Open jessesuen opened 5 days ago
The data structure that drives this is composed on the back end since #2389. I looked at this briefly today, hoping it would be a quick fix, but it's quite complex as it turns out, and I believe the structure of this data will have to be revised (to be less map-centric and more list-centric). @Marvin9 and I will have to tag-team this so the front-end and back-end agree on the new data structure.
@krancour Do you know why did we move this calculation from UI to backend?
As I dig into this, what I see is, API construct the view for UI but then there is no reaction of any changes (eg. new freight, freight moves to stage) from either side to re-construct or update the view. Thats what lead to refresh page every time to see any changes
iirc, @rbreeze had said that UI performance improved a lot by offloading this to the back end.
Description
The image history is randomly listed and changes on every refresh:
https://github.com/user-attachments/assets/9a1f8290-bd35-486a-92a5-4225c7d6b0a3
We should sort by semantic version if we detect that they are semvers (we can just look at the first one to determine this), and then fall back to lexographical if it's not a semvers.
Version