Oriano de Stefani

This looks exactly like what i would need too. With the SOAP source every action is mapped to mutation Type which is just wrong but i get why its happending....

i'm currently quite busy at work, but i'd love to pick the vue integration up again in a bit. Iirc it wasn't far away from a functional implementation without the...

As you can see here: https://github.com/thecodingmachine/graphqlite/blob/3b97f9d020297658e32833e9b3d49e7b27c650dd/src/Mappers/RecursiveTypeMapper.php#L212-L223 Your type somehow is not yet part of the typeRegistry, might be a ordering issue in the recursive type mapper?

You are right, this seems quite naive. The interface type name is generated in this function called on line 214 above: https://github.com/thecodingmachine/graphqlite/blob/3b97f9d020297658e32833e9b3d49e7b27c650dd/src/NamingStrategy.php#L31-L34 It should actually check if there is a...

@oojacoboo what do you mean by two mappings?

I'd consider implementing a DescriptionFieldMiddleware using the existing FieldMiddlewares as described in https://github.com/thecodingmachine/graphqlite/issues/439#issuecomment-1031227899 by @withinboredom since it would be possible to have the new Annotation not ever trying to read...

@withinboredom have a look at this PR #466 this would be possible to implement when afterwards. alltought factories are probably still handled in a diffrent way, i did not touch...

I'm not sure, but that might be the case that a null value will cache forever. It seems odd that the globTTL is always set but the mapTTL not.

I've been thinking about this lately and feel like a Namespace Attribute on Controllers would be the most logical. Imagine: ```php #[GQL\Namespace] class UserController { #[GQL\Query] public function user()... #[GQL\Mutation]...

I mean yes its possible, but but hacky. And the goal of graphqlite is great DX, with the Namespace Attribute it can be backwards compatible and easily added to existing...