We need to record series dropped by the filter program in the database somehow, so that we can derive the "last tested patch" from them as well as from pending and tested patches. Otherwise we end up re-running the filter on dropped series, possibly generating reports with every run, until a patch that passes the filter shows up, which we can add to the database, and thus advance the notion of the "last tested patch" past it.
We need to record series dropped by the filter program in the database somehow, so that we can derive the "last tested patch" from them as well as from pending and tested patches. Otherwise we end up re-running the filter on dropped series, possibly generating reports with every run, until a patch that passes the filter shows up, which we can add to the database, and thus advance the notion of the "last tested patch" past it.