Parsing: chunking

[rbalicki2]

In the pico world, it would be great to divide parsing of selection sets into a chunking step and then actually parsing, e.g.

field Query.Foo {
  bar
  baz # not defined
  syntax.error
}

If we hover on bar, then what should happen is that we do something like

• get_chunks(some_identifier)
  • get_type("Query") -> query_object_id
  • get_field(query_object_id, "Foo")
  • find which chunk contains bar, i.e. the first chunk. This is done via comparing row/col or whatnot
  • get_field(query_object_id, "bar")
  • etc

i.e. at no point do we actually get get_field(query_object_id, "baz") and determine that it doesn't exist!

Chunking for iso literals:

• we know that fields must end with a closing curly, comma, or line break
• we know that curlies must match
• we know that fields cannot start with @
• that should be enough to divide the above into chunks: "bar", "baz", "syntax.error"

Chunking for GraphQL schema:

• should be easy enough to do, also lower priority, though
• but basically apply similar heuristics: go until closing matching curly or another keyword

I'm not sure how IDs work in this case, but presumably, we can chunk and find? idk