Validator does not catch missing value parameters

[scf37]

if declared hello with (a: string, b: string) following will pass execution successfully:
{hello(a: "4")}
{hello(b: "4")}
{hello}

code (sorry for Scala):

val queryType = newObject()
        .name("helloWorldQuery")
        .field {
          newFieldDefinition()
            .`type`(GraphQLString)
            .name("hello")
            .argument {
              GraphQLArgument.newArgument()
                    .name("a")
                  .description("description")
                  .`type`(GraphQLString)
                  .build()
            }
            .argument {
              GraphQLArgument.newArgument()
                  .name("b")
                  .description("description")
                  .`type`(GraphQLString)
                  .build()
            }
            .dataFetcher(env => "world")
        }
    .build()

    val schema = GraphQLSchema.newSchema()
        .query(queryType)
        .build()

    val graphQL = new GraphQL(schema)

    val result = graphQL.execute("{hello(a: \"4\")}")
    System.out.println(result.getData)
    System.out.println(result.getErrors)

[bbakerman]

Can you expand more on what your expectations are here?

I think your are right in that the spec says 👍

http://facebook.github.io/graphql/#sec-Validation.Arguments

that arguments with a non null "type" should not allowed to be non null at runtime.

[scf37]

I expected error about missing arguments.
Does nullable type of field argument imply default value of null?

[bbakerman]

I think I have it wrong. The library requires you to wrap the type in a GraphQLNonNull type first

so

 new GraphQLNonNull(GraphQLString)

Thats the equivalent of the JS implementation where you use !

rollDice(numDice: Int!, numSides: Int): [Int]

So by default the Java API assumes its not required

[bbakerman]

GraphQLNonNull should have a factory for static import so you could do

 GraphQLNonNull.nonNull(GraphQLString)

and hence

 .name("a")
 .type(nonNull(GraphQLString))

[okorz001]

It seems this issue could be closed?