Open jorroll opened 6 years ago
Hmm, looking at some of the tests, looks like same goes for MutationType
as well.
Oops, looks like I'm wrong in saying that the current format breaks the spec. The spec has an example with type MyQueryRootType
as the root query type.
While I'm still confident in saying that the current name is unconventional (I can't remember ever seeing a root query type not named Query), I'm going to close this issue.
It's causing issues with my schema stitching, but it sounds like I should figure out another way to solve the issues as this behavior is valid.
We can easily fix it. No problem. Thanks for the feedback.
Btw. sorry for the silence on your other issues, there is a lot going on with GraphConnect I pick it up again after next week.
No worries! I actually saw that GraphConnect was happening and figured I wouldn't hear from anyone for a week or two 👍
At the moment, neo4j-graphql appears to place auto-generated root fields in a special
QueryType
type. It's unclear if this is a bug or intentional. As shown in the spec, the root level query type is expected to have the form:Using introspection on the auto-generated schema, it does not appear as though there is a root level
Query
type.Update
I probably misspoke: fields have levels to them, but type definitions do not. The
schema
object hasquery
andmutation
root fields, its just that those fields are typed incorrectly. In addition to being unconventional (spec breaking?), this also impacts schema stitching.