Greg Lucas

Could you please copy/paste the code here using triple backticks in a code block, and you can copy-paste images into the issue as well. It should be a **minimal** self-contained...

I believe @mathause has the solution. I assume that using these lats and lons shows the same issue you're seeing? ```python lats = np.linspace(-90, 90) lons = np.linspace(-180, 180) ```...

It should only **expand** the limits I believe. If you set the limits before the scatter, it does respect those limits as far as I can tell. Additionally, you can...

There is no longer a direct PROJ dependency check with v0.21 that was just released, so that error shouldn't happen anymore. For conda something like this should work to make...

I think grabbing it from the source is better whenever possible, that way we get any updates they have. It also looks like there are multiple "versions" of etopo, focusing...

I agree the current situation is less than ideal and there should be a way to handle this easier within Cartopy. I opened #1920 with a proposed fix for this.

There is a linked PR with a proposed fix that no one has commented on: #1926

I think this is running into the GEOS library linkings between the two libraries seen in #805. This only happens with the shapely wheels as far as I can tell....

There are a lot of open issues relating to Molleweide: https://github.com/SciTools/cartopy/search?q=Mollweide&type=issues Some of them indicate polygons are getting flipped inside out during the transformation and having issues near the poles....

It looks like this may be the same issues as https://github.com/SciTools/cartopy/issues/1333 and there does appear to be a PR to fix that: https://github.com/SciTools/cartopy/pull/1334 although I haven't looked at it.