Hi @asaf. I reported a similar issue on this #81. I tend to agree. The property name that is the source of the exception should be made available. I created my own version for the time being until this gets addressed.

I am confused by what the desired behavior is as well. On v0.1.1, the ProperyPath was set to the child that violated the schema. After upgrading to v0.2.0, the path is now set to the parent.

I'm not sure if this change was intended or if a fix (I only had to move a single line of code) would be welcome?

I just noticed this issue as well (and that there hasn’t been a tagged release in quite some time).

It would be good if the behavior was documented if working as intended.

I'll aim to review the spec and look into it by the end of the week if time permits to move this along.

@Arqu any news on this? thanks!

I've looked into it and apparently neither is really up to spec anymore. As the output format is not really correct per spec.
Taking the more useful route, I'd say the better way would be to show the property path and not just the parent path until this is patched.

For anyone interested, the KeyError structure needs to be updated (and surrounding logic) to contain (at a minimum) keywordLocation, instanceLocation & error which would contain both options and satisfy the basic output format per spec. This is more looking ahead to the 2020 spec as for the 2019-09 it's not clearly defined and only mandates the "errors" field.