Velocity computed using negative time offset if no valid position at positive offset. #5349
Conversation
|
Welcome @klingerj!!! |
|
To test this, load Test coverage of |
|
As predicted, your first merge conflict... 😄 |
# Conflicts: # CHANGES.md
|
I'm still seeing incorrect camera behavior in this branch.
The camera will still radically change position. |
|
Thanks for catching this @mramato! @klingerj looks like we only tested this with the Molniya satellite, not the ISS. The Molniya uses a high-Earth "North-Up" view (these lines), but the ISS and GeoEye use the VVLH view (these lines). In the latter case, the We should also test with trucks (Vehicle.czml) and facilities, although I don't expect any problems there due to relative lack of velocity. |
Source/DataSources/EntityView.js
Outdated
| @@ -73,7 +82,11 @@ define([ | |||
| var inertialCartesian = Matrix3.multiplyByVector(toInertial, cartesian, updateTransformCartesian3Scratch5); | |||
| var inertialDeltaCartesian = Matrix3.multiplyByVector(toInertialDelta, deltaCartesian, updateTransformCartesian3Scratch6); | |||
|
|
|||
| Cartesian3.subtract(inertialCartesian, inertialDeltaCartesian, updateTransformCartesian3Scratch4); | |||
| if (invertVelocity) { | |||
| Cartesian3.subtract(inertialDeltaCartesian, inertialCartesian, updateTransformCartesian3Scratch4); | |||
There was a problem hiding this comment.
Looking at this more closely, the result here gets fed straight into Cartesian3.magnitude and not used again. So I don't think the invertVelocity flag needs to be checked here at all, as magnitude yields an absolute value. The flag does need to be checked in the VVLH section below for yBasis.
…offset-fixed # Conflicts: # CHANGES.md
|
Thanks @klingerj! There was a CHANGES conflict but I took care of it. |
Fixes #4949