When updating existing features, if the assigned_to_user field contains information that does not match an existing user, the feature will change ownership to the default assignee. This should instead result in a warning that the CSV import contained inaccurate data AND the feature should not be changed.
Good news: the CSV import no longer silently erases the "Assigned To" field if the "Assigned To" value in the CSV file is invalid!
Bad news: the CSV import now silently fails to update the "Assigned To" field if the "Assigned To" value in the CSV file is invalid.
That's much better than the old behavior. But the new behavior still gives no indication to the user that the CSV Import failed to update the "Assigned To" field.
Thanks for the status update! I'm glad this is now likely to be implemented.
Under the current behavior, data is lost and the user is never informed of the error. To me, that is unacceptable behavior.