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.
Changelogs:
http://mypy-lang.blogspot.com/2019/09/mypy-730-released.htmlhttp://mypy-lang.blogspot.com/2019/10/mypy-0740-released.html
New errors:
src/_pytest/recwarn.py:77: error: Missing return statement
src/_pytest/recwarn.py:185: error: "bool" is invalid as return type for "__exit__" that always returns False
src/_pytest/recwarn.py:185: note: Use "typing_extensions.Literal[False]" as the return type or change it to "None"
src/_pytest/recwarn.py:185: note: If return type of "__exit__" implies that it may return True, the context manager may swallow exceptions
src/_pytest/recwarn.py:185: error: Return type "bool" of "__exit__" incompatible with return type "None" in supertype "catch_warnings"
src/_pytest/recwarn.py:230: error: "bool" is invalid as return type for "__exit__" that always returns False
src/_pytest/recwarn.py:230: note: Use "typing_extensions.Literal[False]" as the return type or change it to "None"
src/_pytest/recwarn.py:230: note: If return type of "__exit__" implies that it may return True, the context manager may swallow exceptions
src/_pytest/recwarn.py:230: error: Return type "bool" of "__exit__" incompatible with return type "None" in supertype "catch_warnings"
The errors are due to this new error:
https://mypy.readthedocs.io/en/latest/error_code_list.html#check-the-return-type-of-exit-exit-return
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.
The typing module on these versions have these issues:
- `typing.Pattern` cannot appear in a Union since it is not considered a
class.
- `@overload` is not supported in runtime. (On the other hand, mypy
doesn't support putting it under `if False`, so we need some runtime
hack).
Refs #5751.
Bruno Oliveira
Daniel Hahler
Ran Benita
Anthony Sottile
Yoav Caspi
Philipp Loose
Victor Maryama
Victor Maryama
Hynek Schlawack
Bruno Oliveira
Andrzej Klajnert
aklajnert
Michael Goerz
Christian Neumüller
linchiwei123
Robert Holt
Ran Benita