Get schema values synchronously from the AST

[jdesrosiers]

In @hyperjump/json-schema, you can compile a schema once and use that compiled schema to validate against many instances much more efficiently. The compile step is asynchronous and the interpret step is not. I'd like to make returning the result from getErrors an output option for the interpret step, but I can't because getErrors is asynchronous. That's why I added the validate function here (#203) for now where the interpret step can have a different API.

I want to find a way to make the getErrors function synchronous so it can be used in @hyperjump/json-schema.

The asynchronous part is getting values from the schema. Those values should all be in the compiled schema AST. So, it should be possible to build something to get values synchronously from the AST instead of getSchema/Schema.value.

[Ahmedmhmud]

I agree with you, there is no need to getting values for the schema while we have the compiled AST already there.
We can add a small AST resolver built once per compiled schema. It will create a fast lookup index keyed by schemaLocation (and parent location for sibling keywords) so error handlers can synchronously fetch the compiled keyword value they need. So the resolver walks AST once, stores the value of the location, then handlers just do sync reads.

After that, flipping to sync logic is the easy part: We convert handlers to use the resolver, then remove await from getErrors and its call sites. This gives us a sync error path while keeping behavior the same.

If you want, I can work on it.