The tty1 login hook ran systemctl --user start graphical-session.target, which systemd refuses (RefuseManualStart=yes), so none of the NixOS session's services started. The units in infra/systemd were wanted by graphical.target, a system target, although weft-appd's needs the session user's runtime and home directories and the compositor publishes its socket to the user manager; run as system units the shell could never learn the socket. weft-session.target binds graphical-session.target and wants the compositor, shell and weft-appd, in the NixOS configuration and as a user unit in infra/systemd; the login hook starts it. The services are part of the graphical session, the shell no longer hardcodes the compositor's socket, and the compositor's units unset WAYLAND_DISPLAY and DISPLAY so a restarted compositor does not take them for an enclosing session and pick the nested backend. As a user unit, weft-appd drops ProtectSystem= and PrivateTmp=, which a user manager implements with PrivateUsers=yes, under which the setuid mount helper cannot mount verified images. docs/building.md adds the VM's disabled graphics and its mount helper, which the Nix store cannot hold setuid, to what still keeps the image from a working desktop. Rendered from the NixOS configuration, and as installed from infra/systemd, the units pass systemd-analyze --user verify.
7 KiB
Building WEFT OS
Prerequisites
Linux (x86_64 or aarch64). WEFT OS is built and validated on Linux only; runtime components require Linux kernel interfaces.
System packages (openSUSE):
sudo zypper install -y \
libwayland-devel libxkbcommon-devel \
libglvnd-devel libgbm-devel libdrm-devel \
libinput-devel seatd-devel libudev-devel systemd-devel \
pkg-config clang cmake python3
Rust toolchain: pinned in rust-toolchain.toml. Run rustup show to confirm the active toolchain matches.
Workspace crates (non-Servo)
cargo build --workspace \
--exclude weft-servo-shell \
--exclude weft-app-shell
These crates do not require Servo and build in under two minutes.
weft-compositor, weft-servo-shell, weft-app-shell (Linux)
cargo build -p weft-compositor
cargo build -p weft-servo-shell
cargo build -p weft-app-shell
Without --features servo-embed, the servo-shell and app-shell binaries compile for type and lint checks but exit with an error at startup; they never report readiness. This is the default and the CI baseline.
Servo embedding (optional, slow)
cargo build -p weft-servo-shell --features servo-embed
cargo build -p weft-app-shell --features servo-embed
This fetches and compiles the Servo and Stylo forks at the revisions recorded in Cargo.lock (see crates/weft-servo-shell/SERVO_PIN.md). Expect 30–60 minutes on a clean build. Servo's dependencies include SpiderMonkey (C++), which requires clang and python3.
Frame check
tests/frame/check_frame.py runs weft-compositor nested on Xvfb with the
winit backend, starts a Servo host showing tests/frame/reference.html, and
checks the presented pixels. It needs Xvfb, xwd, ImageMagick's convert
and a Mesa OpenGL driver (llvmpipe is sufficient).
cargo build -p weft-compositor
cargo build -p weft-servo-shell -p weft-app-shell \
--features weft-servo-shell/servo-embed,weft-app-shell/servo-embed
python3 tests/frame/check_frame.py --host system
python3 tests/frame/check_frame.py --host app
python3 tests/frame/check_frame.py --host app --slow-style 3
--host app also requires the page on screen within two seconds of
weft-app-shell printing READY; --slow-style delays a stylesheet that
reveals the page so premature readiness becomes visible. The stylesheet is a
file next to the page that the harness keeps from being opened with a file
lease, so the delay must stay below /proc/sys/fs/lease-break-time.
Logs and the last capture are written to target/frame-check.
tests/frame/check_input.py uses the same nested desktop to send real X
pointer and keyboard input (with xdotool) through weft-compositor to
weft-servo-shell showing tests/frame/input.html, and checks that the
page reacts to clicks on either side of a boundary and to the a and
Shift+a keys. Logs and the last capture are written to target/input-check.
python3 tests/frame/check_input.py
The application checks run weft-appd with weft-runtime (built with
wasmtime-runtime) and weft-app-shell (built with servo-embed) on the
same nested desktop and drive the demo apps with real keyboard input:
python3 tests/frame/check_counter.py # Counter round trip through the session bridge
python3 tests/frame/check_notes.py # Notes stores exactly what is typed, refuses stale saves
python3 tests/frame/check_app_data.py # Notes data from the earlier layout is moved and kept
python3 tests/frame/check_confinement.py # an app page reaches only its UI files and its session bridge
Demo apps (wasm32-wasip2)
Each demo is a standalone crate in examples/. Pre-built app.wasm binaries are committed; they are reproducible with the toolchain pinned in rust-toolchain.toml. To rebuild:
rustup target add wasm32-wasip2
cd examples/org.weft.demo.counter
cargo build --release
# output: target/wasm32-wasip2/release/app.wasm
cp target/wasm32-wasip2/release/app.wasm app.wasm
rm -rf target
The signature covers every file in the package directory, so remove the crate's target directory before signing it again. The demo packages are signed with the test key in examples/keys, which is a fixture, not a signing authority:
weft-pack sign examples/org.weft.demo.notes --key examples/keys/weft-sign.key
weft-runtime with Wasmtime
cargo build -p weft-runtime --features wasmtime-runtime,net-fetch
Without --features wasmtime-runtime, the runtime compiles for type, lint and unit checks (the portable CI configuration) but exits with an error when asked to run a component; it never reports readiness.
Signing packages
weft-pack generate-key ./keys
weft-pack sign ./examples/org.weft.demo.counter --key ./keys/weft-sign.key
weft-pack verify ./examples/org.weft.demo.counter --key ./keys/weft-sign.pub
NixOS VM (requires Nix with flakes)
Build the VM image:
bash infra/vm/build.sh
Run with QEMU:
bash infra/vm/run.sh
See infra/nixos/weft-packages.nix for the package derivations. Its outputHashes hold one hash per git source in Cargo.lock (the Servo and Stylo forks); every package vendors them, so they must match the pinned revisions for any Nix package to build.
Logging in on tty1 starts the user target weft-session.target, which binds graphical-session.target (that one refuses a manual start) and wants weft-compositor, then weft-servo-shell with the system UI page its package installs, and weft-appd with the runtime, app shell, file portal and mount helper packages. Run as a systemd user service, the compositor publishes its socket as WAYLAND_DISPLAY to the user manager before reporting readiness, so the units after it, and the app shells weft-appd starts, connect to it. The units in infra/systemd/ are the same session for an installation outside Nix, as user units under weft-session.target. The image is not yet bootable to a working desktop:
- the Servo source hash in
outputHashesis still a placeholder, so no package builds until a Nix build reports it; - the shell packages do not enable
servo-embed, soweft-servo-shellandweft-app-shellexit at startup; - the VM runs without graphics (
virtualisation.graphics = falseininfra/nixos/configuration.nix, for a serial console), so the session's output is not shown; WEFT_MOUNT_HELPERnames the mount helper's store path, which cannot be setuid, so verified images cannot be mounted; a setuid wrapper would be needed.
CI
Jobs run on every push to main and on pull requests:
portable— fmt, clippy, tests for the crates that need no Linux system libraries (excludes Wayland crates)linux-only— clippy and tests forweft-compositor,weft-servo-shell,weft-app-shellruntime-grants— clippy and a build ofweft-runtimewithwasmtime-runtime,net-fetch, thentests/runtime/check_grants.py, which runs a probe component under each kind of capability grantservo-embed-linux—cargo check --features servo-embedfor servo-shell and app-shellcontributor-toolkit—python .agents/scripts/verify.py toolkit