jac-testing

Write `test "name" { }` blocks alongside your code using plain assert statements, then run them with `jac test main.jac` or `jac test -d tests/`. Use `-t "name"` to run one test, `-x` to stop on first failure, and `-v` for verbose output.

Spawn the walker inside a test with `result = root spawn MyWalker();` and assert on `result.reports`, a list of everything the walker reported in order. You can also assert graph shape directly, such as `assert len([root-->]) == 2;`.

Tests in a file share one persisted root, and nodes hung off root persist to .jac/data between runs, so stale state corrupts later runs. Run `jac clean --all --force` before testing and write assertions that count only nodes the test itself created.

A `.test.jac` file is an annex that must pair with a same-basename module, so `people.test.jac` requires `people.jac` and runs via `jac test people.jac`. Also avoid the `test_` filename prefix, which collides with Python's test-module import machinery.

Use JacTestClient from jaclang.runtimelib.testing in pytest: create it with `JacTestClient.from_file("app.jac", base_path=str(tmp_path))`, register a user, and post to walker endpoints. The tmp_path base keeps each test's persisted graph isolated.

Do not hit a real model in tests; use MockLLM with canned outputs so results are deterministic. This pairs with the jac-by-llm skill, which documents MockLLM configuration for `by llm()` functions.