Tony Ghita

Hey, thanks for the contribution! I'm still getting up to speed with the implementation details of execution but I'd like to get this feature in. I believe we have a...

I haven't experienced a need for the selection set in resolvers, but I can see how this may be important if your resolvers directly communicated with a database. Is anyone...

@salmana1 We don't interact with any databases directly—instead we make service requests to hundreds of data services which front databases. The databases are denormalized (generally joins happen in the application...

I think only if the specification explicitly called for it. They do demonstrate this type of usage in an example, albeit implicitly: http://facebook.github.io/graphql/#sec-Directives-Are-Unique-Per-Location

I think adding a stacktrace is a good idea. I _believe_ the preference for this library is to keep external dependencies minimal. Maybe there exists a way to get the...

Yes I would consider that an option 👍

And later, we'll need to add a `subscription` resolver as well

@neelance would you mind if we started work on a pull-request to resolve this issue? What are your thoughts on the approach we should take?

The SDL RFC was merged into the specification recently, and I think approaches queries vs. mutations vs. subscriptions in this way: https://github.com/facebook/graphql/pull/90/files#diff-fe406b08746616e2f5f00909488cce66R167

We don't yet have a PR open for this specific change. I'd be happy to work through it with someone if they were to open a PR. I'd like to...