I made a mistake merge. Reverting it in c778945806 undid the state
on master, but now I realize it crippled the git merge mechanism.
As the merge contained a mix of commits from `master..staging-next`
and other commits from `staging-next..staging`, it got the
`staging-next` branch into a state that was difficult to recover.
I reconstructed the "desired" state of staging-next tree by:
- checking out the last commit of the problematic range: 4effe769e2
- `git rebase -i --preserve-merges a8a018ddc0` - dropping the mistaken
merge commit and its revert from that range (while keeping
reapplication from 4effe769e2)
- merging the last unaffected staging-next commit (803ca85c20)
- fortunately no other commits have been pushed to staging-next yet
- applying a diff on staging-next to get it into that state
Announcements:
- https://docs.mesa3d.org/relnotes/20.2.0.html
- https://docs.mesa3d.org/relnotes/20.2.1.html
I've rebased the patches accordingly and avoided:
meson.build:320: WARNING: Platform `surfaceless` is now always selected; setting this option will be an error in Mesa 20.3
meson.build:324: WARNING: Platform `drm` is now automatically selected; setting this option will be an error in Mesa 20.3
This will also fix the list in the configuration summary:
EGL/Vulkan/VL platforms: x11 surfaceless wayland drm surfaceless drm
The effective change is that "drirc.d/00-mesa-defaults.conf" will now be
installed to $out instead of $drivers and not appear under
"/run/opengl-driver/share/" anymore.
This is done to fix#89421, i.e. so that Mesa will find and load the
included DriConf defaults. The other files in "$drivers/share" will not
be moved to $out as some Nixpkgs logic might depend on them being
available under "/run/opengl-driver/share/" (e.g. addOpenGLRunpath).
Overriding it just for .drivers had the advantage of limiting rebuilds,
but it seems less clean and apparently it can interact a bit surprisingly
with some other overrides. /cc #94444.
Also this will get removed once patchelf gets updated.
Without this, the radv cache uuid would fall back to using the
timestamps of the radv and llvm shared libraries, which are fixed in
/nix/store. This caused cache collisons, which resulted in crashes
(e.g. #92807).
Quoting from the splitString docstring:
NOTE: this function is not performant and should never be used.
This replaces trivial uses of splitString for splitting version
strings with the (potentially builtin) splitVersion.
This fixes the darwin build, while also using Meson’s auto features as
much as possible. As a result, we avoid using having to specify
default drivers and instead delegate that to Mesa’s build system.
Removed other flags that were specified to the default in Mesa.
The -fno-common is needed to address undefined symbol _lp_dummy_tile
in the build.
