Objects don't revert to unsnapped positions if connections were unsuccessful

[phet-steele]

This one takes admittedly precise movements, so prepare yourself.

1. Take out two fixed length components, don't connect them.
2. Start by taking finger A and dragging one (by its body) until it snaps to the other. Don't release, instead HOLD FINGER A VERY STILL.
3. Take finger B and move the other element away. You have effectively moved the would be targetVertex out from under the would be oldVertex.
4. Without moving finger A at all, release it from the screen. If finger A's element jumped at all before this step, you've failed and need to restart. We don't want to move this element from its snapped position.

Now taking the other element and trying to snap it to finger A's element, you'll see that it tries to snap to the unsnapped position, which looks weird since finger A's element never reverted to its unsnapped position.

(ignore that the video is on phettest)
Seen on iPad Air 2 iOS 9.3.5. For phetsims/qa/issues/51.
URL: https://www.colorado.edu/physics/phet/dev/html/circuit-construction-kit-dc/1.0.0-dev.166/circuit-construction-kit-dc_en.html
Version: 1.0.0-dev.166 2017-10-11 16:23:32 UTC
Features missing: touch
Flags: pixelRatioScaling
User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/61.0.3163.100 Safari/537.36
Language: en-US
Window: 1920x1014
Pixel Ratio: 2/1
WebGL: WebGL 1.0 (OpenGL ES 2.0 Chromium)
GLSL: WebGL GLSL ES 1.0 (OpenGL ES GLSL ES 1.0 Chromium)
Vendor: WebKit (WebKit WebGL)
Vertex: attribs: 16 varying: 32 uniform: 1024
Texture: size: 16384 imageUnits: 16 (vertex: 16, combined: 80)
Max viewport: 16384x16384
OES_texture_float: true
Dependencies JSON: {"assert":{"sha":"928741cf","branch":"master"},"axon":{"sha":"b8dd4e51","branch":"master"},"babel":{"sha":"45eca7d2","branch":"master"},"brand":{"sha":"cfca902d","branch":"master"},"chipper":{"sha":"66763c36","branch":"master"},"circuit-construction-kit-common":{"sha":"7d852b4c","branch":"master"},"circuit-construction-kit-dc":{"sha":"22604c0e","branch":"master"},"dot":{"sha":"84bc5146","branch":"master"},"joist":{"sha":"3eff115f","branch":"master"},"kite":{"sha":"9689a3de","branch":"master"},"phet-core":{"sha":"054b2d6d","branch":"master"},"phetcommon":{"sha":"5b0c8dd6","branch":"master"},"query-string-machine":{"sha":"c74e454e","branch":"master"},"scenery":{"sha":"db7ea875","branch":"master"},"scenery-phet":{"sha":"a90363b5","branch":"master"},"sherpa":{"sha":"f0d5f750","branch":"master"},"sun":{"sha":"d57850fd","branch":"master"},"tandem":{"sha":"f399e7ee","branch":"master"},"twixt":{"sha":"a2eaa10a","branch":"master"}}

[ariel-phet]

Seems like a serious edge case. Marking as deferred, unassigning.