# justfile for oy-cli — autonomous OpenCode agent and deterministic repository workflows
#
# Run `just` or `just --list` to see available recipes.
#
# Quick start:
# just dev # fast checks (fmt + cargo check)
# just check # standard local checks plus the mdBook site
# just docs # build the mdBook site into book/
# just fix # auto-fix formatting and clippy lints, then check
# just run -- --help
#
# Requires: cargo, rustc >= 1.96, just, and mdbook. `mise install` provides
# them. Optional CI-parity recipes below require cargo-nextest and/or nightly Miri.
_default:
@just --list
# === Development checks ===
# Fast development check: format + cargo check (no recompilation across subcommands).
dev: _fmt-check
cargo check --locked
# Standard local check suite. Uses stable Cargo only so it works after `mise install`.
check: _fmt-check _clippy _test _rustdoc _book _help-smoke _installer-smoke
@echo "✓ local checks passed"
# Optional CI-parity suite. Requires cargo-nextest and nightly Miri.
ci: _fmt-check _clippy _nextest _miri _rustdoc _book _help-smoke
@echo "✓ CI-parity checks passed"
# Auto-format, apply clippy suggestions, update lockfile, then run the local suite.
fix: _fmt _clippy-fix
cargo update --workspace
@just check
# Validate the local LLM evaluation corpus without provider/model calls.
eval:
python3 scripts/eval_runner.py validate
# Build user/contributor documentation into book/.
docs: _book
# Run local prompt evaluations. Example: just eval-run --dry-run --task zuko-remote-pty-precision-audit
eval-run *args:
python3 scripts/eval_runner.py run {{args}}
# Compare two completed eval runs. Example: just eval-compare .tmp/eval/runs/base .tmp/eval/runs/new
eval-compare baseline candidate:
python3 scripts/eval_runner.py compare {{baseline}} {{candidate}}
# === Individual checks (available as standalone targets) ===
# Check formatting (no changes).
_fmt-check:
cargo fmt --check
# Apply formatting in-place.
_fmt:
cargo fmt
# Run clippy with deny-warnings.
_clippy:
cargo clippy --all-targets --locked -- -D warnings
# Auto-apply clippy fixes.
_clippy-fix:
cargo clippy --all-targets --locked --fix --allow-dirty --allow-staged
# Run all non-doc tests with stable Cargo, then run rustdoc examples/tests.
_test:
cargo test --all-targets --locked
cargo test --doc --locked
# Run all non-doc tests with nextest, matching CI's test runner.
_nextest:
cargo nextest run --all-targets --locked --profile ci
# Run focused smoke tests under Miri on nightly to catch undefined behavior.
_miri:
cargo +nightly miri test --locked miri_smoke
# Build docs with deny-warnings (no deps).
_rustdoc:
RUSTDOCFLAGS="-D warnings" cargo doc --no-deps --locked
# Build the mdBook user/contributor site.
_book:
rm -rf book
mdbook build
cp docs/install.sh book/install.sh
sh -n book/install.sh
test -f book/index.html
test -f book/getting-started.html
test -f book/reference.html
# Smoke-test the CLI help output.
_help-smoke:
cargo run --locked -- --help
cargo run --locked -- setup --help
cargo run --locked -- mcp --help
cargo run --locked -- run --help
cargo run --locked -- audit prepare --help
cargo run --locked -- audit finalize --help
cargo run --locked -- review prepare --help
cargo run --locked -- review finalize --help
cargo run --locked -- audit --help
cargo run --locked -- review --help
cargo run --locked -- enhance --help
cargo run --locked -- recover --help
cargo run --locked -- doctor --help
cargo run --locked -- upgrade --help
# Exercise installer sequencing and pins with a fake mise executable.
_installer-smoke:
sh scripts/test_install.sh
# === Release preparation ===
# Verify the crate can be packaged for publishing.
package:
cargo package --locked
# === Run the binary ===
# Run oy with arguments. Example: just run -- mcp
run *args:
cargo run --locked -- {{args}}