1. Open any document
2. Blank it
3. Type some text
4. Dump the model into the console (ve.instances[0].documentModel.data)

Expected output: [ 'F', 'o', 'o', 'b', 'a', 'r' ]
Actual output: [ 'F', ['o', ve.AnnotationSet([])], ['b', ve.AnnotationSet([])] ...

The pre-annotations code is applying the empty set as a pre-annotation to every character except the first. This shouldn't happen, it should just not annotate if it's got nothing to annotate with.

Version: unspecified
Severity: normal