MNT: Uncap dependencies #998
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #998 +/- ##
=======================================
Coverage 89.54% 89.54%
=======================================
Files 26 26
Lines 3395 3395
Branches 619 619
=======================================
Hits 3040 3040
Misses 207 207
Partials 148 148 ☔ View full report in Codecov by Sentry. |
There was a problem hiding this comment.
Wow thank you @effigies!
You just saved me a couple of hours of work at Brainhack next week!
I have looked at the reports (for next time, you can see them in the artifacts panel in CircleCI) and the bokeh side of things seems to be working well.
Thank you once again! I'm glad it was a simpler fix than expected.
That said, should we have a minimum version of bokeh? I imagine that bokeh moved from plot_width to width at some point? I'm happy to approve as is, since we've had this issue open for way too long.
|
I think a decent policy is to support two years back, but I reverted because the width and height options have been present the whole time I could see. I can set a minimum, or it can be a separate issue. Your call. |
|
If the Happy to leave it as is then. Thank you Chris! |
With ME-ICA/tedana#998 in a release, we can let bokeh and its dependencies upgrade to their latest versions and start running tests on Python 3.12.
I looked at #869 and #766 and I think I roughly understand the problems they were trying to fix. I think the jinja2 one was a short-lived incompatibility that just removing it should fix. The other one appeared to affect the reports. Running
make three-echoin an environment with the latest bokeh produced one complaint that resolved to bokeh/bokeh#12543, and otherwise just worked.Closes #976.