When generating typedefs for interfaces / extending other typedefs (and removing the duplicate fields, HaxeGenerator.hx) -- I'm not sure this is the correct way to do it. Because the extended field may have a type that is a (super? sub?)-set of the interface. I think that's what the above link is saying.
Make some tests. See if it's valid .gql. See if HxGen handles it.
http://facebook.github.io/graphql/October2016/#sec-Object-type-validation
When generating typedefs for interfaces / extending other typedefs (and removing the duplicate fields, HaxeGenerator.hx) -- I'm not sure this is the correct way to do it. Because the extended field may have a type that is a (super? sub?)-set of the interface. I think that's what the above link is saying.
Make some tests. See if it's valid .gql. See if HxGen handles it.
Update README's wrt covariance.