graphql: Return better error message if a type only contains ID field

[pawanrawal]

Fixes GRAPHQL-449

---

This change is 

Docs Preview:

[abhimanyusinghgaur]

Reviewed 4 of 4 files at r1.
Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @MichaelJCompton and @pawanrawal)

---

graphql/e2e/custom_logic/custom_logic_test.go, line 226 at r1 (raw file):

		name: String!
	}
	type Movie {

Why did we need to change this? @remote types won't be checked by the rule right as we store them in dgraph?

---

graphql/schema/gqlschema_test.yml, line 2051 at r1 (raw file):

Quoted 4 lines of code…

      input AuthorUpdate {
        id: ID!
        name: String!
      }

Someone may want to identify a query input with just an ID. So, input types should still be allowed with only ID!, as we don't store them in dgraph.

---

graphql/schema/gqlschema_test.yml, line 2057 at r1 (raw file):

Quoted 13 lines of code…

  -
    name: "@custom directive with variable definitions in operation in graphql"
    input: |
      type Author {
        id: ID!
        age: Int!
        name: String! @custom(http: {
          url: "http://google.com/",
          method: "POST",
          graphql: "query ($id: ID!, $age: Int!) { getAuthor(id: $id, age: $age) }",
          skipIntrospection: true
        })
      }

Why did we remove this?

[abhimanyusinghgaur]

Reviewable status: all files reviewed, 3 unresolved discussions (waiting on @MichaelJCompton and @pawanrawal)

---

graphql/e2e/custom_logic/custom_logic_test.go, line 226 at r1 (raw file):

Previously, abhimanyusinghgaur (Abhimanyu Singh Gaur) wrote…

Why did we need to change this? @remote types won't be checked by the rule right as we store them in dgraph?

sorry, read that line as: Why did we need to change this? @Remote types won't be checked by the rule right as we don't store them in dgraph?

[pawanrawal]

Reviewable status: 2 of 4 files reviewed, 3 unresolved discussions (waiting on @abhimanyusinghgaur and @MichaelJCompton)

---

graphql/e2e/custom_logic/custom_logic_test.go, line 226 at r1 (raw file):

Previously, abhimanyusinghgaur (Abhimanyu Singh Gaur) wrote…

sorry, read that line as: Why did we need to change this? @Remote types won't be checked by the rule right as we don't store them in dgraph?

Actually this seemed a bit weird that a @remote type has custom fields. We don't resolve custom fields within remote type yet. Not remote types themselves are responses of custom fields. Maybe we should even throw an error for a remote type having custom fields for now?

---

graphql/schema/gqlschema_test.yml, line 2051 at r1 (raw file):

Previously, abhimanyusinghgaur (Abhimanyu Singh Gaur) wrote…

      input AuthorUpdate {
        id: ID!
        name: String!
      }

Someone may want to identify a query input with just an ID. So, input types should still be allowed with only ID!, as we don't store them in dgraph.

Ok

---

graphql/schema/gqlschema_test.yml, line 2057 at r1 (raw file):

Previously, abhimanyusinghgaur (Abhimanyu Singh Gaur) wrote…

  -
    name: "@custom directive with variable definitions in operation in graphql"
    input: |
      type Author {
        id: ID!
        age: Int!
        name: String! @custom(http: {
          url: "http://google.com/",
          method: "POST",
          graphql: "query ($id: ID!, $age: Int!) { getAuthor(id: $id, age: $age) }",
          skipIntrospection: true
        })
      }

Why did we remove this?

Thanks for flagging, got removed because of bad merge.

[abhimanyusinghgaur]

Reviewed 2 of 2 files at r2.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @MichaelJCompton and @pawanrawal)

---

graphql/e2e/custom_logic/custom_logic_test.go, line 226 at r1 (raw file):

Previously, pawanrawal (Pawan Rawal) wrote…

Actually this seemed a bit weird that a @remote type has custom fields. We don't resolve custom fields within remote type yet. Not remote types themselves are responses of custom fields. Maybe we should even throw an error for a remote type having custom fields for now?

Yes, I think we should throw that error. It would be better.

[MichaelJCompton]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @abhimanyusinghgaur and @pawanrawal)

---

graphql/e2e/custom_logic/custom_logic_test.go, line 226 at r1 (raw file):

Previously, abhimanyusinghgaur (Abhimanyu Singh Gaur) wrote…

Yes, I think we should throw that error. It would be better.

Yeah, sounds right to throw an error if there's custom things in remote types until we support it. Can you add a card to jira.

[pawanrawal]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @abhimanyusinghgaur and @MichaelJCompton)

---

graphql/e2e/custom_logic/custom_logic_test.go, line 226 at r1 (raw file):

Previously, MichaelJCompton (Michael Compton) wrote…

Yeah, sounds right to throw an error if there's custom things in remote types until we support it. Can you add a card to jira.

Created GRAPHQL-486