proposal-error-cause VS graphql-spec

:::note When using remote APIs, we often have the possibility to generate the types automatically from a JSON schema for REST APIs, from protobuf files for gRPC-based APIs, from a database schema, etc. You might even be using an external API through an SDK that already provides you with all types. In such cases, the creation of specialized Error classes is not mandatory. However, it might still be a good idea to do so to provide application-specific errors rather than bubbling up 3rd-party low-level errors. For such cases, the upcoming Ecma TC39 proposal for Error Cause is useful as it allows to chain errors. Polyfills exist: Pony Cause or error-cause. :::

Error Cause : .cause property on all Error types slides.

The impact and cause provides the most value when paired with the other, and that's what Error Cause enables and what Pony Cause is is a ponyfill for and provides helpers for.

Well, we have error-cause on stage-3 for the same and with which we could do something like:

Error Cause: Enhancing errors with a distinct "cause".

Error Cause

GraphQL's first draft release was 8 years ago. [1]

It's first non-draft release was 5 years ago. [2]

It's first release under a community foundation was 2 years ago. [3]

[1] https://spec.graphql.org/July2015/

[2] https://github.com/graphql/graphql-spec/releases/tag/June201...

[3] https://github.com/graphql/graphql-spec/releases/tag/October...

I'm a big believer in GraphQL (in fact, at time of writing I'm #2 contributor to the GraphQL spec itself) so it pains me that a tool I built doesn't always have easy ways to achieve the "versionless schema" design that GraphQL encourages when it comes to making significant breaking changes to your underlying database tables. (Personally, I think you should aim for your database schema itself to be versionless, but this is not always possible.) Of course you can build your PostGraphile schema over views instead of tables, but views have their own problems that I won't go into here…

I created a proposal for Map type but didn’t make it through.

https://github.com/graphql/graphql-spec/pull/888

The issue with GraphQL is it tries to appease too many masters.

Similar to jsx. The language isn’t evolving.

The good thing is the spec is (almost) frozen, so there’s many implementations, the bad is it can encompass the flexibility of json schema can do.

Longer thread - Subscriptions RFC: Are Subscriptions and Live Queries the same thing?

https://github.com/graphql/graphql-spec/issues/284

You’ve probably figured it out by now, but for others who may be in a similar position; GraphQL is a specification (with various implementations) and you can read up on the spec here: https://spec.graphql.org/

If you're asking whether GraphQL supports anonymous objects that can be arbitrarily nested then no, it doesn't.

show this feature request some love https://github.com/graphql/graphql-spec/issues/174

* Performance: It's just hard to track down what makes an operation slow. The waterfall nature of resolvers is a big contributor

[1] https://github.com/graphql/graphql-spec/issues/488

:::note GraphQL Union is available for Types only, not for Inputs. However, the oneOf directive will bridge the gap in the future.