Age | Commit message (Collapse) | Author |
|
Needs a couple of follow-up commits for workarounds, which hopefully
aren't needed anymore at a later point.
Fixes #6068
|
|
|
|
With https://github.com/python/mypy/pull/9614, mypy now tries to collect
all files in tests/, but it fails due to:
tests/end2end/conftest.py: error: Duplicate module named 'conftest' (also at 'tests/end2end/features/conftest.py')
tests/end2end/conftest.py: error: Are you missing an __init__.py? [misc]
We should probably add __init__.py files to tests/ at some point...
See #6059 and #5249
|
|
|
|
See https://github.com/PyCQA/pylint/issues/3760
|
|
|
|
|
|
We test Qt 5.15.2, but that's based on a newer Chromium version:
Qt 5.14.x -> Chromium 77
Qt 5.15.0 -> Chromium 80
Qt 5.15.2 -> Chromium 83
Some behavior changes between Chromium versions (see e.g. #5915), so it
makes sense to test both.
Picking 5.15.0 rather than .1 due to renderer process crashes in
5.15.1 (#5721). Doesn't matter much on the CI as we don't use WebRTC,
but if someone wants to downgrade with mkvenv.py for some reason, .0
will be more useful than .1.
|
|
Since the virtualenv v20.0.0 rewrite, those are hopefully not needed
anymore.
Closes #5373
|
|
|
|
|
|
See #4800
|
|
|
|
|
|
This reverts commit 1b13ee9429f43ae2bbd0e7e8e584e6470374f20a.
This worked locally, but apparently not on CI. Not worth the trouble.
|
|
|
|
|
|
|
|
|
|
Looks like mypy uses curses to get colors, so there's really nothing we can do
here on GitHub Actions.
|
|
- Set MYPY_FORCE_TERMINAL_WIDTH to get 180-column output (rather than 80), so
that error messages fit on one line.
- Set MYPY_FORCE_COLOR to get colored output
- Adjust problemmatcher regex accordingly
|
|
This should also color tox output, and it doesn't conflict with PYTEST_ADDOPTS
being used for coverage options.
|
|
|
|
Also reenables some tests which have been problematic on Travis/AppVeyor but
maybe work fine nowadays.
|
|
|
|
For some reason, pytest on GitHub doesn't display any colors.
|
|
|
|
With GitHub Actions, we can now easily install tox *and* eslint in the same
environment.
|
|
TERM being missing was the reason that eslint doesn't output colors unless
--color was given.
For some reason, forcing color via --color breaks GitHub Action's problem
matchers, while this approach works just fine.
|
|
|
|
Not done by default with "tox -e mypy" because it disables mypy's caching,
causing its runtime to go from <1s to ~15s:
https://github.com/python/mypy/issues/9041
See #1456
|
|
|
|
See #5368
|
|
For some reason, pip fails now when trying to "update" the existing source
clone...
|
|
- tox -e mkvenv and friends
- config.bind(key, None)
- :yank markdown
|
|
Fixes #4918
|
|
|
|
This was mainly needed because PyQt 5.14.1 is a manylinux2014 wheel and needs
an updated pip, see #5013
|
|
This is needed to upgrade pip so it can install PyQt 5.14.1 manylinux2014
wheels.
|
|
|
|
Also use Qt 5.14 for linters/etc. as it seems to work fine there.
|
|
Rationale: https://twitter.com/codewithanthony/status/1188958327318794241
|
|
|
|
|
|
We can't use specific type ignores yet as there is no pyflakes release with
this change in: https://github.com/PyCQA/pyflakes/pull/455
|
|
This reverts commit 548b5a17d72b42df997055c945c5e42e80b31af3.
This reverts commit fe32e00a4e1b5a7cfd49638501a072df243636c2.
Due to QTBUG-77364 (broken keyboard focus), we can't upgrade to qtbase 5.13.1.
However, with 5.13.0 (and QtWebEngine 5.13.1), we run into two more issues:
- Segfaults in QQuickWindowPrivate::updateDirtyNode() when logging into a
Google account (especially on macOS).
- Qt 5.13.0 with the old PyQt5 exit scheme segfaults on exit on Windows.
However, pyqt5_enable_new_onexit_scheme() was added in PyQt 5.13.1.
See https://bugreports.qt.io/browse/QTBUG-77364
See https://www.riverbankcomputing.com/news/pyqt-5131
Reopens #4918
|
|
|
|
This reverts commit 56a2c5050892ad93492e217a21df77b6cc0013a8.
Fixes #4918
|
|
|
|
See #4928
|