Repository navigation
Release 2022.11.0 #287
Description
| Release version | 2022.11.0 |
| Planned release date | 2022-11-15 |
| Status | On-track |
| Release manager | @jrbourbeau |
Best effort
Try to close before the release but will not block the release
- Allow
bokeh=3distributed#5648 - Allow
bokeh=3dask#9659 - Fix
sort_valueswithTimestampdata dask#9642 - Avoid
pandasconstructors indask.dataframe.coredask#9570 - ...
Blocker
Issues that would cause us to block and postpone the release if not fixed
- ...
- ...
Activity
jrbourbeau commentedon Nov 10, 2022
Adding dask/dask#9570 to the "Best effort" list
jakirkham commentedon Nov 10, 2022
JFYI RAPIDS (US) will be on holiday tomorrow
jrbourbeau commentedon Nov 10, 2022
Thanks for the heads up @jakirkham. US-based folks at Coiled will be off too. In light of that, I'll propose we bump the release to next Monday so more folks will be around to help if any follow-up is needed for the release (this isn't usually needed, but does happen rarely). I also don't think there's anything so pressing sitting in
mainthat can't wait until next Monday.jrbourbeau commentedon Nov 14, 2022
I wanted to surface some thoughts on dask/distributed#5648 here dask/distributed#5648 (comment). I'm proposing we include that PR in the release today. With the changes in that PR, user code won't break when
bokeh=3is installed. There are a couple of plots with CSS/sizing issues (it's still not clear if this is an issue withdistributedorbokeh), but I'd rather have a couple of plots be off than have users running intoImportErrors. If folks have thoughts on this topic, I'd appreciate whatever feedback you have.jakirkham commentedon Nov 14, 2022
Would those changes still work with Bokeh 2 or does Bokeh 3 become the new minimum requirement?
jrbourbeau commentedon Nov 14, 2022
Yeah, everything is backwards compatible with
bokeh=2. The only dependency change we're making is removing thebokeh<3constraintjrbourbeau commentedon Nov 14, 2022
Status update: It looks like there's agreement around including dask/distributed#5648 in the release. The
bokeh=3.0.2release fixes some CI failures we seeing in dask/distributed#5648.bokeh=3.0.2is on PyPI but isn't yet onconda-forge(but is close conda-forge/bokeh-feedstock#85). I'll propose we push the release back one more day to givebokeh=3.0.2time to land on conda-forge. We can then merge dask/distributed#5648 / dask/dask#9659 and release.jrbourbeau commentedon Nov 15, 2022
Okay, all outstanding issues for this release have been addressed. I'm going to start pushing out the release in the next hour
jrbourbeau commentedon Nov 15, 2022
2022.11.0 is on PyPI and conda-forge PRs are open:
jakirkham commentedon Nov 16, 2022
Conda & PyPI packages are up.
Currently waiting on the Docker image update ( dask/dask-docker#262 ).
charlesbluca commentedon Nov 16, 2022
Docker images are updated with dask/dask-docker#263
jrbourbeau commentedon Nov 16, 2022
Thanks for you work on getting that across the finish line @charlesbluca!