tracker-extract seem unable to handle a specific PDF
I'm running Fedora Workstation 28, which use tracker by default. Since upgrading to Fedora 28 recently, tracker-extract seem to keep running over and over on the same set of files (PDFs containing book chapters).
To diagnose I stopped tracker with
tracker daemon -t
and then I started tracker-extract with
TRACKER_VERBOSITY=3 /usr/libexec/tracker-extract -v 3
This gave me a rather long output, which I'm attaching here. Hopefully this is useful for determining why tracker-extract keeps running over and over on the same files.