Fix snap-to-route crash #106
Conversation
| List<RouteLeg> upcomingLegs = routeProgress.directionsRoute().legs() | ||
| .subList(routeProgress.legIndex() + 1, routeProgress.directionsRoute().legs().size()); | ||
| for (RouteLeg leg : upcomingLegs) { | ||
| if (leg.steps() != null && leg.steps().size() >= 2) { |
There was a problem hiding this comment.
Wouldn't it be easier to use a for(i= routeProgress.legIndex() + 1 ; ...)? That way we don't need sublists.
In general since returning null is acceptable, I am not sure if we shouldn't just return null if the next leg has no steps or < 2 steps? But I guess both approaches would be fine as this seems to be an edge case either way.
There was a problem hiding this comment.
Wouldn't it be easier to use a for(i= routeProgress.legIndex() + 1 ; ...)? That way we don't need sublists.
Sure. Make sense!
In general since returning null is acceptable, I am not sure if we shouldn't just return null if the next leg has no steps
If we return NULL the bearing is set to 0, this means the puck and camera is turn around to zero position, and will turn back for next step.
This also happens on the last position. Maybe we should store the latest bearing and re-use them instead of return a 0 bearing. In this case we don't need the loop-stuff.
What do you think?
There was a problem hiding this comment.
How about we use the actual bearing of the location? Or yeah, we could use the last bearing, I think this should be fine for most cases as well.
There was a problem hiding this comment.
I like the idea of fallback to the real location. But this is often not route matched and we have also the issue #63. So I would implement the last bearing usage for now, but I think we can improve this later after some other topics to the location.
|
I removed the loop now. If the necessary points not available, the logic is fallback to the last calculated bearing. If them also not available it will fallback to the real location bearing. |
I discovered that legs that contains only a single step will cause a
IndexOutOfBoundsExceptioncrash.Cause was only a wrong if-condition. But to prevent a wrong bearing of
0, I adjust the fetching of the upcoming step.