BUG: Crashes 3 times during 12 km running activity ("follow course" routing). Wrong Recovery Time calculation after end of activity, and too short distance calculation!

BUG Report:
===========
Garmin Product: Forerunner 945 LTE
SW Version: 5.07 (=latest version)

Date: 2022-07-03

Title: Crashes 3 times during 12 km running activity ("follow course" routing). Wrong Recovery Time calculation after end of activity, and too short distance calculation!


Expected behaviour:
-------------------
The watch should not crash during an activity, not even when routing is active.
And IF it crashes anyway, I expect that no course segment is unnecessarily(!) missing afterwards (i.e. distance should not be implausible!).
And IF it crashes anyway, I expect that the recovery time calculated after the end of the activity is correctly taking into account the complete activity and not calculating a much too low value.


Observed behaviour:
-------------------
One main bug (B1) and two "follow-up bugs" (B2, B3):

B1: The watch crashed 3 times during my first proper running activity in which I used the "follow course" routing method for following a course that I had previously planned in Garmin Connect web browser.
After each crash it took about one minute until the watch was back and I could resume the activity. The crashes happened after ca. 6.8 km, 7.0 km and 11.0 km.

B2: After the first crash, after the watch had rebooted, the distance was indicating 6.65 km, although last time I had looked at the distance BEFORE the crash it was already indicating 6.70 km, and the crash happend about 100 meters later at around 6.80 km. I did not pay attention to this in crashs nb. 2 & 3 but suspect the same effect (note that crash 2 was a very short time after crash 1). The total course distance was hence shorter than expected, because of missing distances after each crash. We can see these "missing distances" in the "heart rate plot over distance", which has weired anomalities at the points where the crashes happened - see the diagram/screenshot attached!

B3: After the end of the activity, the watch indicated a recovery time of only 9 hours. This activity was quite exhausting for me and these 9 hours are WAY too little, as I know from ample experience of earlier similar trainigs with the same watch! Compared to other, comparable activities, I would have expected a Recovery Time of more than 40 hours, but at least certainly much more than 24 hours. I suspect that the revocery time algoritm was confused by the crashes and did not take into account the complete activity but only parts - maybe just the part between last crash and final destination was taken into consideration, or at least the section from start until the first crash was disregarded, or the section from first crash till the final destination was disregarded.   


Steps to reproduce:
-------------------
1.) Planning an 11.7 km round course in Garmin Connect Web browser using "Freehand" mode. I used many manual points to draw the route, esp. at crossings where I wanted to enforce turn signals. At such locations (like Y-shape crossings that normally do not generate turn signals due to much less than 90 degree turns) I created a modified course such that I got 90 degree turns to enforce the turn signals (which worked by the was). Note that none of the crashes happened near such special crossings that contained many points on the manually modelled course, so I do not know that there is a causal relation.

In Addition to the normal route I also saved 7 manual course points on this course (but the crashes did not occur near any of these course points, so not sure if there is a causal relation).

2.) I started a running activity and used navigation to route along this pre-planned course. The routing setting was set to "follow course", not "use map".

3.) During this runnning activity I saved a few positions - but the crashes did not happen directly near any of these saved positions, they seemed unrelated, so no sure if there is a causal relation.

4.) During this runnning activity I switched the data field display now and then between map and numerical display (distance/pace/time/height), but the crashes did not occur while I was actively switching but always in between, so not sure if there is a causal relation.

Planned Course:

Example of a tailored Freehand course planning to enforce a turn signal at Y-chape crossing by drawing a >90 degree turn:

Recorded Heart Rate over Distance with anomalities on the distance scale, relating to the "follow-up bug" nb. B2 explained above.