I think that make sense for the teams that already have experience with AWS services

@joshua Manual query and mutations, no request pipeline, manual database creation, etc.

I agree with @jhony0311. I've you're already part of the AWS borg and understand their other services (Cognito / IAM), it makes sense. I like the modularity of GraphCool where I can let them run it, let them cluster it, or go running to Digital Ocean or bare metal.

Simply compare your

types.graphql

with the full schema you can download, and see what you have to do manually with AWS AppSync

@agartha I just want through the AppSync tutorial. Most of the database creation is automated if desired. It looks like you can insert Lambda to make a pipeline.

@agartha Yes, it looks like everything else AWS, very feature rich, but with a high learning curve. However, they do mention auto database setup based on your GraphQL schema as an option: https://docs.aws.amazon.com/appsync/latest/devguide/provision-from-schema.html#aws-appsync-provision-from-schema

Even if database creation can be partly automated, it's still based on a full schema definition. Queries and mutations will not be generated automatically

Yes, there’s not nearly as much generation behind the scenes.

I'm more interested in their AppSync client and how they're caching offline mutations and rewalking the query tree when the client comes back online

@agartha what do you mean by request pipeline?

Before and after hooks

Got it

What I hate the most is that they forked the Apollo codegen repo, and call it aws-appsync-codegen.

Is there any guidance/docs/anything on setting graph.cool up on Azure?

Is there a way to clear or reset a local graphcool db data?

What's the best way to access to the local db?

does any one else have a problem when running locally

graphcool deploy --dry-run

, it always tells me my functions are wrong but

graphcool deploy

works fine, and the functions work correctly?

is there any bigger project on github what is using apollo-client? 🙂

Hey guys, I want to create a resolver that returns information about the logged in user. An endpoint similar to the

/me

endpoint as seen in many APIs. As per Graphcool's online examples, issuing a query from a resolver to the CRUD Api for the logged in user would look something like this:

const query = `
    query me($id: ID!) {
      viewer {
        User(id: $id) {
          < fields from the client go here >
        }
      }
    }
  `;
  const variables = {
    id: event.context.auth.nodeId,
  };
  const api = fromEvent(event).api('relay/v1');
  const response = await api.request(
    query,
    variables,
  );

My question is: Is there a more idiomatic way to do this in Graphcool? And would it make sense for the client to pass up a variable (potentially named

fragment

) that will be used inside the query to fetch the fields the client is interested in?

Does anyone know what version of know the console uses for Hooks and Function?

It is

v4.8.6

Im looking for a best practice server example to follow for my react app im building. It needs authentication(social, email/password), crud actions, and i need to set up elasticsearch for a complex search and filter page. Anyone know of one?

Hey everyone! There's something weird going on wiht the website. https://cl.ly/o441

Not sure if it's only me or there's an actual problem with it. Just wanted to let you know.

Is it possible to have a hook function (e.g.

operationBefore

) fire in response to a change in a relationship? For example, I have

Department

and

Employee

types, and a relationship between the two to define managers. I want a hook to fire in response to the

addToDepartmentManagers

mutation. I've tried hooking into both

Department.update

and

Employee.update

operations but neither get triggered.

I've also tried the generated

AddToDepartmentManagersPayload

, but that causes

deploy

to throw an error.

Hi everyone, may I ask if it's possible to deploy an instance of graphcool using external MySQL instance with current graphcool-cli?

a