Hi
We’ve had longstanding issues with reordered positions not being saved properly, and also tasks moved between kanban columns not working and resetting.
I’ve looked into this some more now and for the reordered position it seems to be a corruption of the positioning system. The database uses a position double to store the order of things.
In our case it apparently got to a point where there are just numerous tasks with the same position value 3870920346.6308727 in one case. I don’t quite know how this happened, like lots of moving positions around, or maybe it’s a legacy issue and this developed over time after updating Vikunja. One note in general, we have very big projects.
In any case, I’d love to see a solution to this. Maybe Vikunja could detect a case like this when the position is the same and to a full reassign of all position values in a list to fix this?
Or as a hotfix is there maybe a database script I could run to fix this for currently broken setups?
Thanks a lot for looking into this
There is now (#1998) a cli command to repair broken task positions and a bunch of heuristics that should automatically prevent and recalculate broken positions when they happen. Please check with the latest unstable build if the problem still persists.