Elmer de Looff

Digging in to this a bit more, it seems as if the `resolution_scope` of the `RefResolver` provided by `jsonschema` has the wrong scope to resolve the URL correctly. That said,...

This could be worked around by means of a validator, adding an explicit `"default"` key to the dict to pop off and construct a new defaultdict from: ```py class Model(pydantic.BaseModel):...

It sure looks like one of the paths is checked but the other one is left open. I'm curious if we really need two checks, or if the `null()` object...

Apologies for the delay. The change looks good but the testing needs a small change to separate the two actual tests.

Triggering `set` events from a nested change sounds like a pretty aggressive and potentially impactful thing to do by default. However, being able to opt in to behaviour like this...

The `changed` method on the `NestedMutableList/NestedMutableDict` should be the last in the chain to get hit, resulting in exactly one call regardless of the level of nesting (all objects below...

I don't know the exact behaviour of `get_history`, assuming this is a SQLAlchemy function, but assuming it gets the previous object from the session, what you are showing is expected...

Where we stand now, I think the actual sorting is tested pretty exhaustively. All of the text/value transforms are tested, loading is tested and the behavior of numbers in `data-value`...

I think this is resolved (for now anyway). I can't really think of any parts that seriously lack testing. There are some minor bits like the texts and classes that...

Sounds good. I think there's some room for improvement (making things easier) on how the headers (and the sortable body) are detected by making use of `` and `` elements...