr/X11R6/lib/modules/drivers/*_drv.o) > - the Mesa driver (/usr/X11R6/lib/modules/dri/*_dri.so) > and also: > - the libGL (/usr/X11R6/lib/libGL.so) > > Is this sufficient?
I belive so. In some respects it is a function of how divergent the XFree on peoples machines is from what's in DRI CVS. DRI's basically fucked up backwards compatibility on every front until quite recently - we now have backwards compatibility for kernel modules. But there are other interfaces the 3d driver deals with, specifically the DRI extension protocol to the X server. Thankfully this hasn't changed in a long time, and if it ever does, we know now to do it in a backwards-compatible way. My guess is this is going to be sufficient then for most if not all cases. Have you looked at Alan H's tools for sanity-checking DRI installs? I've always thought that these might be useful when trying to write an installer for DRI binary distributions. Keith _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel