| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
Sorry, something went wrong.
|
|
||
|
|
||
| @pytest.mark.skipif(sys.platform == 'darwin', | ||
| reason="Fails on darwin") |
There was a problem hiding this comment.
This is related to the svg to png conversion. The resulting svg file is identical to a local test, but the png is slightly different.
Sorry, something went wrong.
Sorry, something went wrong.
There was a problem hiding this comment.
I think this will be fixed by one of the last commits in #23559 which adds quotes to the font name in the expected svg.
I think they were always mal-formed (missing quotes on the family name), but inkscape (probably actually librsvg) is getting crankier about it. At some point we fixed the bug of not putting in quotes so the rest image is correct and the expected image is wrong 🤦🏻 .
Sorry, something went wrong.
There was a problem hiding this comment.
I think this will be fixed by one of the last commits in #23559 which adds quotes to the font name in the expected svg.
Seems like that didn't help. Should I simply go back to skipping (or possible better xfail-ing if that can be done on a platform-specific basis)?
Sorry, something went wrong.
There was a problem hiding this comment.
xfail for now and I'll look into if I can reproduce this on a local mac.
Sorry, something went wrong.
| print('SUCCESS', flush=True) | ||
|
|
||
|
|
||
| @pytest.mark.skipif(sys.platform == 'darwin', reason="Does not work on darwin") |
There was a problem hiding this comment.
Possibly the wrong signal is sent on OSX?
Sorry, something went wrong.
| print('SUCCESS', flush=True) | ||
|
|
||
|
|
||
| @pytest.mark.xfail(sys.platform == "darwin", reason='Fails on OSX') |
There was a problem hiding this comment.
I don't think this should fail on darwin? It may on CI, but I think we should put that somewhere else if possible... Especially because this will fail me locally.
Running on darwin I get:
lib/matplotlib/tests/test_backend_qt.py::test_sigint[show-kwargs0] PASSED [ 5%] lib/matplotlib/tests/test_backend_qt.py::test_sigint[pause-kwargs1] PASSED [ 7%]
Sorry, something went wrong.
There was a problem hiding this comment.
I have pyqt5, version 5.15.4 installed in this env if that is any help.
Sorry, something went wrong.
There was a problem hiding this comment.
You can check for 'CI' in os.environ if you think it's going to be CI-specific.
Sorry, something went wrong.
There was a problem hiding this comment.
Interesting, I thought I saw these locally as well. I'll try to check tomorrow when I get home.
Sorry, something went wrong.
There was a problem hiding this comment.
I think it started failing with qt6. The current CI uses qt5 and there it passes.
Sorry, something went wrong.
There was a problem hiding this comment.
Now it XPASS:es, but it is not clear which qt-version is used.
XPASS lib/matplotlib/tests/test_backend_qt.py::test_sigint[pause-kwargs1] Fails on OSX
XPASS lib/matplotlib/tests/test_backends_interactive.py::test_interactive_thread_safety[{'MPLBACKEND': 'qtagg', 'QT_API': 'PySide2'}]
XPASS lib/matplotlib/tests/test_backends_interactive.py::test_interactive_thread_safety[{'MPLBACKEND': 'qtcairo', 'QT_API': 'PyQt5'}]
XPASS lib/matplotlib/tests/test_backends_interactive.py::test_interactive_thread_safety[{'MPLBACKEND': 'qtcairo', 'QT_API': 'PyQt6'}]
XPASS lib/matplotlib/tests/test_backends_interactive.py::test_interactive_thread_safety[{'MPLBACKEND': 'qtcairo', 'QT_API': 'PySide6'}]
XPASS lib/matplotlib/tests/test_backends_interactive.py::test_interactive_thread_safety[{'MPLBACKEND': 'qtcairo', 'QT_API': 'PySide2'}]
I'll remove that line in the web interface and see what happens (so please squash merge...).
Sorry, something went wrong.
There was a problem hiding this comment.
There seems to be some xfail that can be removed in test_interactive_thread_safety as well, but I cannot really figure out which...
Sorry, something went wrong.
|
So it seems to be possible to build wxPython on OSX, but quite time consuming (about an hour). I guess it really doesn't make sense to even try to install it (still a 73 MB download) for the OSX tests. Will sort it out properly from a computer with a recent git (which this one isn't). |
Sorry, something went wrong.
|
Why are we building wxPython now; there should be wheels? |
Sorry, something went wrong.
|
There are no wheels for OSX. Here, I was basically just trying it out, but bad idea and better to skip even trying to install wx on OSX. (Unless we want to provide our own OSX wheel, but I have no idea how to do that. Should open an issue at them, asking for OSX wheels though since it seems to, at least, build without problems.) |
Sorry, something went wrong.
|
I moved this to 3.7 so we do not backport it. Lets keep going with the current CI on the bug-fix branch (if anything we need to freeze the packages more aggressively). |
Sorry, something went wrong.
|
It seems like despite it looking to be wheels for wxPython on OSX, the following happens. With: # But wheels for OSX
if [[ "macOS" == 'macOS' ]]; then
python -mpip install --upgrade wxPython &&
python -c 'import wx' &&
echo 'wxPython is available' ||
echo 'wxPython is not available'
fi
so plain "pip install wxPython" Collecting wxPython
Downloading wxPython-4.2.0.tar.gz (71.0 MB)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 71.0/71.0 MB 21.7 MB/s eta 0:00:00
Preparing metadata (setup.py): started
Preparing metadata (setup.py): finished with status 'error'
error: subprocess-exited-with-error
× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [8 lines of output]
Traceback (most recent call last):
File "<string>", line 2, in <module>
File "<pip-setuptools-caller>", line 34, in <module>
File "/private/var/folders/24/8k48jl6d249_n_qfxwsl6xvm0000gn/T/pip-install-h25qkvuc/wxpython_a713a81ec5204ad6bcbe911a560b6c3e/setup.py", line 27, in <module>
from buildtools.config import Config, msg, opj, runcmd, canGetSOName, getSOName
File "/private/var/folders/24/8k48jl6d249_n_qfxwsl6xvm0000gn/T/pip-install-h25qkvuc/wxpython_a713a81ec5204ad6bcbe911a560b6c3e/buildtools/config.py", line 30, in <module>
from attrdict import AttrDict
ModuleNotFoundError: No module named 'attrdict'
[end of output]
note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed
× Encountered error while generating package metadata.
╰─> See above for output.
note: This is an issue with the package mentioned above, not pip.
hint: See above for details.
wxPython is not available
|
Sorry, something went wrong.
|
I do not fully get why wxPython-4.2.0-cp38-cp38-macosx_11_0_universal2.whl is not deemed appropriate for 11.6.8 and Python 3.8.13 though. |
Sorry, something went wrong.
|
The latest version has GTK3 installed for OSX, but gives the following failing tests: FAILED lib/matplotlib/tests/test_backends_interactive.py::test_interactive_backend[toolbar2-{'MPLBACKEND': 'gtk3cairo'}]
FAILED lib/matplotlib/tests/test_backends_interactive.py::test_interactive_backend[toolmanager-{'MPLBACKEND': 'gtk3cairo'}]
FAILED lib/matplotlib/tests/test_backends_interactive.py::test_interactive_backend[toolbar2-{'MPLBACKEND': 'gtk3agg'}]
FAILED lib/matplotlib/tests/test_backends_interactive.py::test_interactive_thread_safety[{'MPLBACKEND': 'gtk3agg'}]
FAILED lib/matplotlib/tests/test_backends_interactive.py::test_interactive_backend[toolmanager-{'MPLBACKEND': 'gtk3agg'}]
FAILED lib/matplotlib/tests/test_backends_interactive.py::test_figure_leak_20490[time_mem0-{'MPLBACKEND': 'gtk3cairo'}]
FAILED lib/matplotlib/tests/test_backends_interactive.py::test_figure_leak_20490[time_mem1-{'MPLBACKEND': 'gtk3agg'}]
FAILED lib/matplotlib/tests/test_backends_interactive.py::test_figure_leak_20490[time_mem0-{'MPLBACKEND': 'gtk3agg'}]
FAILED lib/matplotlib/tests/test_backends_interactive.py::test_blitting_events[{'MPLBACKEND': 'gtk3agg'}]
FAILED lib/matplotlib/tests/test_backends_interactive.py::test_figure_leak_20490[time_mem1-{'MPLBACKEND': 'gtk3cairo'}]
Not much info in the transcript. |
Sorry, something went wrong.
|
Moved to 3.8, but if this is passing we can obviously merge it. |
Sorry, something went wrong.
|
Needs a rebase, as the logs are gone, though I may have looked before and found nothing else to add. |
Sorry, something went wrong.
|
It looks like there's a Traceback in stderr, but it's too long to see. Might have to run with -vvv to get pytest to print the whole thing? |
Sorry, something went wrong.
|
@oscargus are you going to push this through? |
Sorry, something went wrong.
|
I believe this was effectively replaced by #27723. |
Sorry, something went wrong.
| Back | FazBrowse Home | New Git URL |
PR Summary
#23584 (comment)
Also, inkscape and the extra test dependencies are added to OSX so the number of skipped tests is drastically decreased.
ffmpeg can be installed, but is rather time consuming (lots of additional dependencies).
dvipng cannot be installed via brew (but basictex can and GhostScript is already installed).
PR Checklist
Tests and Styling
Documentation