Iain Beeston

@Relequestual @Julian Should we get formal approval from the authors? Is there any way to get one of them to 👍 this PR? Furthermore, @Julian can you confirm that I...

Incidentally, these tests will fail on a lot of implementations, as shown by the paper. I've also tested them on the implementation I contribute to (https://github.com/ruby-json-schema/json-schema/) and even the latest...

Thanks @DomagojVrgoc !

I didn't include test T3 from [the paper](http://cswr.github.io/JsonSchema/spec/why/) because I couldn't see a succinct way of doing it... The original test checks that a schema for an object ignores value-based...

So - rather than simply ignore those properties, we would fail validation because the schema itself is invalid? (Is that correct?) Is that how we handle invalid schemas in the...

Yeah I think it's worth raising as another issue. FYI the ruby json-schema gem validates the user's schema against the appropriate meta schema (eg http://json-schema.org/draft-04/schema), before validating the data against...

Thanks for your very thorough feedback @Relequestual @epoberezkin and @awwright - I've removed several tests that used invalid schemas, and corrected one test that @epoberezkin pointed out had a mistake...

@epoberezkin I've added these tests to draft6 now (in addition to draft4)

I just noticed this as well. I've also noticed that calling `set_depth_for_self_and_descendants!` on the root doesn't calculate the depth correctly either. At the moment my workaround is to call `rebuild!`...

The build is failing because travis is trying to use bundler 2.0 with ruby < 2.3. #459 should fix that