Closed aaronayres35 closed 1 year ago
Do we know which data range/mapper is giving the bad values? I'm worried that there may be other places which are relying on the datalow <= datahigh invariant and so this may just push the error elsewhere.
In any case additional logging of the case datalow > datahigh (equal is OK) may be worthwhile.
Failure is unrelated - the same changes are needed as in upstream pacakages for ubuntu-latest
This PR includes a defensive programming "fix" to #824 to prevent plot crashing. The real fix would be to track down where the bad inputs are coming from. However, regardless I think it makes sense for auto_ticks to be more robust to bad inputs as was discussed on the issue.
Also, in the same downstream project, when an application is resized to be very small and an inspector pane visibility is toggled, the plot some how ends up in a situation where
datalow > datahigh
inchaco.axis.py
and we are hitting theRunTimeError
. I believe it makes sense to handle this case the same that we do for wheredatalow == datahigh
(namely there is no plot we could show so we have no ticks). I am unable to cause the downstream project's plots to crash when working off of this branch.