Phil Elson

Confirming that the ``transform_vectors`` algorithm is based on "derivatives of source coordinate system". This is documented at https://scitools.org.uk/cartopy/docs/latest/crs/index.html#cartopy.crs.CRS.transform_vectors, though there is no doubt that it could be made even more...

In putting together the summary above, I had the following example: ``` src = ccrs.PlateCarree() nx, ny = 6, 5 xs = np.linspace(*src.x_limits, nx+2)[1:-1] ys = np.linspace(*src.y_limits, ny+2)[1:-1] # u...

I've thought over this some more, and now wonder if cases 1 and 2 are actually the same thing, and we are looking at a bug that is acute for...

Couple of thoughts for myself: * should I deal with pchanged and prop observers * should I prevent certain properties (non-stylistic ones, such as data, transform etc.)

No need to do this for v2.1, but I'm definitely keen to do it asap.

Hi friends! 😄 Just to say that this can also be triggered by coordinates of "zero-terminated bytes" (non-unicode strings). Perhaps this is reasonable to not support it, since in the...

Came across this one today also.

Thanks for merging. I'm happy to see this in - it is a small change, but makes life so much more pleasant when you have an IDE and static-analysis at...

Closing this out, since there isn't really much value following the improvements in #1048. There were a few typos fixed etc., which I may try to get in later on....

Clearly we stalled on this one (totally understandable). I was hoping to at least review what is here already, but it seems that the fundamentals aren't yet working (e.g. ``_lookup``...