Xila

Appearance

Validate a Core change ​

Run the checks that exercise the behavior and targets your change affects. Unless stated otherwise, run commands from Core's root inside its configured Devbox shell.

Before you begin ​

Complete development setup and install any extra prerequisites for your selected tests. Cargo-make initializes font generation before its tasks.

Check formatting ​

Use read-only checks before review:

bash
cargo make format-rust -- --check
cargo make format-toml --check
npm exec prettier -- --check "**/*.json"

Use cargo make format to apply the formatters while editing.

Check the affected targets ​

For host-facing changes:

bash
cargo make check-host
cargo make clippy-host
cargo make doc-host

For guest-library changes:

bash
cargo make check-guest
cargo make clippy-guest
cargo make doc-guest

Use check, clippy, and doc for both sets. These tasks check the Linux host and wasm32-unknown-unknown guest library; build the actual WASI guest separately when it changes.

Test WASM runtime ​

For graphics catalog, runtime, and alternate guest-word translation changes:

bash
cargo test -p graphics_bindgen --target x86_64-unknown-linux-gnu
cargo test -p wasm --features graphics --target x86_64-unknown-linux-gnu
cargo test -p wasm --no-default-features --features host,graphics,memory_64 --target x86_64-unknown-linux-gnu

Calculator is a wasm32 guest; its acceptance test is not run in the memory_64 configuration. Portable translation coverage and Miri commands are documented in executables/wasm/tests/translation/README.md. Add llvm-tools-preview to the pinned stable toolchain for coverage and miri to the pinned nightly for Miri.

Use cargo make test for workspace tests, and the command reference for the combined gates.

Browser behavior ​

Build Calculator before building the example:

bash
cargo build -p calculator --target wasm32-wasip1 --release
cargo make generate-fonts
(cd examples/wasm && RUSTUP_TOOLCHAIN="$XILA_NIGHTLY_TOOLCHAIN" trunk build)

Use trunk serve instead of trunk build and open the served URL to exercise rendering, input, 1 + 2 = 3, and window closing. For storage changes, check save/reload after a completed save as well; compilation alone does not verify persistence.

OPFS storage tests ​

Install wasm-bindgen-cli matching the resolved wasm-bindgen dependency and a Chrome/Chromium WebDriver matching your browser:

bash
CARGO_TARGET_WASM32_UNKNOWN_UNKNOWN_RUNNER=wasm-bindgen-test-runner \
CHROMEDRIVER=/path/to/chromedriver \
cargo test -p drivers_wasm --target wasm32-unknown-unknown --test opfs

For a non-default browser executable, set WASM_BINDGEN_TEST_WEBDRIVER_JSON to a JSON file containing {"goog:chromeOptions":{"binary":"/path/to/chromium"}}. These tests use real OPFS files and cover MBR/LittleFS save/reopen behavior.

Verify completion ​

Confirm your selected commands completed successfully and record the target, toolchain, and test scope. For native UI changes, also exercise rendering, input, and closing from a Wayland session.

Rust host tests, browser tests, and Miri do not establish embedded hardware correctness or firmware RAM/flash costs. Validate the relevant board/toolchain when the change affects hardware behavior.