| FazBrowse GitHub Viewer | Trending | | Home |
| Tools: [Download Repo ZIP] [Original HTTPS Page] |
|
Can the issue at #6789 also be triggered by e.g. patches at nonfinite positions? It feels like the logic of ignorining nonfinite positions should go into tight_layout itself instead (having thought approximately for 5s about the issue, so I may completely wrong). |
Sorry, something went wrong.
| """ | ||
| self.set_x(xy[0]) | ||
| self.set_y(xy[1]) | ||
| if not np.all(np.isfinite(xy)): |
There was a problem hiding this comment.
This will act as a latch and this text will awlays be invisible
Sorry, something went wrong.
|
I guess you could have a scale that brings infinity to a finite value? :-) (although that would probably crash other parts of mpl) |
Sorry, something went wrong.
|
Closing this, since the eventual fix will be different. If anyone else wants to pick this up feel free. |
Sorry, something went wrong.
| Back | FazBrowse Home | New Git URL |
This fixes #6789. The problem was there was some text at +/- inf locations, which mean that when the tight_layout calculation was being done an inf was being passed through.
I think it's reasonable to set text with co-ordinates that aren't finite to not visible as I presume it'll never be seen, although this currently just hides the problem without a warning or error. Perhaps this should raise a warning or error too? Thoughts appreciated!