Hi Adam,
thanks indeed. As Hardy mentioned, we know there are several limitations in the support we have for this in Lucene 3, and have big plans for Lucene 4. The main problem in this area is to be honest that we don't have a performance test covering the faceting functionality. If you'd like to contribute such a test, that will for sure influence the results of the upcoming new mayor version.

In the meantime though, I can trust your report and indeed the fix seems trivial so I don't mind integrating this improvement for the time being.