fix(test runner): avoid internal error for step end without begin
Consider the following scenario:
- Test finishes and starts tearing down fixtures.
- Fixture teardown starts a step S and then times out.
- We declare the test finished (with timeout).
- Dispatcher shuts down the worker and spins a new one for a retry.
Additionally, it clears steps information for the test to be
ready for the new retry. Step S information is lost.
- Meanwhile, during worker teardown, the step S does
actually finish (usually with an error), and we send stepEnd for S.
- Dispatcher does not know what to do with step S end and
prints an internal error.
The fix is to ignore certain messages from the shutting down worker that failed.
When sharing a context between tests and using `'on-first-retry'` we
could end up with tracing still running in non-retried tests. That's
extra overhead without a reason.
Using a worker fixture forces a new worker. This might be unexpected
when part of the test file runs in one worker, and another runs
in another worker. Top-level use of worker fixtures is still fine.
- Uses some auto fixtures to set default options and instrumentation on BrowserType.
- Moves screenshot, trace and video to worker-scoped fixtures.
- Throws in page/context when used from beforeAll/afterAll.
- Plumbs around BrowserType to be accessible from Browser and BrowserContext.
Each hook gets its own test scope. This is not too useful for
object fixtures like `page` (although one can use a page in
`beforeAll` to save storage state), but much more useful for option
fixtures like `viewport`.
We used to sort based on workerHash, and that changes depending on
the exact worker fixtures list. Now we replace workerHash with
an ordinal when constructing the TestGroup list to preserve the
natural order.
- `TestResult.startTime`
- `Suite.location` is optional now
- `Test.status()` renamed to `Test.outcome()` to differentiate against a
`Test.expectedStatus` and `TestResult.status` of the different type.
This makes our suite structure the following:
```
Root(title='') > Project(title=projectName) > File(title=relativeFilePath) > ...suites > test
```
Removed `fullTitle()` because it is not used directly by anyone.
Default reporters now report each test as
```
[project-name] › relative/file/path.spec.ts:42:42 › suite subsuite test title
```
- `Location` with `file`, `line` and `column`.
- `fullTitle` does not include project name.
- `titlePath` method.
- All methods of `Reporter` are optional.
- Removed `Test.skipped` property that is superseeded by `Test.status()`.
- Replaced `Suite.findTest()` with `Suite.allTests()`.
- Removed `Test.suite` property.
We now have Suites and Tests. When running multiple projects the whole
suite is cloned for each project. Same happens for repeatEach.
This simplifies the reporters API, but there is still room for improvement.
JSON reporter continues to produce old json output.
This avoids problems with `toMatchSnapshot('../../dir/file.png')`
where we append this path to `snapshotDir` and end up in some random
place.
Also added a note to documentation.
This fixes two issues:
- Sudden worker process exit is properly accounted for.
- We can stop() workers willy-nilly, e.g. after reaching maxFailures.
Details:
- DonePayload does not send `reamining` anymore, and worker does not track it.
- Instead, `Dispatcher._runJob` track remaining tests and acts accordingly.
- Upon worker exit, we emulate a fatal error for all remaining tests.
Drive-by:
- Do not report onTestBegin after reaching maxFailures to avoid confusion.
Before, we did report onTestBegin, but not onTestEnd.
- List reporter aligned between "running" and "finished" state - it was
one character misplaced.
- Added a SIGINT test.
Stop wrapping/prepending error messages so that we do not loose the stack trace. For this, update a few manually thrown errors with better messages (usually including a file path).
Speed up locations by doing manual `sourceMapSupport.wrapCallSite()` for a single call site. Performance gain in the runner process with 100 files x 100 tests each:
- 25% on the fresh run without babel cache;
- 80% on the cached run where babel is almost instant.
Also some obvious cleanups around stack traces (removing unused code).
This fixes an issue where we incorrectly labeled and assigned ids for tests
that declared tests in require'd files or used test wrappers.
See new tests for examples.
- Source now lives at `src/test`.
- Former folio tests live at `tests/playwright-test`.
- We use `src/test/internal.ts` that exposes base test without
Playwright fixtures for most tests (to avoid modifications for now).
- Test types live in `types/testFoo.d.ts`.
- Stable test runner is installed to `tests/config/test-runner` during `npm install`.
- All deps including test-only are now listed in `package.json`.
Non-test deps must also be listed in `build_package.js` to get included.
