fast figure-updates in jupyter-notebooks can cause glitches

[raphaelquast]

Sometimes images are not updated fast enough so that artefacts of animated artists remain after blitting...

• this seems to be a general issue with blitting in nbbagg, webagg and ipympl
  check progress of: Stale information drawn by nbAgg matplotlib/matplotlib#19116

[banesullivan]

I've been having some fun with EOmaps and was trying to use the ipympl MPL backend and see this for an example from the docs. I'm curious if the difference I'm seeing here is due to this?

from eomaps import Maps
m = Maps(Maps.CRS.PlateCarree(central_longitude=-60))
m.add_feature.preset.ocean()

m_i = m.new_inset_map(xy=(5, 45), radius=10,
                      plot_position=(.3, .5), plot_size=.7,
                      boundary=dict(ec="r", lw=4),
                      indicate_extent=dict(fc=(1,0,0,.5),
                                           ec="r", lw=1)
                      )
m_i.add_indicator_line(m, c="r")

m_i.add_feature.preset.coastline()
m_i.add_feature.preset.countries()
m_i.add_feature.preset.ocean()
m_i.add_feature.cultural.urban_areas(fc="r", scale=10)
m_i.add_feature.physical.rivers_europe(ec="b", lw=0.25,
                                       fc="none", scale=10)
m_i.add_feature.physical.lakes_europe(fc="b", scale=10)

ipympl	inline

[raphaelquast]

@banesullivan Woha, this should not happen... It is in fact a issue in EOmaps itself and not related to blitting problems in ipympl. (I'll track it in #214) ...Thanks for reporting!

The problem with blitting really only concerns very fast screen updates (e.g. callbacks that draw on mouse-movement).
There seems to be a race-condition somewhere in the backend that causes a mixup in the order of received
diff-images... and this in turn causes plot-artefacts from previous diff-images to remain on screen.