bump RTD to py39 #2730
bump RTD to py39 #2730
Conversation
|
"Problem in your project's configuration. Invalid "python.version": expected one of (2, 2.7, 3, 3.5, 3.6, 3.7, 3.8, pypy3.5), got 3.9" 🤦 |
|
https://docs.readthedocs.io/en/stable/config-file/v2.html#build-beta-specification apparently we were using an old-style (deprecated) version specification |
| m103_loc = mast.Mast.resolve_object("M103") | ||
| print(m103_loc) | ||
| assert m103_loc.separation(SkyCoord("23.34086 60.658", unit='deg')).value == 0 | ||
| assert round(m103_loc.separation(SkyCoord("23.34086 60.658", unit='deg')).value, 10) == 0 |
There was a problem hiding this comment.
Why was this change included in this RTD config PR?
| python: "mambaforge-4.10" | ||
|
|
||
| conda: | ||
| environment: docs/rtd_environment.yaml |
There was a problem hiding this comment.
Would be nice to not rely on any conda stuff for CI. Do you plan to have this temporarily, or this is now our solution for docs builds?
There was a problem hiding this comment.
Great question. I did this because it seemed the most direct way to get RTD to build again. I'm totally happy to return to a non-conda approach. I was unable to figure one out - graphviz specifically caused problems.
There was a problem hiding this comment.
And did you find out why it stopped working the first place? As grapviz wasn't really causing issues recently
There was a problem hiding this comment.
This error started appearing spontaneously:
https://readthedocs.org/projects/astroquery/builds/20444648/
"Could not import extension sphinx.builders.linkcheck (exception: urllib3 v2.0 only supports OpenSSL 1.1.1+, currently the 'ssl' module is compiled with OpenSSL 1.0.2n 7 Dec 2017. See: urllib3/urllib3#2168)"
My best read was that there is a problem caused by using old python3.8. We then had to upgrade to the new RTD configuration system, and then that caused problems with graphviz until I switched to the astropy conda-based version.
There was a problem hiding this comment.
yeap, but graphviz can be installed into the image directly without adding conda into the picture. #2732 should fix it.
3 participants
Attempt to fix #2729 outstanding CI failure