GraphQL Stack Traces are (Needlessly) Unhelpful;

Here is a stack trace I recently got:

Error: Cannot return null for non-nullable field User.id.
at completeValue (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:560:13)
at completeValueCatchingError (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:495:19)
at resolveField (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:435:10)
at executeFields (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:275:18)
at collectAndExecuteSubfields (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:713:10)
at completeObjectValue (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:703:10)
at completeValue (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:591:12)
at completeValueCatchingError (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:495:19)
at resolveField (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:435:10)
at executeFields (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:275:18)
at collectAndExecuteSubfields (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:713:10)
at completeObjectValue (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:703:10)
at completeValue (/home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:591:12)
at /home/me/projects/my-project/server/node_modules/graphql/execution/execute.js:492:16

The problem is, it's entirely made up of lines from `graphql`, which means it's entirely worthless for helping me understand where the error occurred (ie. the entire thing stack traces are _supposed_ to do). All I know is that it involved a resolver which somehow involved a User ID.

I don't mean to sound ungrateful for awesome free software (that is the best way to build an API today) ... and I _very much am_ appreciative of all the hard work that has gone into this project ... but would it kill you to at least include the query and variables used in the stack trace?

At the moment it really feels like the library is saying "Oh you messed up? Well, too bad: you're on your own now chump! That's you what you get for making a mistake." ;) But surely you _want_ to help users understand errors, no? And surely GraphQL has the query/variables available to include in the error, since it receives both long before the error occurs?