FIX: contour sticky edges match dataLim#2414
Conversation
greglucas
left a comment
greglucas
left a comment
There was a problem hiding this comment.
Great job tracking this down! Thanks for the quick fix.
|
@rcomer have you not signed the new CLA yet? I can't find a way to refresh that check now either... |
I got a pass from signing it as I'm publicly a member of the Met Office GitHub organisation. But the checker seems broken. @trexfeathers any ideas? |
|
You can recheck using this URL: https://cla-assistant.io/check/SciTools/cartopy?pullRequest=2414 I have sometimes needed to try a couple of times when the service is patchy. I'm keen to find a good place to share this information. Any thoughts would be welcomed. This could include a note in each repository if we can't think of a good central place |
|
@trexfeathers Is there a way to add that in the "Details" section? It was just stalled previously, so if there were a way to re-run it from this page directly that would be ideal. But a link somewhere else might be useful, although I'm guaranteed to forget about where that link is I'm sure ;) |
I think @trexfeathers already had told me about it... |
Probably not, since at this point it hasn't run. All GitHub knows it should expect something with that name. I got the URL from the comments that are posted when people still need to sign, so CLAAssistant are providing the information whenever they can. It's my fault for making it a required check. This is the only place I can think of but that's very hidden at the point that a note would be needed: cartopy/.github/PULL_REQUEST_TEMPLATE.md Lines 31 to 35 in 11bb8d3 |
|
Regarding CLA reboot instructions, I have now written a wiki page here: |
3 participants
Rationale
Fixes #2413. The problem seems to be that we have a very small mismatch between the artist's sticky edges and the axes dataLim. Within Matplotlib's contour code, these are set equal.
I have tested locally with both mpl nightly and mpl 3.8.4. Note that matplotlib/matplotlib#28393 was included in mpl 3.9.1 so our CI should already use the new version.
Implications