After the most recent sync, I'm seeing the following warning when running any docker-compose command:
▶ docker-compose /usr/lib/python3.9/site-packages/numpy/ma/core.py:4466: SyntaxWarning: "is" with a literal. Did you mean "=="? if self.shape is ():
Note that this doesn't seem to impact the execution of the command, but it is slightly annoying to see the warning every time.
| R2189 numpy | |||
| R2189:9791b3af0cca Update to 1.21.2 | |||
You get that warning by running docker-compose alone, without argument?
I don't see any warning like that at least with what I tried --help, --version.
Our current version of numpy doesn´t fully support python 3.9. We are on the last version of numpy that supports python 2. We may have to update numpy to the latest version and package numpy-legacy separately for python2 applications that still need it.
You get that warning by running docker-compose alone, without argument?
@algent yes, running any docker-compose sub-command (or none at all) triggers this warning. It'd be good to get a confirmation from someone else that this is indeed reproducible for multiple people.
We may have to update numpy to the latest version and package numpy-legacy separately for python2 applications that still need it.
@joebonrichie if this wouldn't be too much of a burden, and T8207 doesn't seem closable anytime relatively soon, then that does sound like it might be a good solution. With that being said, I think it'd be a good idea to first check what python2 packages do actually depend on numpy, as if its a low amount (or 0), then that might make other solutions more viable.
printrun, mypaint, displaycal reply on py2 numpy. Printrun is on rc8 for py3 support, mypaint has an update available for py3 but is blocked by gimp and displaycal is seemingly never going to receive py3 support.
I think we should update numpy to latest version and package python2-numpy separately. Our current numpy version is relatively ancient by now and only officially supports python 3.7, and I doubt it'll work correctly when python 3.10 arrives. When the time comes we can deprecate python2-numpy when nothing relies on it.