Consider whether ComparisonReports need to include a ComparisonResult for all differences, or just the first.

[tgregg]

The documentation implies that ComparisonResults should be generated for all differences. It's not clear that this provides much benefit, and risks adding complexity to the implementations and polluting the ComparisonReport.

Once a decision is made, the documentation should be updated to be explicit, and implementations updated as necessary.

ion-c's CLI, for example, only includes the first difference in its ComparisonReport, allowing it to short-circuit on the first difference.

[wesboyt]

+1 for short circuiting.

[tgregg]

Reporting on all differences would really just be for human consumption, because all comparison reports generated by the CLIs for use by ion-test-driver are expected to be empty.

The benefit would be that such a human could view and fix all of the comparison violations at once without having to re-run the driver. But I don't think this is a big enough benefit to justify complicating the implementation and potentially increasing the time the tests take to run.