Roman Ivantsov

File enterprise-attack, the following 2 IDs are referenced in relationship objects but are not defined in the file: malware--c19cfc89-5ac6-4d2d-a236-70d2b32e007c malware--c41a8b7c-3e42-4eee-b87d-ad8a100ee878

Spec references RFC7231 which is obsoleted

4

Here are couple of things I found that might be an issue: Section [Response](https://github.com/graphql/graphql-over-http/blob/main/spec/GraphQLOverHTTP.md#response) > A server must comply with RFC7231. First, it would be nice to provide RFC title...

Default Root Op type names: removed duplicate statement; clarified that only Query type required

10

Default Root Op type names: removed duplicate statement, added that only Query type is required. @benjie argued previously that the second sentence (starting with Likewise...) is not a duplicate, it's...

added mutation op type to examples 193, 194

17

The examples clearly show mutating fields (changeBirthday); since mutating fields are allowed only on top level, we clearly see the 'top' level of mutation request; however, there is no op...

Change example for directive order to use Query or Mutation, not type system

1

[Directives Order Is Significant](https://spec.graphql.org/draft/#sec-Language.Directives.Directive-order-is-significant) sub-section is in section "2. Language", which talks about Query language. The directive order example surprisingly uses a type definition, while type system and its SDL...

field merging - field TYPES must not differ

3

Minor correction, Editorial.

Removed parallel execution requirement for merging fields

7

The main point here is to remove 'when executed in PARALLEL'; service is free to execute the request 'normally' - in any order it wants, parallel or not. But I...

provided explicit ref to Value Completion section

1

the current version mentions executor (what is that by the way?), and then: > this is covered in that section which section is not clear, we do not have 'executor'...

Fixed capitalization of subsections in chapter 7. Response

1

Changed capitalization of subsections to first-cap style like in the rest of the doc

Changed 'must NOT BE' to 'must not be'

1

There are 38 occasions of phrase 'must not be' in the spec; this is the only one with such a strange capitalization