Attempt to fix Matplotlib issue by resticting to 2.x#3175
Attempt to fix Matplotlib issue by resticting to 2.x#3175DPeterK merged 1 commit intoSciTools:masterfrom
Conversation
|
Hi @tomgreen66, thanks for submitting this PR and getting the Iris tests passing again! For us to be able to progress this PR we're going to need you to sign a CLA, for which there are instructions here (see the "Contributors" bullet point). Thanks! |
|
Hi - thanks for the pointer. Just completed CLA and sent it in. Thinking about it the main issue was in Cartopy so maybe I should submit the change to them as well? I assume the dependency analysis will work if set in Iris but seems neater to fix it in Cartopy as well. I would love to be able to submit a fix to make it work with Matplotlib 3 but this is a good first fix. |
Thanks! We'll try and get that processed our end soon.
Yes please! (And, indeed, thanks for having done so.) This is certainly a good starting point for getting the tests passing again, but evidently not where we'd like to be with supporting the latest matplotlib. From experience though, upgrading Iris or cartopy to support the latest matplotlib can be quite a lot of effort! |
|
🎉 Thanks very much @tomgreen66 - this should be really helpful for moving lots of Iris PRs forward! |
Matplotlib 3 was released a few days ago. Seems pulling in latest install from conda breaks a lot of functionality due to Matplotlib 3 not being supported. Cartopy seems to have the biggest issues but might be other issues elsewhere.
For now suggest restricting to 2.x version of Matplotlib.