API/FIX: mark figure as not stale on draw attempt #7451
Conversation
| renderer, self, dsu, self.suppressComposite) | ||
| try: | ||
| renderer.open_group('figure') | ||
| # prevent triggering call backs during the draw process |
There was a problem hiding this comment.
Doesn't this comment refer to the self._stale = True line that was removed? I guess it possibly does nothing since it shouldn't have that underscore, but the comment should go too, if it's not going to be left in.
There was a problem hiding this comment.
no, the self._stale should not have been removed. The reason it reaches in and uses the private variable is to not re-trigger the drawing logic.
There was a problem hiding this comment.
actually, on a bit further consideration that line does not matter. In the first pass at the stale system which did not propagate 'stale' state past an already stale state, however during the development (to support live updates in a plain python prompt) we made the 'staleness' always propogate all the way up (which may be related to performance issues in #6664) so that you could put a call back on the stale transition to draw idle (in IPython we use their 'post-execute' hook to kick the drawing)
tacaswell
modified the milestones:
2.0.1 (next bug fix release),
2.0 (style change major release)
Addresses jupyter/notebook#1881 where exceptions during the draw can cause an infinite loop.
tacaswell
force-pushed
the
fix_infinite_redraw_loop
branch
from
02c1167
to
4928219
Compare
|
pulled out of line comment so it does not get folded: In the first pass at the stale system which did not propagate 'stale' state past an already stale state, however during the development (to support live updates in a plain python prompt) we made the 'staleness' always propogate all the way up (which may be related to performance issues in #6664) so that you could put a call back on the stale transition to draw idle (in IPython we use their 'post-execute' hook to kick the drawing) |
tacaswell
modified the milestones:
2.0 (style change major release),
2.0.1 (next bug fix release)
|
Add doc note. I could be convinced that everything should behave this way (not just the figure). |
tacaswell
force-pushed
the
fix_infinite_redraw_loop
branch
from
b98257f
to
7a7480d
Compare
|
The docs failure looks like a python3 only thing that slipped through in the mandlebrot example. |
|
@tacaswell this looks fine to me. Is it ready to merge? |
|
Yes. This gets around a nasty bug of the infinite loop variety, |
|
merged v2.x to master |
Addresses jupyter/notebook#1881 where
exceptions during the draw can cause an infinite loop.