Make parseResult public again

[fwal]

We currently use Apollo as de-/serialiser for our graphql requests within our pre-existing network stack and it works really well.

We used this method for achieving this but it was recently made internal.

[apollo-cla]

@fwal: Thank you for submitting a pull request! Before we can merge it, you'll need to sign the Meteor Contributor Agreement here: https://contribute.meteor.com/

[martijnwalraven]

I'd like to better understand your use case. Which part of your app calls parseResult? To integrate with a pre-existing network stack, wouldn't it be enough to implement your own NetworkTransport?

Maybe you could also chime in on #6?

[martijnwalraven]

I'm a little hesitant to make parseResult public because its current implementation includes result normalization, and if you only want to deserialize a result, that is wasted effort that could be avoided by implementing result parsing yourself (which should only take a few lines of code).

What do you think?

[MrAlek]

@martijnwalraven Parsing the result ourselves would require implicit knowledge of implementation details (such as that the main body is under the data key).

Also, responseName on Field is internal right now so it can't be done without modification anyway. The cleanest API would be to keep parseResult public.

if let dataEntry = body["data"] as? JSONObject {
    let reader = GraphQLResultReader(variables: operation.variables) { field, object, info in
        return (object ?? dataEntry)[field.responseName]
    }
    data = try Operation.Data(reader: reader)
}

[martijnwalraven]

@MrAlek: Let's make it public for now then, with the understanding that the API may change again in the future. I also still think it's valuable to continue the conversation in #6 to see if there isn't a better way to do this.

[MrAlek]

@martijnwalraven Thanks! And yes, of course, I'm working on a second integration from our end currently (using ApolloClient) and will get back to you with some thoughts based on that.

[martijnwalraven]

Great! I'm also in the process of writing some API documentation, so that may also help.