FPGATrackSim: fix EFTRACK-743, check again if hits are in regions

This is something that the recent mapping changes seems to have broken. We had removed the mapping of hits from the RawToLogical tool; part of that included a check that hits actually be in a given eta/phi region.

The check wasn't removed entirely, it just got moved into later in the data prep algorithm. Unfortunately I/we didn't notice that this is subsequently after spacepoint formation, meaning that spacepoints could be formed from a combination of hits, one inside the region and one outside the region.

We then later test to make sure spacepoints are "in the region" before forming roads. Unfortunately, this check uses the coordinates of the inner paired hit, so it wasn't catching this.

This was causing a very odd error to be omitted from the eta pattern filter tool; as part of this commit I added more information to that debug message, without which I wouldn't have been able to catch this. It was also causing a crash when converting FPGATrackSim spacepoints to offline spacepoints; probably for exactly the same reason. Running with this fix seems to have fixed the crash.

Tagging @imaznas, @jahreda, @wcastigl. This should resolve EFTRACK-743.