Tracing errors is hard

[switchtrue]

Tracing errors can be incredibly difficult as graphene collects the error messages and returns them. This is great for production but not during development. For example, I get the following back from a query and have no idea what is causing the error:

{
  "errors": [
    {
      "message": "'unicode' object has no attribute 'get'"
    }
  ],
  "data": {
    "users": null
  }
}

Would it be possible to include some sort of debug flag that can raise exceptions with stacktraces on error?

[switchtrue]

I believe I've raised this in the wrong place it should be in Flask-graphql. Apollogies.

[TimOrme]

@Mleonard87 Did you ever find a way to do this in flask-graphql? Having the same issue

[ekampf]

@Mleonard87 graphene generate the response json not Flask. It collect said errors as part of its execution... should probably improve it and add trace etc

[reverie]

The discussion at graphql-python/graphql-core#237 might be helpful