Comment 12 for bug 149300

I understand, hunting down all those keywords would be very intensive. So do I understand correctly that in the current situation the index slowly gets polluted? Or is there some kind of "cleanup" process that periodically weeds out deleted files?

If it doesn't exist yet, perhaps something like that can be implemented using a black list of deleted files (which is updated upon deletion of files). This can then be taken into consideration by the search tools and can be used for cleaning up the database in idle time.

Then again, I am by no means acquainted with the Tracker code and internals so this might be a stupid suggestion. There may be all kinds of difficulties that arise by implementing such a scheme.