Skip to content
This repository was archived by the owner on Nov 14, 2026. It is now read-only.

Latest commit

 

History

History
30 lines (21 loc) · 4.27 KB

README.md

File metadata and controls

30 lines (21 loc) · 4.27 KB

@systemfsoftware/effect-e2e-cli

The command line for @systemfsoftware/effect-e2e.

Install

Add it as a devDependency. It is a workspace tarball of the effect-e2e flake (packages.<system>.effect-e2e-cli), taken from nix.systemfsoftware.com through a flake input. It ships the effect-e2e bin and has no module entry, so nothing imports it. It requires Node >=24 (its engines field).

Usage

effect-e2e init --dir <suite> --engine web --gateway openrouter scaffolds a suite: e2e.config.ts, vitest.config.ts, runner.ts, global-setup.ts, a deterministic example and an agent example under tests/, a test:e2e script and a .e2e/ ignore entry. It asks for any choice not given as a flag, refuses to replace an existing file without --force, and installs the suite's dependencies unless --no-install is given.

The example tests init writes are titled in the Should_<Behaviour>_When_<Condition> form that @systemfsoftware/oxlint-config-recommended requires. init writes no lint configuration, so a suite linted with that preset declares the end-to-end role in its lint config: settings: { '@systemfsoftware/oxlint-plugin-test-discipline': { role: 'e2e' } }.

effect-e2e run --dir <suite> [files...] runs the suite under Vitest in the same process and prints one report: the one the run's global setup merged into the output directory (.e2e/ unless the config or --output moves it); without --dir it uses the nearest directory holding e2e.config.ts. --reporter picks list (the default), json, junit (writes junit.xml) or markdown (writes summary.md). The process exits with the report's code: 0 passed, 1 failed, 3 blocked, 4 error, 130 interrupted, and 2 when nothing reached a report: a config the runner refuses, a missing suite root, conflicting flags, a vitest.config.ts without the effect-e2e global setup, or a Vitest load error. A refused config never reads an older report.

Flag Effect
--target <name>, --agent <name> Run one target; make the named agent the default
--tag <tag> Run only tests carrying the tag (the suite declares its tags)
--no-cache, --strict-cache Cache mode off or read-only
--workers <n> Vitest's maxWorkers; overrides the config's workers
--retries <n> Retries per failed test; overrides the config's retries
--timeout <ms> Whole-test deadline and session budget
--output <dir> Where the report and artifacts go, relative to the suite
--trace off|on|retain-on-failure When the trace page is kept
--ai-trace Write ai-trace.json, the masked model calls, into each failed test's result directory, whatever --trace says
--last-failed Rerun what the previous report in this run's output directory did not pass or skip; a green previous run, or none, starts nothing and exits 0

Apache-2.0. Part of systemfsoftware.