matplotlib 3.0 animation blitting bug?

This looks like a bug to me, but maybe someone knows better...

After updating to matplotlib 3.0, I get the following problem on animations
that use blitting: A horizontal bar appears at the bottom and a vertical bar
on the right of function animation plots. The attached screenshot shows the
problem for the "The Bayes update? example from the matplotlib animation
examples:
https://matplotlib.org/gallery/animation/bayes_update.html#sphx-glr-gallery-animation-bayes-update-py.

This same problem occurs on all animation files I?ve tried that use
blitting. None of them had the problem before I updated.

Interestingly enough, when I save the animation to an mpeg4 file, there is
no problem in the saved movie file.

This occurs on a system configured as follows:

macOS: 10.13.6
Python: 3.7.0 Anaconda distribution
matplotlib: 3.0.0
<http://matplotlib.1069221.n5.nabble.com/file/t1291/bayes_update_frame.png>

···

--
Sent from: http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html

I should have stated that I'm using the Qt5Agg backend, which seems to be
where the problem lies. The animation works without a problem for the TkAgg
backend.

···

--
Sent from: http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html

David,

Is this on a retnia display? If you re-size while the animation is running
does it fix it's self?

When saving to disk we re-render each frame 'from scratch' so I am not
surprised that it does not show up there.

Tom

···

On Tue, Oct 16, 2018 at 10:37 AM DJP <djpine at gmail.com> wrote:

I should have stated that I'm using the Qt5Agg backend, which seems to be
where the problem lies. The animation works without a problem for the TkAgg
backend.

--
Sent from:
http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
https://mail.python.org/mailman/listinfo/matplotlib-users

--
Thomas Caswell
tcaswell at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20181019/de73f5b8/attachment-0001.html>

That does help to localize the bug as Qt has support for high-dpi (to
support full resolution figures we have to first ask the desktop to not
give us effectively half-resolution and then internally double the rendered
resolution) where as tk does not. It looks like something is getting
confused about what parts of the buffer need to be updated / restored /
cleared during blitting.

Tom

···

On Fri, Oct 19, 2018 at 8:09 PM David J Pine <djpine at gmail.com> wrote:

Hi Tom,

Yes, it is on a retina display. I can try it on another display when I'm
at work tomorrow. However, resizing the window while running the animation
does not fix the problem.

David

On Fri, Oct 19, 2018 at 6:57 PM Thomas Caswell <tcaswell at gmail.com> wrote:

David,

Is this on a retnia display? If you re-size while the animation is
running does it fix it's self?

When saving to disk we re-render each frame 'from scratch' so I am not
surprised that it does not show up there.

Tom

On Tue, Oct 16, 2018 at 10:37 AM DJP <djpine at gmail.com> wrote:

I should have stated that I'm using the Qt5Agg backend, which seems to be
where the problem lies. The animation works without a problem for the
TkAgg
backend.

--
Sent from:
http://matplotlib.1069221.n5.nabble.com/matplotlib-users-f3.html
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
https://mail.python.org/mailman/listinfo/matplotlib-users

--
Thomas Caswell
tcaswell at gmail.com

--
Thomas Caswell
tcaswell at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20181019/db8770fe/attachment.html>