losing stack trace, anyway to retain it ? about apollo-errors

[AdelBachene]

I have the same problem eventhough adding the source error in the internalData and data

       data: {
            sourceMessage: (error.message) ? error.message : "",
            sourceStatusCode: (error.code) ? error.code : ""
        },
        internalData: {
            sourceError: error
        }

I could still see it in all catches but I lost when it arrives at format error Im able to see the message only

 import {isInstance as isApolloErrorInstance, formatError as formatApolloError} from 'apollo-errors';

    graphqlExpress((req) => 
                 ({
                    schema: schema,
                    context: {
                        req: req
                    },
                    formatError: formatApolloError,
                    debug: true,
                    tracing: true,
                    cacheControl: false
                })
        )

-------------------------------> I found out that my issue was because of the schema stitching where it will swallow the original error and my work around for this is:

import {isInstance as isApolloErrorInstance, formatError as formatApolloError} from 'apollo-errors';

                   formatError: (error) => {
                        let errors: any[] = [];
                        error.originalError.errors.forEach((currentError) => {
                            let currentOriginal = currentError.originalError;
                            if (isApolloErrorInstance(currentOriginal)) {
                                Logger.Instance().gatewayError(currentOriginal);
                            }
                            errors.push(formatApolloError(currentOriginal, false));
                        });
                        return (errors.length > 1) ? errors : errors[0];
                    }

[thebigredgeek]

@appsolutegeek are you using schema stitching as well?

[iangregsondev]

no, I used another library to merge my schema...

[codegoblin]

@thebigredgeek I am also curious how one should go about solving this problem. Let's say my model layer is calling a connector and that connector throws an error. What's the proposed way to bubble up that original error's stacktrace to the graphql response's errors array? If you use an ApolloError in the resolver to mask the original connector error, it would seem that the returned stacktrace is pointing to the resolver error rather than the original trouble spot in the connector.

[thebigredgeek]

PRs welcome on this. Closing for now