Right now LLMs are asked to do:
cargo build --workspace --all-features --all-targets --quiet
cargo test --workspace --all-features --quiet
cargo clippy --workspace --all-features --quiet -- -D warnings
cargo doc --workspace --all-features --quiet
cargo fmt --all --quiet
cargo publish --dry-run --quiet
Such that when users of the template use LLMs, they will avoid failing the PR quality checks.
This could however be folded into one command, e.g.
cargo build && cargo test && cargo clippy
This will make it faster to run, and use up less of the user's tokens, as it is one request instead of multiple.
In addition, only run the checks if editing code.
Right now LLMs are asked to do:
Such that when users of the template use LLMs, they will avoid failing the PR quality checks.
This could however be folded into one command, e.g.
cargo build && cargo test && cargo clippyThis will make it faster to run, and use up less of the user's tokens, as it is one request instead of multiple.
In addition, only run the checks if editing code.