You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Oh I see why. Compounds in that compounds set have no names. Actions in snapshot identify the target object by its name (in this case by tuple of dataset name and compound name). It was done in this way more than two years ago because IDs were unreliable.
I THINK we store also the ID of given object so if don’t find it by its name we can try to find it by its ID as a some kind of fallback mechanism.
And also I THINK that IDs are reliable now because this unreliability was caused by the mechanism of updating the target. Previously it worked in a way that update target was simply deleted and uploaded again. Now the target is just updated so the IDs should be stable. But @alanbchristie knows more about this than I do. @phraenquex So IDs should be stable. Should I implement this fallback mechanism? If I start NOW then tomorrow by the end of the day it should be in staging. It's not hard change but I need to go and at least check in code a LOT of stuff (all action definitions, update functions restoring stuff, check restoring stuff when switching between snapshots in 'Project History' tab)
Bug from #894: colour categories assigned to RHS selected compounds don't persist in the snapshot.
The text was updated successfully, but these errors were encountered: