ExecutionResult.formatted does not format errors

When I call formatted on an ExecutionResult, the errors portion of the dict is not formatted and instead returns a list of GraphQLErrors. GraphQLError is not serializable, but it does have a formatted property which is, why is formatted not called on the list of GraphQLErrors?

As a work around, I have to call formatted on my ExecutionResult and then rebuild the dict and format the errors:

    formatted = result.formatted

    data = formatted.get('data')
    errors = ([err.formatted for err in formatted.get('errors')] 
              if formatted.get('errors') is not None 
              else None)
    extensions = formatted.get('extensions')

    if extensions:
        body = dict(data=data, errors=errors, extensions=extensions)
    else:
        body = dict(data=data, errors=errors)

Ideally ExecutionResult.formatted would perform this error formatting for me.

Environment: python 3.8 graphql-core 3.1.4

[Cito]

Don’t worry. The execution results and errors are not automatically formatted. We’re only talking about the case when you access the formatted property of these objects. With other words, if you access the errors of the execution result, they will be still a list of GraphQLErrors, if you access the errors of the formatted execution result, they will be a list of formatted GraphQLErrors.

[AlecRosenbaum]

Right, sorry let me rephrase.

If GraphQLError’s are automatically formatted when being put into the result, is there any way of accessing traceback data? I had thought the accepted way was by looking at original_error.__traceback__, which would no longer be possible if this was changed.