Refactor #4117
Updated by ttereshc over 5 years ago
The implementation of RpmContentUnitSaver ErratumContentUnitSaver needs to not rely on attaching attributes to models and then reading those attributes back later. For background see comments around this one: https://pulp.plan.io/issues/4060#note-12 That is currently done in the rpm code here: https://github.com/pulp/pulp_rpm/blob/844336c1c55427851861491a50746c1c2af48d4e/pulp_rpm/app/tasks/synchronizing.py#L293-L327
.