Conversation
|
In view of #11350 I'd be interested to know what use case people have for |
dstansby
left a comment
There was a problem hiding this comment.
This looks good to me. For the test, instead of adding a new image (we try and avoid that) could you instead add a manual check that the axes limits change to include the arrow after the axes are autoscaled?
|
Looking forward to this fix ! thanks |
|
This will also need an API change note. There are likely users who are relying on this not auto-scaling and they need to be notified. |
|
@liujordan please don't give up ! I want your fix :p |
|
Taking over from @liujordan. This was previously a student project at University of Toronto Scarborough with Dr. Anya Tafliovich (@atafliovich). |
|
A major change I made was removing the |
anntzer
left a comment
There was a problem hiding this comment.
Please remove the unused image file and squash your commits.
|
Looks like something went wrong with the rebase? :( |
|
Trying to fix it now. |
|
I was unable to fix this branch, so restarted in another PR #15392 |
|
Closing for other pr |
can try mptcpanalyzer -dDEBUG "plot dss examples/client_2_filtered.pcapng 1 --dest=Server" "quit" There is an issue with autoscaling not working because FancyArrows are drawn as artists rather than patches see matplotlib/matplotlib#13788
PR Summary
When ax.arrow() is called, a FancyArrow object is being added to the current axes as an artist and not a patch. So when autoscale_view is called, which loops through all Patches and scales the view based on the minimum and maximum of those, the FancyArrow isn't being considered.
The solution is a two-step process. First, we have to add the arrow as a Patch, and second, we have to call autoscale_view() after adding it on the axes.
Resolves #12712
PR Checklist