TTFF value is wrong when GPS is automatically started on startup about gpstest

[barbeau]

After testing GpsStatus.getTimeToFirstFix() with GPS Benchmark running in the background, it seems there is still an issue with TTFF not being calculated properly by the GPS engine when there are multiple activities with LocationListeners registered. However, since we no longer use the MyLocationOverlay, there isn't the "counting" effect within GPSTest anymore. Therefore, we can stick with the official GpsStatus.getTimeToFirstFix() and see if we can get the multiple activities issues resolved with the Android platform team.

Quote from my original write-up on mikeandroid#3:

TTFF for the GpsStatus.getTimeToFirstFix() in the underlying GPS engine seems to be calculated by:

TTFF =  the most recent GpsStatus.GPS_EVENT_FIRST_FIX time - start time of this activity monitoring GPS

Since the GpsStatus.GPS_EVENT_FIRST_FIX is triggered by another activity (in this case the GpsMapActivity continuously registering the MyLocationOverlay on each GPS fix), the resulting behavior is the TTFF continuously climbing by 1 sec every second.

It seems that the underlying problem is a bug in how the GPS engine reports TTFF when multiple Activities are interacting with the engine. It seems that TTFF should be reported independently for each Activity, with the TTFF being reported as 0 for an Activity B that starts listening to GPS updates when another Activity A is already listening to GPS updates (i.e., the two apps are listening to GPS concurrently). And, Activity A TTFF should never change based on another Activities actions of listening to GPS, if Activity A was the Activity that triggered the original GPS engine startup.