Pier Giuseppe Fogli

I'm just chiming in to let interested people know that I've been working on this task during the last week or so. I managed to produce a working example on...

> There is some initial work in [70ef683](https://github.com/Alexander-Barth/NCDatasets.jl/commit/70ef6830497674461c33114940ae58d439c7f827) in the branch MPI. Is it new? I don't remember seeing it when I checked last week. Anyway, it's almost identical to...

Continuing here #1804. Although this behavior is not a `set_extent` bug, I would like to point out a possible solution. As pointed out by @pelson the default rectangular axes has...

> @pgf your situation is similar to the one I wrote about for #1367 above. The interpolator in trace.pyx doesn't sample enough points to determine the extent in projected space...

> I disagree---your work and #1367 reveal a bug in `set_extent` that may be solved by increasing the number of points sampled by trace.pyx. What you've presented is a workaround...

@kdpenner @dopplershift I've already experimented with the `threshold` value with no success. I've got some improvement reducing the value at: https://github.com/SciTools/cartopy/blob/d12c86ca228365c61d5fb8b7121b49201f19fea8/lib/cartopy/trace.pyx#L449 Reducing from `1.0e-6` to `1.0e-8` results in some more...

@kdpenner Just to clarify, for the Robinson projection I've changed: https://github.com/SciTools/cartopy/blob/7c75f5ab6a397bca3197774b063820b6aabe87f9/lib/cartopy/crs.py#L1874-L1877 to: ```python self._threshold = 1e4 @property def threshold(self): return self._threshold ``` and then modified #1367 with: ```python print(ax.projection._threshold) ax.projection._threshold=1.e-2...

> and I haven't changed `threshold` there. The problem was probably fixed by changing the extent geometry to a `Polygon` from a `LineString`. @kdpenner This makes sense. Indeed sounds like...