Update linux_dmabuf_v1 to version 5.#4001
Conversation
We don't need to pass it through from KMS, the EGL version will be correct.
RAOF
left a comment
There was a problem hiding this comment.
This is a libmirplatform ABI break, so we need to do that dance.
Otherwise looks good (and works, too). Can we get a WLCS test? 😉
|
I've updated the ABI (please check if correct) and taken a task to look at WLCS next pulse. |
|
Sorry for not getting to this today! I got waylaid while trying to test it by #4047 |
RAOF
left a comment
There was a problem hiding this comment.
Turns out I did manage to get to this today!
Testing indicates that it's working fine. I've re-triggered the tests, as they seem to be unrelated (and possibly fixed by Saviq's recent wlcs change?)
AlanGriffiths
left a comment
There was a problem hiding this comment.
Are we safe dropping linux_dmabuf_unstable_v1? Won't we encounter clients that expect the old protocol extension?
The interface name hasn't changed - it is still "zwp_linux_dmabuf_v1". The protocol name dropped the unstable suffix, but this only affects the .xml file name. So all existing clients will continue to work. |
|
This was blocked by TICS stating that the addition to |
There was a problem hiding this comment.
We're breaking ABI, so we can consolidate the symbols into a single 2.22 stanza.
(Sorry, didn't think of that last time)
2dfc669 to
6c3b304
Compare
# Breaking changes:
- Consolidation of EDID parsing
- `User-..` and `DisplayConfigurationOutput::edid` replaced with
`::display_info`
- `mir::graphics::Edid` removed
- `miroil::Edid::Descriptor` removed
- Nvidia 565+ breaks EGLStream support (#4147)
Use `--platform-display-libs mir:atomic-kms --platform-rendering-libs
mir:gbm-kms` to use our (unstable) Nvidia-on-GBM support.
We'll follow up with a point release to make this the default.
# ABI summary:
- miral ABI unchanged at 7
- mircommon ABI unchanged at 11
- mircore ABI unchanged at 2
- miroil ABI bumped to 8
- mirplatform ABI bumped to 32
- mirserver ABI bumped to 65
- mirwayland ABI unchanged at 5
- mirplatformgraphics ABI unchanged at 23
- mirinputplatform ABI unchanged at 10
# Enhancements:
- static_config_display: add property-based display config support
(#3872)
- Implement Simulated Secondary Click (#3873)
- DRM: Explicit synchorisation support (#3717)
- live configuration facilities (#3995)
- Atomic KMS: enable bypass (#3595)
- fix: move minimized windows to the back of the Z stack (#4000)
- Implement the 'overlay_cursor' flag for WlrScreencopy (#4013)
- Support Nvidia through the atomic-kms/gbm platforms (#3979, #4060,
#4070)
- Implement hover click (#3928)
- server: add `--vt-switching` option (#4033)
- Improve support for bespoke Wayland extensions (#4018)
- Update linux_dmabuf_v1 to version 5. (#4001)
- Implement bounce keys (#3934)
- Implement slow keys (#3945)
- Implement magnifier accessibility feature (#3976)
- feature: implement sticky keys accessibility feature (#4076)
- Integrate some Rust code into the Mir build (#4022)
- (#4078) package mir-platform-graphics-stub and mir-platform-input-stub
(#4080)
- graphics: introduce EDID parsing in `DisplayConfigurationOutput`
(#4099)
- Add live_config support to miral::Keymap (#4088)
- Add tracking keymap from locale1 (#4111)
# Bugs fixed:
- Don't raise the parent of windows sent to the back (#4008)
- Update wl_output::geometry to include the correct transformation
(#4062)
- `BasicScreenshooter` does not take the output transformation into
account (#4086)
- Screenshots wrong when display rotated (again) (#3909)
- Docking kgx left/right in miriway is misplaced (#4023)
# Documentation:
- HACKING.md: Add some intructions for MESA debugging (#3992)
- doc: add Debugging section to CI docs (#4034)
No description provided.