This makes use of the MultiSourceReader provided by the graphQL library.
Using this reader allows to read a single document from multiple files,
while correctly associating line numbers and errors to their original
source file.

This doesn't bring any overhead because the method Parser.parseDocument(String) that was previously used already created a MultiSourceReader itself.
Note that the previous method didn't associate the source file name with the content used for parsing, while the new one does.

It allows to solve cross-reference problems with union types for instance.

Resolves:
https://github.com/kobylynskyi/graphql-java-codegen/issues/27Co-authored-by: NJoffrey Bion <joffrey.bion@booking.com>