Issue with using wrong pyrcc
See original GitHub issueš Bug
I updated napari
to post- #1229 and napari
started failing when trying to build icons on Windows after importing napari:
https://github.com/napari/napari/blob/96e9722419a5b0adddbc14fe12c69a7d769f8758/napari/resources/build_icons.py#L178
Cause
This may be a non-issue for others, but became an issue for me. Basically, each of these binaries were called:
'pyrcc5.bat', 'pyrcc5', 'pyside2-rcc'
and my vanilla python virtual environment didnāt have pyrcc5.bat
, but my base anaconda install was on path (bottom of list), and was the only source of pyrcc5.bat
. This then failed, because it was found, but didnāt have pyqt5-sip
I believe. Rearranging so that pyrcc5
was tried first fixed the issue for me.
This is a weird bug because I happened to have an Anaconda install on my path that happened to have pyrcc5.bat
, and is kind of my fault. I donāt know if it makes sense for you to rearrange the binaries to try, or ensure itās coming from the same environment. At least this is an explanation for anyone else getting a similar error:
subprocess.CalledProcessError: Command '['pyrcc5.bat', '-o', 'XXX\\lib\\site-packages\\napari\\resources\\_qt_resources_0.3.1_pyqt5_5.14.0.py', 'XXX\\lib\\site-packages\\napari\\resources\\res.qrc']' returned non-zero exit status 1.
Issue Analytics
- State:
- Created 3 years ago
- Comments:23 (16 by maintainers)
@tlambert03 You were right, I think
python 3.7
outside conda was causing the issue. Here is what I did to fix it,C:\Users\ajn16\AppData\Roaming\Python
After that,
Thanks for figuring this out @tlambert03 and hope it is helpful for others who find this thread š
Best, Ajit.
I see⦠if pip doesnāt see
numpy>=1.10.0
from PyPI when it tries to install napari, it will likely overwrite whatever is already there. You might have better luck with conda, which is pretty good about compatibility with MKL. cheers