In the ideal scrum world, shippable code, which is successfully unit-tested, does not fail; That in reality, at least after a first manual acceptance test is no longer that way, I think rather corresponds to our experience. My question is: How do you handle errors that occur in Acceptance Testing, in parallel with running / upcoming sprints?
Do you generally create new user stories from bugs, or do you track them separately in a bugtracking system?
In the case of new user stories, I find their estimation somewhat problematic, since the effort to troubleshoot is, in my opinion, more difficult to estimate than those for implementing new stories.
In addition, a running sprint should remain untouched, but what about serious mistakes that more or less prevent a meaningful continuation of acceptance testing?