Shane Krueger

We should be able to write support for the `@oneOf` directive on top of the existing framework. It's simply a custom directive coupled with a validation rule that asserts that...

Having this directive exist is fully spec-compliant, just like how `@length` is. And moreso than the directive introspection feature we added, which is not spec compliant. So I'm just saying...

@Socolin If you're referring to OneOf Input Objects, see https://github.com/graphql/graphql-spec/pull/825 . Once the RFC is merged we will work on changes in this library for native support.

If you are trying to sanitize HTML fields, I would write a custom scalar - e.g. `HtmlGraphType` - which would perform sanitization to remove javascript or whatever before it hits...

I would generally agree but I think we should maintain backwards compatibility in some manner - not necessarily source compatibility but some option or easy code snippet to retain old...

Sure. Looks good to me

@dgmoex It sounds like you have two different schemas, with one using MyEntityPublicType and one using MyEntityAdminType, correct? The problem is not that both of the graph types are being...

My first thought is that the cache should be implemented entirely as IConfigureExecution and we can remove IDocumentCache. But there’s more to consider. I’ll look later tonight

I believe this will only bury the problem deeper and make it more difficult to identify later. The problem still exists even in the user's code -- a multithreaded initialization...

Rereading your comments... > I consider the case with singleton graphtype and scoped schema/schemas like "rather normal". But... this is impossible. During initialization, middleware modifies all graph types attached to...