The previous test was better in that it used fakes to test all of the
real code paths. The problem with that is that it makes it impossible to
simplify the code with `isinstance` checks. So let's just simulate the
issue directly with a monkeypatch.
It currently fails with a TypeError, and was not updated since 2013 -
therefore it can be assumed that it is not important to support it.
```
____________________ ERROR collecting test_nested_marks.py _____________________
…/Vcs/pluggy/src/pluggy/hooks.py:286: in __call__
return self._hookexec(self, self.get_hookimpls(), kwargs)
…/Vcs/pluggy/src/pluggy/manager.py:93: in _hookexec
return self._inner_hookexec(hook, methods, kwargs)
…/Vcs/pluggy/src/pluggy/manager.py:337: in traced_hookexec
return outcome.get_result()
…/Vcs/pluggy/src/pluggy/manager.py:335: in <lambda>
outcome = _Result.from_call(lambda: oldcall(hook, hook_impls, kwargs))
…/Vcs/pluggy/src/pluggy/manager.py:87: in <lambda>
firstresult=hook.spec.opts.get("firstresult") if hook.spec else False,
…/Vcs/pytest/src/_pytest/python.py:235: in pytest_pycollect_makeitem
res = list(collector._genfunctions(name, obj))
…/Vcs/pytest/src/_pytest/python.py:404: in _genfunctions
self.ihook.pytest_generate_tests.call_extra(methods, dict(metafunc=metafunc))
…/Vcs/pluggy/src/pluggy/hooks.py:324: in call_extra
return self(**kwargs)
…/Vcs/pluggy/src/pluggy/hooks.py:286: in __call__
return self._hookexec(self, self.get_hookimpls(), kwargs)
…/Vcs/pluggy/src/pluggy/manager.py:93: in _hookexec
return self._inner_hookexec(hook, methods, kwargs)
…/Vcs/pluggy/src/pluggy/manager.py:337: in traced_hookexec
return outcome.get_result()
…/Vcs/pluggy/src/pluggy/manager.py:335: in <lambda>
outcome = _Result.from_call(lambda: oldcall(hook, hook_impls, kwargs))
…/Vcs/pluggy/src/pluggy/manager.py:87: in <lambda>
firstresult=hook.spec.opts.get("firstresult") if hook.spec else False,
…/Vcs/pytest/src/_pytest/python.py:130: in pytest_generate_tests
metafunc.parametrize(*marker.args, **marker.kwargs)
…/Vcs/pytest/src/_pytest/python.py:965: in parametrize
function_definition=self.definition,
…/Vcs/pytest/src/_pytest/mark/structures.py:111: in _for_parametrize
if len(param.values) != len(argnames):
E TypeError: object of type 'MarkDecorator' has no len()
!!!!!!!!!!!!!!!!!!!! Interrupted: 1 error during collection !!!!!!!!!!!!!!!!!!!!
```
This is a useful utility to abstract the caching property idiom.
It is in compat.py since eventually it will be replaced by
functools.cached_property.
Fixes#6131.
It is a little too obscure IMO, but the reason I want to drop it is that
type checking has no hope of understanding such dynamic constructs.
The warning argument wasn't used.
This fixes/removes the previous hack of re-trying with minimum width,
which fails short when it splits strings.
This inherits from `pprint.PrettyPrinter` to override `_format` in a
minimal way to always dispatch, regardless of the given width.
Code ref: 5c0c325453/Lib/pprint.py (L170-L178)
Harden one test where it is tested.
All tests testing this:
testing/acceptance_test.py:184(TestGeneralUsage::test_not_collectable_arguments)
testing/acceptance_test.py:373(TestGeneralUsage::test_direct_addressing_notfound)
testing/acceptance_test.py:403(TestGeneralUsage::test_issue134_report_error_when_collecting_member[test_fun.py::test_a])
testing/acceptance_test.py:420(TestGeneralUsage::test_report_all_failed_collections_initargs)
testing/test_config.py:1309(test_config_blocked_default_plugins[python])
(via https://github.com/blueyed/pytest/pull/88)
Works around:
_____ ERROR collecting testing/io/test_saferepr.py _____
src/_pytest/python.py:502: in _importtestmodule
mod = self.fspath.pyimport(ensuresyspath=importmode)
.venv38/lib/python3.8/site-packages/py/_path/local.py:701: in pyimport
__import__(modname)
<frozen importlib._bootstrap>:991: in _find_and_load
???
<frozen importlib._bootstrap>:975: in _find_and_load_unlocked
???
<frozen importlib._bootstrap>:671: in _load_unlocked
???
src/_pytest/assertion/rewrite.py:136: in exec_module
source_stat, co = _rewrite_test(fn, self.config)
src/_pytest/assertion/rewrite.py:288: in _rewrite_test
co = compile(tree, fn, "exec", dont_inherit=True)
E File "…/Vcs/pytest/testing/io/test_saferepr.py", line 45
E None()
E ^
E SyntaxError: 'NoneType' object is not callable; perhaps you missed a comma?
This causes INTERNALERRORs with pytest-django, which uses
`pytest.fail` (derived from `BaseException`) to prevent DB access, when
pytest then tries to e.g. display the `repr()` for a Django `QuerySet`
etc.
Ref: https://github.com/pytest-dev/pytest-django/pull/776
Follow-up to 946434c61 (#5924).
Before this patch the test would look like this:
{'env': {'sub...s wrapped'}}}} == {'env': {'sub...}}}, 'new': 1}
Omitting 1 identical items, use -vv to show
Right contains 1 more item:
{'new': 1}
Full diff:
{
'env': {'sub': {'long_a': 'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa',
- 'sub1': {'long_a': 'substring '
+ 'sub1': {'long_a': 'substring that gets wrapped'}}},
? +++++++++++++++++ ++++
+ 'new': 1,
- 'that '
- 'gets '
- 'wrapped'}}},
}
indicate current outcome/status with color of percentage indicator
Fix type annotation, refactor _write_progress_information_filling_space
Keep code in _get_main_color as similar as possible to how it was before
Write test
Make black-compliant
Fix error in newly introduced test_collecterror
Make tests more readable by using constants and f-strings
Remove accidentally added monkeypatch
Make Python 3.5-compatible, add changelog entry
Add newline at the end of changelog file
Remove the `--fulltrace` arg from the `Option` fixture used in several
tests, but not checked for. Only use it with `test_keyboard_interrupt`.
(removes 8 tests, coverage not affected)
Without this, the second time it tries to stop in a parametrized
function it raises instead:
`ValueError: --trace can't be used with a fixture named func!`
Implementation idea, test (and changelog tweaks) thanks to blueyed
Co-Authored-By: Ronny Pfannschmidt <opensource@ronnypfannschmidt.de>
Co-Authored-By: Daniel Hahler <git@thequod.de>
Failed for me due to different indent (?) - not reproducible:
> ???
E Failed: nomatch: 'E TypeError: write() argument must be str, not bytes'
…
E and: '> def mode(self):'
E and: 'E TypeError: write() argument must be str, not bytes'
…
E remains unmatched: 'E TypeError: write() argument must be str, not bytes'
In order to make the LiteralOutputChecker lazy initialization more
amenable to type checking, I changed it to match the scheme already used
in this file to lazy-initialize PytestDoctestRunner.
This is important when used with ``pytester``'s ``runpytest_inprocess``.
Since 07f20ccab `pytest testing/acceptance_test.py -k test_doctest_id`
would fail, since the second run would not consider the exception to be
an instance of `doctest.DocTestFailure` anymore, since the module was
re-imported, and use another failure message then in the short test
summary info (and in the report itself):
> FAILED test_doctest_id.txt::test_doctest_id.txt - doctest.DocTestFailure: <Do...
while it should be:
> FAILED test_doctest_id.txt::test_doctest_id.txt
This avoids mutating the original list to reflect on InvocationParams,
which is supposed to be an immutable snapshot of the state of pytest.main()
at the moment of invocation (see pytest-dev/pytest-xdist#478).
Massage text input for difflib when comparing pformat output of
different line lengths.
Also do not strip ndiff output on the left, which currently already
removes indenting for lines with no differences.
Before:
E AssertionError: assert ['version', '...version_info'] == ['version', '...version', ...]
E Right contains 3 more items, first extra item: ' '
E Full diff:
E - ['version', 'version_info', 'sys.version', 'sys.version_info']
E + ['version',
E + 'version_info',
E + 'sys.version',
E + 'sys.version_info',
E + ' ',
E + 'sys.version',
E + 'sys.version_info']
After:
E AssertionError: assert ['version', '...version_info'] == ['version', '...version', ...]
E Right contains 3 more items, first extra item: ' '
E Full diff:
E [
E 'version',
E 'version_info',
E 'sys.version',
E 'sys.version_info',
E + ' ',
E + 'sys.version',
E + 'sys.version_info',
E ]
The current idiom is to use:
assert re.match(pat, result.stdout.str())
Or
assert line in result.stdout.str()
But this does not really give good results when it fails.
Those new functions produce similar output to ther other match lines functions.
attrs 19.2 deprecated cmp in favor of the dataclass-ish eq/order duo.
This causes deprecation warnings that in turn break some of the cool new deep
object comparisons. Since we at attrs expected this to be a problem, it shipped
with helpers to write backward and forward compatible code.
This PR uses that and avoids changed to minimal versions.
Using Path().resolve() is better than py.path.realpath because
it resolves to the correct path/drive in case-insensitive file systems (#5792):
>>> from py.path import local
>>> from pathlib import Path
>>>
>>> local('d:\\projects').realpath()
local('d:\\projects')
>>> Path('d:\\projects').resolve()
WindowsPath('D:/projects')
Fix#5819
We find that the --pastebin option to pytest sometimes fails with "HTTP
Error 400: Bad Request". We're still investigating the exact cause of
these errors, but in the meantime, a failure to upload to the pastebin
service should probably not crash pytest and cause a test failure in the
continuous-integration.
This patch catches exceptions like HTTPError that may be thrown while
trying to communicate with the pastebin service, and reports them as a
"bad response", without crashing with a backtrace or failing the entire
test suite.
`argparse.HelpFormatter` looks at `$COLUMNS` only, falling back to a
default of 80.
`py.io.get_terminal_width()` is smarter there, and could even work
better with https://github.com/pytest-dev/py/pull/219.
This ensures to use a consistent value for formatting the ini values etc.
Previously it would say:
> assert '123456789012...901234567890A' == '1234567890123...901234567890B'"
This makes it look like the "3" might be different already.
This is clearer, and it is OK to have potentially one less char in the
right one:
> assert '123456789012...901234567890A' == '123456789012...901234567890B'"
Previously, the test failed when the directory was not present,
which could have been caused for example by invoking the tests
with PYTHONDONTWRITEBYTECODE=1.
Fixes https://github.com/pytest-dev/pytest/issues/5664
Map `BdbQuit` exception to `outcomes.Exit`.
This is necessary since we are not wrapping `pdb.set_trace` there, and
therefore our `do_quit` is not called.
This reverts commit dfe54cd82f.
The idea in the commit was to simplify the code by removing the check
and instead letting it TypeError which has the same effect.
However this type error is caught by mypy, and rather than ignoring the
error we think it's better and clearer to go back to the previous
explicit check.
These doctests were expected to fail, but they were failing because of a
silly bug (I forgot to replace "{expression}" with the actual expression
to be tested), not because of the thing they were meant to be testing.
Then I had to fix one of the testcases because it was actually matching:
>>> 3.0 #doctest: +NUMBER
2.99
The doctest is saying that the actual output should match to 2 decimal
places, i.e. within 0.01 -- which it is, so it passes. I changed the
expected output to 2.98 and now it doesn't match (as we expect).
Add a very lax mypy configuration, add it to tox -e linting, and
fix/ignore the few errors that come up. The idea is to get it running
before diving in too much.
This enables:
- Progressively adding type annotations and enabling more strict
options, which will improve the codebase (IMO).
- Annotating the public API in-line, and eventually exposing it to
library users who use type checkers (with a py.typed file).
Though, none of this is done yet.
Refs https://github.com/pytest-dev/pytest/issues/3342.
Add a very lax mypy configuration, add it to tox -e linting, and
fix/ignore the few errors that come up. The idea is to get it running
before diving in too much.
This enables:
- Progressively adding type annotations and enabling more strict
options, which will improve the codebase (IMO).
- Annotating the public API in-line, and eventually exposing it to
library users who use type checkers (with a py.typed file).
Though, none of this is done yet.
Refs https://github.com/pytest-dev/pytest/issues/3342.
When enabled, floating-point numbers only need to match as far as the
precision you have written in the expected doctest output. This avoids
false positives caused by limited floating-point precision, like this:
Expected:
0.233
Got:
0.23300000000000001
This is inspired by Sébastien Boisgérault's [numtest] but the
implementation is a bit different:
* This implementation edits the literals that are in the "got"
string (the actual output from the expression being tested), and then
proceeds to compare the strings literally. This is similar to pytest's
existing ALLOW_UNICODE and ALLOW_BYTES implementation.
* This implementation only compares floats against floats, not ints
against floats. That is, the following doctest will fail with pytest
whereas it would pass with numtest:
>>> math.py # doctest: +NUMBER
3
This behaviour should be less surprising (less false negatives) when
you enable NUMBER globally in pytest.ini.
Advantages of this implementation compared to numtest:
* Doesn't require `import numtest` at the top level of the file.
* Works with pytest (if you try to use pytest & numtest together, pytest
raises "TypeError: unbound method check_output() must be called with
NumTestOutputChecker instance as first argument (got
LiteralsOutputChecker instance instead)").
* Works with Python 3.
[numtest]: https://github.com/boisgera/numtest
These attributes can be used to access the unchanged arguments passed
to pytest.main().
The intention is to use these attributes to initialize workers in
the same manner as the master node is initialized in pytest-xdist.
Daniel Hahler
Ran Benita
Anthony Sottile
Bruno Oliveira
Gregory Lee
NNRepos
Tibor Arpas
MarcoGorelli
Florian Bruhin
Patrick Harmon
David Szotten
Joshua Storck
Kale Kundert
Michael Krebs
AnjoMan
Daniil Galiev
Yoav Caspi
Victor Maryama
Hynek Schlawack
Tomáš Chvátal
Andrzej Klajnert
Bruno Oliveira
Hugo
aklajnert
Michael Goerz
Christian Neumüller
Robert Holt
Thomas Grainger
dmitry.dygalo
Bruno Oliveira
Samuel Searles-Bryant
Semen Zhydenko
Kaiqi
Miro Hrončok
Ran Benita
Ronny Pfannschmidt
David Röthlisberger
Zac Hatfield-Dodds