* fix(capture): map the DMA-BUF frames niri actually sends
`uur run` reported "No Wayland frames arrived within 120 seconds" on a
session where the portal handshake succeeded. The session log showed the
stream never leaving negotiation:
portal session ready (video node 73)
stream state: connecting
stream state: paused
stream state: error - no more input formats
stream state: unconnected
Three separate defects in `uur-pw-capture`:
1. The format filter omitted `modifier`. niri marks that property
MANDATORY (SPA flags 0x18), and SPA filtering fails unless both sides
carry a mandatory property -- which surfaces as "no more input
formats". The filter now advertises LINEAR, the layout the frame
protocol already assumes.
2. `on_process` only handled `data != NULL` and `SPA_DATA_MemFd`. niri
delivers SPA_DATA_DmaBuf planes, which carry an fd and no pointer, so
every frame was dropped in silence. Since the session negotiates
LINEAR, that fd is directly mappable without GBM/EGL; the helper now
mmaps it. A non-linear modifier is refused instead of publishing
garbage.
3. DMA-BUF producers leave `chunk->size` and `maxsize` at 1. Both the
mapping length and the row count were derived from those, so the mmap
covered one byte and `rows` came out zero -- publishing an all-black
frame. Both now use the negotiated stride and height.
Verified on niri 26.04 with Wine 11.19: the stream reaches `streaming`,
~60 fps (1195 frames in 20 s), and the published frame decodes to the
real desktop rather than black.
* fix(capture): validate DMA-BUFs and recover interrupted screen streams
Use immutable transport files across geometry changes and remap Wine readers safely. Bound stream recovery and preserve portal stop-sharing intent. Refs #29.
* fix(capture): watch Mutter layout changes and support older build headers
* fix(capture): keep display changes outside the failure retry budget
* fix(capture): honor portal stop signals during GNOME recovery
Subscribe before starting on the portal-owned D-Bus connection and accept the real Session.Closed dictionary body. Correlate Mutter's close-before-MonitorsChanged ordering without restarting an explicit Stop Sharing session.
Verified in a private GNOME 51.1 namespace with real GBM DMA-BUF, 100/125 percent scaling, portrait and landscape, and native Stop Sharing. Existing 48 tests and strict clippy pass. Real Wine GDI reader captured 120 final-run frames across four layout changes with no hook fallbacks. No test scripts added.
---------
Co-authored-by: Xuc Pan <[email protected]>
* fix(capture): choose a cursor mode the portal advertises
`uur run` on a Wayland session aborted during capture negotiation:
Error: Portal request didn't succeed: Other
Error: ZBus Error: org.freedesktop.portal.Error.InvalidArgument:
Unavailable cursor mode 2
`negotiate` hardcoded `CursorMode::Embedded` (2). The portal frontend
validates that against the backend's `AvailableCursorModes` and rejects
anything outside it, so any backend that does not offer Embedded could
never start a session. A frontend whose backend property cache is empty
reports `AvailableCursorModes = 0` and then rejects every mode, which is
what made this fail even on backends that do support Embedded: restarting
`xdg-desktop-portal` restored the flags from 0 to 7 and mode 2 was
accepted again.
`preferred_cursor_mode` now reads the advertised flags and picks Embedded,
then Hidden, then Metadata. If none are advertised it fails with an
actionable message naming the portal restart, instead of surfacing a raw
D-Bus error. Older portals that do not implement the v2 property keep the
previous behaviour.
The same hardcoded mode in the RemoteDesktop input path is routed through
the shared helper.
Adds four tests: preference order, subset fallback, the empty-advertisement
regression, and source-type mapping.
* fix(capture): preserve portal errors and report cursor fallback
---------
Co-authored-by: Xuc Pan <[email protected]>
* fix(wine): select the client root that actually holds bin/
`uur run` aborted with "WebView2 bootstrapper missing from the client
directory" on a complete install.
`find_client_dir` searched depth-first for `GameViewerServer.exe` and
returned the first hit, special-casing only a parent named `bin`. NetEase
ships the payload twice: as `<root>/bin/` and as a sibling
`<root>/<version>/`. When readdir order reached `<root>/<version>/` first,
that directory was returned as the client root. It has no `bin/`
subdirectory, so every `join("bin")` lookup missed -- the WebView2
bootstrapper and the terminal proxies -- and the session bailed out.
`deploy_terminal_proxy` then created a spurious `bin/` inside the version
directory and wrote proxies into it.
Now every directory holding `GameViewer.exe` is collected and ranked:
`bin/GameViewerServer.exe` scores highest because that is the layout uur
drives, then `bin/MicrosoftEdgeWebview2Setup.exe`, then a top-level
server. Ties break on path depth, then lexically, so the choice no longer
depends on readdir order. A versioned copy is still used when it is the
only candidate.
Adds four tests covering both layouts, readdir stability, and the
fallback.
* fix(wine): require complete payloads and support flat client layouts
---------
Co-authored-by: Xuc Pan <[email protected]>
* fix(wine): make setup detect a provisioned prefix
Two defects made `uur setup` fail on a fully installed prefix.
1. The service check matched `gameviewerserver.exe`, but the Windows
service NetEase registers uses `GameViewerService.exe` as its
ImagePath. `GameViewerServer.exe` is the streaming server and never
appears in the service registration, so `client_service_registered`
and `query_client_service` always returned false. A completed
install was reported as "GameViewerService has no valid registered
executable" and setup re-downloaded and re-installed every run.
Both checks now share one `SERVICE_IMAGE` constant, and the
`system.reg` parse is factored into a pure `service_image_path`
helper so it is testable without a Wine prefix.
2. `wineboot --init` returns before wineserver flushes
`system.reg`/`user.reg` to disk (measured ~7s on Wine 11), so a
freshly created prefix was misread as uninitialized and setup
aborted with "Wine prefix initialization failed (exit status: 0)".
Added a bounded `flush_prefix` barrier that waits for `wineserver -w`
to return, then polls for the registry files.
Verified on Wine 11.19: a clean install into an isolated prefix
succeeds end to end, re-running setup takes the already-provisioned
fast path, and `setup --repair` verifies without reinstalling.
* fix(wine): bound prefix readiness waits without blocking on services
---------
Co-authored-by: Xuc Pan <[email protected]>
wpaperd writes configured image paths verbatim into the state symlink, so
targets relative to its working directory dangle under symlink semantics
and link.canonicalize() fails for every output. Read /proc/<pid>/cwd where
the process is matched and resolve relative targets against it; absolute
targets keep filesystem semantics.
wpaperd mirrors the current wallpaper of every output as a symlink
under $XDG_STATE_HOME/wpaperd/wallpapers, rewritten on each change.
Identify the wpaperd process in the runtime provider scan and resolve
the newest link, joining relative targets with $HOME.
Use the configured AUR key for clone and push with an explicit known-hosts file, and add an AUR-only dispatch mode that avoids rebuilding release packages.
Build conventional artifacts on Ubuntu and the Arch package in an Arch container, merge both artifact sets, generate one checksum manifest, and publish them together. Allow an explicit manual rebuild to refresh AUR metadata.
Trigger packaging only after a maintainer publishes a GitHub Release, keep manual rebuilds tied to an explicit existing tag, and use the selected release tag consistently for assets and AUR metadata.
Declare the Windows API level required by the terminal proxy so Ubuntu 22.04 MinGW exposes CancelSynchronousIo. Patch script shebangs inside the Nix sandbox, install the target-triple Rust binary through the shared staging script, and keep Nix CI resource usage bounded.