Rhumb line generation fails with certain lines

[atlight]

Another rhumb line issue (see also #8042).

Rhumb line generation fails with certain lines. I suspect, but haven't fully confirmed, that these lines are "vertical" (normal to the Earth's surface). This bites anyone with these LineStrings in their GeoJSON.

My interpretation of the situation is that PolylinePipeline.numberOfPointsRhumbLine returns 1, and a one-point rhumb line is indeed generated by PolylinePipeline.generateCartesianRhumbArc, which PolylineGeometry.createGeometry is not expecting and does not handle properly.

Sandcastle example:

var viewer = new Cesium.Viewer("cesiumContainer");
var verticalRhumb = viewer.entities.add({
  polyline: {
    positions: [new Cesium.Cartesian3(-4136082.225644,2896505.853869577,-3883655.777889882),new Cesium.Cartesian3(-4136082.8717214135,2896506.306318757,-3883656.38862544)],
    arcType: Cesium.ArcType.RHUMB,
  },
});
viewer.zoomTo(viewer.entities);

This fails with error:

DeveloperError: Expected right to be typeof object, actual typeof was undefined
Error
    at new DeveloperError (https://sandcastle.cesium.com/CesiumUnminified/Workers/Check-6c0211bc.js:62:13)
    at Object.Check.typeOf.object (https://sandcastle.cesium.com/CesiumUnminified/Workers/Check-6c0211bc.js:281:13)
    at Function.Cartesian3.subtract (https://sandcastle.cesium.com/CesiumUnminified/Workers/Cartesian2-36f5627e.js:536:24)
    at Function.PolylineGeometry.createGeometry (https://sandcastle.cesium.com/CesiumUnminified/Workers/createPolylineGeometry.js:420:31)
    at createPolylineGeometry (https://sandcastle.cesium.com/CesiumUnminified/Workers/createPolylineGeometry.js:558:29)
    at createGeometry (https://sandcastle.cesium.com/CesiumUnminified/Workers/createGeometry.js:60:32)
    at callAndWrap (https://sandcastle.cesium.com/CesiumUnminified/Workers/createTaskProcessorWorker.js:61:25)
    at https://sandcastle.cesium.com/CesiumUnminified/Workers/createTaskProcessorWorker.js:107:9

Browser: Chrome 83

Operating System: Windows 10 18363

[OmarShehata]

Thanks for digging deeper into this @atlight - I believe this is the same issue as reported here: #8764. There's a minimal code example there with a vertical line that shows the issue. I'm going to close this and link your notes to that issue.

[atlight]

Thanks @OmarShehata! Evidently my searching skills could use some improvement 😄