This prevents referring to a generic type without filling in its generic
type parameters.
The FixtureDef typing might need some more refining in the future.
While working on improving the documentation of the
`pytest_runtest_setup` hook, I came up with this text:
> Called to perform the setup phase of the test item.
>
> The default implementation runs ``setup()`` on item and all of its
> parents (which haven't been setup yet). This includes obtaining the
> values of fixtures required by the item (which haven't been obtained
> yet).
But upon closer inspection I noticed this line at the start of
`SetupState.prepare` (which is what does the actual work for
`pytest_runtest_setup`):
self._teardown_towards(needed_collectors)
which implies that the setup phase of one item might trigger teardowns
of *previous* items. This complicates the simple explanation. It also
seems like a completely undesirable thing to do, because it breaks
isolation between tests -- e.g. a failed teardown of one item shouldn't
cause the failure of some other items just because it happens to run
after it.
So the first thing I tried was to remove that line and see if anything
breaks -- nothing did. At least pytest's own test suite runs fine. So
maybe it's just dead code?
_pytest.timing is an indirection to 'time' functions, which pytest production
code should use instead of 'time' directly.
'mock_timing' is a new fixture which then mocks those functions, allowing us
to write time-reliable tests which run instantly and are not flaky.
This was triggered by recent flaky junitxml tests on Windows related to timing
issues.
Flushing on every write is somewhat expensive.
Rely on line buffering instead (if line buffering for stdout is
disabled, there must be some reason...), and add explicit flushes when
not outputting lines.
This is how regular `print()` e.g. work so should be familiar.
TestDurations tests the `--durations=N` functionality which reports N
slowest tests, with durations <= 0.005s not shown by default.
The test relies on real time.sleep() (in addition to the code which uses
time.perf_counter()) which makes it flaky and inconsistent between
platforms.
Instead of trying to tweak it more, make it use fake time instead. The
way it is done is a little hacky but seems to work.
This changes the link anchors in "reference.html", from e.g.
`reference.html#pytest-current-test` to
`reference.html#envvar-PYTEST_CURRENT_TEST`, but I think that is OK, and
not worth adding labels for the old anchors.
Fix mypy errors:
src/_pytest/runner.py:36: error: "addoption" of "OptionGroup" does not return a value [func-returns-value]
src/_pytest/helpconfig.py:64: error: "addoption" of "OptionGroup" does not return a value [func-returns-value]
src/_pytest/terminal.py:67: error: "_addoption" of "OptionGroup" does not return a value [func-returns-value]
src/_pytest/terminal.py:75: error: "_addoption" of "OptionGroup" does not return a value [func-returns-value]
This allows for e.g. Jedi to infer types (it checks the name).
It was only used to support Python 3.5.0/3.5.1, where this is is not
available in the `typing` module.
Ref: https://github.com/davidhalter/jedi/issues/1472
Uses `TYPE_CHECKING = False` in `_pytest.outcomes` to avoid having to
work around circular import.
