[QGIS-Developer] PR about Temporal data, plz test/review

[Richard Duivenvoorde]

Hi devs (or users),

I created a PR [0] which (for me) fixes a bug [1] when working with vector data which have an 'event' with some kind of temporal duration.

Anybody working with Temporal Features: please have a look and test with your data. 
After a positive review I hope that this can be backported to 3.16 too.

The main problem: IF the timeframe of the event fell exact on the timestep-timeframe, those events showed up in 2 timesteps (in my case falsely doubling the result). 

My main 'solution' is to NOT take (both) limits into account for the sliding step-frame selection.
In case of features with a zero-time duration (moments in time, like Positioned Tweets or so), you HAVE to take one limit into account (and we do).

Note that the fix is in QgsTemporalNavigationObject so maybe also affects other feature types.

I hope this does not break anybodies work flow (it shouldn't, because I think this fixes something, but...  you never know).

Thanks in advance.

Regards,

Richard Duivenvoorde

[0] https://github.com/qgis/QGIS/pull/40989
[1] https://github.com/qgis/QGIS/issues/38468