dailz
1d1b5db3c2
Step 5 + 6: split two bench binaries into directory form with sibling
helper modules. Cargo auto-discovers src/bin/<name>/main.rs as binary
<name>; no Cargo.toml change needed.
vaapi_import_bench (973 LOC) -> 6 files:
- main.rs main() + mod declarations
- stats.rs BenchArgs + PipelineMode + FrameStats + impl
- software.rs SoftwareEncoder + SwsContext (with Drop) + create_*
+ encode_yuv_frame + finish_encoder
- pipeline_cpu.rs run_cpu_pipeline
- pipeline_gpu.rs import_frame + build_gpu_filter_graph + run_gpu_pipeline
- util.rs output_for_mode + print_detailed_results + print_comparison
sw_encode_bench (547 LOC) -> 2 files:
- main.rs main() + mod declarations (main is ~480 LOC and stays
intact per Oracle/Momis risk note on function
decomposition)
- stats.rs BenchArgs + FrameStats + impl + pix_fmt helper
Both main.rs files use #[path = "../common/mod.rs"] mod common; to keep
sharing src/bin/common/mod.rs (path adjusted for the new directory depth).
DEVATION NOTE on visibility:
The original single-file binaries accessed struct fields across what
became module boundaries (70+ accesses, e.g. encoder.yuv_frame in
run_cpu_pipeline, sws_ctx.0 in run_gpu_pipeline, stats.frames_encoded
in main, stats.mmap_us in main). Rule 2 forbids widening visibility on
struct fields. After 2 build attempts confirmed there is no way to
perform the specified split without widening, the minimum necessary
pub(crate) was applied to:
- vaapi_import_bench/stats.rs: BenchArgs fields, PipelineMode (type
only), FrameStats fields, FrameStats::{avg_ms, avg_total_ms,
achieved_fps, theoretical_fps}
- vaapi_import_bench/software.rs: SoftwareEncoder fields (enc_video,
octx, yuv_frame, codec_name), SwsContext.0, all four functions
- vaapi_import_bench/pipeline_*.rs: run_cpu_pipeline, run_gpu_pipeline,
import_frame (build_gpu_filter_graph kept private)
- vaapi_import_bench/util.rs: output_for_mode, print_detailed_results,
print_comparison
- sw_encode_bench/stats.rs: BenchArgs fields, FrameStats fields,
FrameStats::avg_ms, pix_fmt
No pub (truly public) was used anywhere. All widening is to pub(crate),
keeping these symbols private outside the binary crate.
Verification (all green):
- cargo build --bins / cargo build --release --bins
- cargo test (79 lib + 3 integration = 82 pass, 1 ignored — unchanged)
- cargo clippy --all-targets -- -D warnings
- cargo fmt --check
- --help smoke test on both binaries
wl-webrtc
Wayland screen capture and encoding tool.
Prerequisites
- Rust toolchain (1.87+; MSRV pinned to match
u32::is_multiple_of/Option::is_none_orusage):rustup default stable - FFmpeg 6.0+ dev libraries with VAAPI support:
- Arch:
pacman -S ffmpeg - Ubuntu/Debian:
apt install libavcodec-dev libavformat-dev libavutil-dev libswscale-dev libva-dev - Fedora:
dnf install ffmpeg-devel libva-devel
- Arch:
- Wayland dev libraries:
- Arch:
pacman -S wayland-protocols - Ubuntu/Debian:
apt install libwayland-dev wayland-protocols - Fedora:
dnf install wayland-devel wayland-protocols-devel
- Arch:
- DRM dev libraries:
- Arch:
pacman -S libdrm - Ubuntu/Debian:
apt install libdrm-dev - Fedora:
dnf install libdrm-devel
- Arch:
Build
cargo build --release
Run
# Basic capture to file
wl-webrtc --output output.mp4
# With custom FPS and bitrate
wl-webrtc --output output.mp4 --fps 60 --bitrate 8000000
# Specify DRM device for hardware encoding
wl-webrtc --output output.mp4 --drm-device /dev/dri/renderD128
# Verbose mode
wl-webrtc --output output.mp4 -v
# WebRTC streaming mode (HTTP signaling server)
wl-webrtc --port 8080 -v
# Force a fresh portal authorization dialog (ignore saved restore token)
wl-webrtc --output output.mp4 --no-persist
# Pin the capture backend instead of auto-detecting
wl-webrtc --output output.mp4 --backend portal # or: --backend screencopy
CLI Arguments
src/args.rsis the authoritative source. Runwl-webrtc --helpfor the live list.
| Argument | Default | Description |
|---|---|---|
-o, --output |
(optional) | Output file path (e.g. output.mp4). Optional when using --port for WebRTC mode. |
--output-name |
auto | Wayland output name to capture |
--fps |
30 | Target frames per second |
--codec |
h264 | Video codec (h264 only for MVP) |
--hw-accel |
vaapi | Hardware acceleration method |
--drm-device |
auto | DRM render device path |
--bitrate |
auto | Target bitrate in bps |
--max-bitrate |
8000000 | Max bitrate cap for WebRTC mode (caps BWE escalation; no effect in MP4 mode) |
--gop-size |
auto | Group of Pictures size |
-v, --verbose |
false | Enable verbose logging |
--backend |
auto | Capture backend: screencopy (wlroots) or portal (KWin/KDE). Auto-detected if omitted. |
--port |
0 | WebRTC HTTP signaling server port. 0 keeps MP4 file output mode. |
--no-persist |
false | Force re-authorization (ignore saved portal restore token) |
--stats |
false | Print per-second pipeline statistics for stutter diagnosis |
Capture backends
The tool supports two Wayland capture backends, auto-detected by default:
- wlr-screencopy (preferred when
zwlr_screencopy_manager_v1is advertised): works on wlroots-based compositors (Sway, Hyprland, etc.). - XDG Portal / PipeWire (fallback when D-Bus ScreenCast is available):
works on KWin/KDE and any compositor that implements the XDG Desktop Portal
screen-cast protocol. The first run shows an authorization dialog; a restore
token is cached under
wl-webrtc/portal-restore-tokenso subsequent runs don't re-prompt (use--no-persistto force a fresh authorization).