test (web, native) – Edge Python

test is a unit-test harness written in pure Edge Python, not a compiled plugin. Import it by bare name, both runtimes resolve it with no manifest. To pin a different version, import it by URL or through a packages.json alias, see Modules.

@fixture registers a factory under its function’s name. @test(description, *uses) registers a test and the fixtures to inject by keyword. raises(ExcType) is a context manager asserting the block raises, and it accepts a class or a tuple of classes. run() executes every registered test, prints a PASS -, FAIL -, or ERROR line per test plus a summary, and raises SystemExit(1) if anything failed, else SystemExit(0). Fixtures are flat and built fresh per test, with no scopes, autouse, or parametrization. One module-level registry holds every test, so run() executes all of them.

from test import fixture, test, raises, run @fixture def user(): return {"name": "Ana"} @test("user has a name", "user") def test_name(user): assert user["name"] == "Ana" @test("division by zero raises") def test_div(): with raises(ZeroDivisionError): 1 / 0 run()

Output · expected


PASS - user has a name
PASS - division by zero raises
2 passed, 0 failed

Fixtures are built fresh per test, so one test cannot leak state into another:

from test import fixture, test, run @fixture def items(): return [] @test("first sees an empty list", "items") def test_first(items): assert items == [] @test("second also sees an empty list", "items") def test_second(items): items.append(1) assert len(items) == 1 run()

Output · expected


PASS - first sees an empty list
PASS - second also sees an empty list
2 passed, 0 failed

A failed assertion prints a FAIL - line with the error, and run() exits nonzero:

from test import test, run @test("two plus two") def test_add(): assert 2 + 2 == 4 @test("this one fails") def test_fail(): assert "a" in "zzz", "letter missing" run()

Output · expected


PASS - two plus two
FAIL - this one fails (AssertionError: letter missing)
1 passed, 1 failed

edge test discovers *_test.py files and drives run() for you, reading the verdict from the exit code.