Tests pass, dev-python/fs based tests are skippable. It appears it may
*never* have any patches merged upstream again, so at this point it may make
sense to just... stop testing fs support on newer pythons.
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>
It was python2 specific and long since treecleaned. For python 3 you
need pygobject, but pandas never made that switch and simply dropped Gtk
clipboard support entirely... back in 2019.
(The whole thing is actually a vendored copy of dev-python/pyperclip.)
postinst message only, no revbump.
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>
Now that tests run at all, I can confirm they pass newer pythons as
well. Can't test 3.14 because qtpy doesn't enable it yet, but at least
we are finally "modern".
Closes: https://bugs.gentoo.org/929800
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>
There is a random segfault in the testsuite. If I run
```
while FEATURES=test emerge -1 vitables; do :; done
```
after a few successful runs, it fails with
```
tests/test_vtgui.py::TestVTGui::test_dockWidget Fatal Python error: Segmentation fault
```
This does NOT fix the tests. It does, however, allow pytest to process
the entire test session and summarize as e.g.
> worker 'gw6' crashed while running 'tests/test_logger.py::TestLogger::test_write'
Bug: https://bugs.gentoo.org/892886
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>
Drastically updated version on account of it includes loads of patches
we were importing from Debian (?) and also actually works with numpy
2.x. In fact, its ::gentoo deps already require `>=numpy-2` therefore it
has so far been impossible to install it in working at all. As such, I
give no credence to the idea of keeping the old version around "in case
people encounter issues with the new release". Not this time, buddy!
In the process, it turns out a particular test is broken. It tries to
test locale support, which would be nice and all except there are some
raw *.ts files floating around in the repo root and no way to actually
compile or install them. To boot, the testcase string wants to hardcode
the es_ES locale. There is no upstream CI or anything.
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>
```
* SED: the following did not cause any changes
* sed -i -e 's:--cov=qtpy --cov-report=term-missing::' pytest.ini || die;
* no-op: -e s:--cov=qtpy --cov-report=term-missing::
```
In commit d27e3d368d, this package added
`-o addopts=` to the pytest invocation, wiping out the effects of this
line that was being sedded. Subsequently, the fact that the sed failed
to apply due to incidental refactoring, went unnoticed. The problematic
`--cov*` arguments are still there, and could be deleted, but is now
redundant.
Signed-off-by: Eli Schwartz <eschwartz@gentoo.org>