Wei Ji

> > Running `black` removes all trailing whitespaces for me. > > No, `black` doesn't remove the trailing whitespace in the docstrings of the file. Does `blackdoc` not remove the...

Thanks for chiming in @snowman2! I see you've opened an issue with making `scipy` optional in `rioxarray` at https://github.com/corteva/rioxarray/issues/413, that will really help slim things down since [SciPy is a...

Wouldn't this be reverting #490? But `test_plot_lines_with_arrows` passes, so does that mean just `GMT_IN` is okay?

Yeah, it looks complicated. The discussion looks to be focused on vectors, I'm not quite sure how grids are dealt with. ~~As an aside, it looks like the issue with...

> Using `GMT_IN` or `GMT_IN|GMT_IS_DUPLICATE` is the safest and the recommended way for PyGMT, but using `GMT_IN|GMT_IS_REFERENCE` can help GMT find more bugs. I believe we want a more stable...

Sorry, what I mean is that one of PyGMT's core features is being able to pass in PyData data structures (e.g. numpy, pandas, xarray) directly into GMT, that is, without...

That, said, we *were* using `GMT_IN` (i.e. `GMT_IS_DUPLICATE`?) before for vectors, so maybe it's not that big of an issue. But I'm not sure if it's wise to duplicate `xarray.DataArray`...

> * `GMT_IN|GMT_IS_DUPLICATE` is the same as `GMT_IN`. In this mode, GMT will make a copy of the input memory, and do all operations (read, write, or free) to the...

The examples get weirder and weirder, from black to gray to white! :rofl: |  |  |  | |:---:|:---:|:---:| | b64f3713017749a80d704ef9683dd140b95b6ea7 | 72f1c24383bf98b9951544400453f96d57ff1a3d | c54d54de25d512f12731acbd55cba3662a3f93e4 |

Related to https://github.com/GenericMappingTools/pygmt/issues/1442