Structure thoughts

[n1ru4l]

What are our goals brainstorming?

• We wanna teach people how we perceive and recommend GraphQL tooling
• There is not just a server, a "server" is made of
  • Schema
  • HTTP logic
  • GraphQL logic (validate, parse, execute, subscribe)
• How can we make building a Graphl Schema type-safe?
  • Codegen!
• Why you want to use envelop
  • Show how and why you would implement logging from scratch
  • Show and implement error masking handling from scratch (e.g. via leaked database exceptions when trying to sign up with the same mail twice)
  • Show how and why you would implement parse caching from scratch
  • Show how and why you would implement validate caching from scratch
  • Show how we move that into an envelop plugin and how it makes stuff cleaner (aka we just taught how to write your envelope plugin)
  • Show how we already have all those plugins and you can just install and use them
• Why you wanna apply authorization rules via SDL
  • Note: I am personally unsure if that is the way we should teach
  • TODO: insert reasons why in here
• error handling using unions, interfaces https://blog.logrocket.com/handling-graphql-errors-like-a-champ-with-unions-and-interfaces/
  • refactor signIn mutation to return a union of Success Error instead of raising an error if an email is already in use
  • input unions via oneOf
  • explain the struggle
  • show envelop oneOf solution
• response caching
  • why? reduce server load
  • show how to use with envelop
• defer and stream
  • why do we need it
  • use it with an example that makes sense
• Live Queries
  • Explain why subscriptions are annoying when you just want to have the latest data
  • TL:DR: https://the-guild.dev/blog/subscriptions-and-live-queries-real-time-with-graphql
  • Propose live queries as potential alternatives with examples where it actually makes sense
  • show json patch capabilities for only sending changed deltas to clients

WIP structure:

• Improve the TypeScript schema development experience
• Better schema modeling
  • Better Error modeling and handling with Union and Interface Types
  • Graphql inspector for detecting breaking changes
• The anatomy of a GraphQL request
• Customizing the GraphQL request
  • Learn how to write validation rules
  • Introduce Envelop
• Making your GraphQL server production-ready
  • Authorization/Authentication
  • Error masking
  • query complexity analysis
  • parse/validate caching
  • response caching
  • error reporting
  • split schema into public and private API
• The bleeding edge
  • defer and stream
  • input unions via oneOf
  • live queries

[saihaj]

Do we want to cover features that still are not part of spec like oneOf and defer and stream?

[Urigo]

yes we do