Julian Berman
Julian Berman
Can you elaborate how this conflicts with adding `$schema`? I've cited sections of the spec in #376 which conflict with your interpretation. Do you have sections which support it? As...
I certainly would rather do so than keep talking about trivial changes. There's nothing different about these tests than existing optional ones, but I won't keep pushing to get them...
I don't believe 1 is spec compliant, but would have to double check we still have language requiring unknown keywords to be ignored. 2 is what this PR was meant...
Hello! The need for tests beyond just valid or invalid is a long-standing issue but one that has to be done carefully given we don't have a formal backwards compatibility...
Side tests are probably best for now yeah until we can express these in the suite, which I hope to do soon. It's important to differentiate between when the *instance*...
There's no real reason other than that it'd be noise (a property needed in all schemas) -- it's also *valid* to leave it out even though it's highly recommended, so...
> Quite the opposite. If you specify with $schema, it applies at that draft. But if you don't, it makes a best guess given the keywords that are used. Right...
> but I would find the ability to outright contradict it to be of dubious value. I didn't mention contradicting -- though yes, mine supports that too. Just specifying it...
Well -- how about this -- if you and @gregsdennis feel confident one can submit it and the other review it that I don't object to -- I *do* think...
That part is/was json-schema-org/json-schema-spec#210 but I don't remember what happened there. But agreed though on we should have them (though I think it'd be nice if we actually did it...