genqlient VS graphql-go

The client support in particular is currently very awkward. For instance, there's this discussion around Hasura in the Khan/genqlient repo: https://github.com/Khan/genqlient/issues/272

An example of a Snowflake API request using Go. This example uses the Go standard library, so it doesn’t require additional dependencies; however, in production, you could use a library like shurcooL/graphql or Khan/genqlient for stronger typing.

I also use https://github.com/Khan/genqlient in tandam with it to generate strongly typed graphql clients. Here is a decent example of a golang CLI using it: https://github.com/superfly/flyctl/tree/master/gql

You might also be interested in genqlient, which Khan Academy released earlier this year. It's closer to what you've done, though I don't think we support uploads because that isn't something we've needed to support.

If you want to just generate a client based on a schema, I highly recommend https://github.com/Khan/genqlient which we use at work.

Although it is supposedly not ready for production, I'm personally using https://github.com/Khan/genqlient from the Khan Academy folks.

I'm not sure if I'm going to stop supporting anything outside of int32 and float64 as that was one of my biggest frustrations when using graphql-go.

I have always preferred https://github.com/graph-gophers/graphql-go. I still write my definitions in .graphql files and pese them with the library. Then you can easily write resolvers or just return model objects if they match the definitions.

For my current project we use ent (https://entgo.it) + gqlgen and its a really nice combo. As well as the two other popular graphql libraries for go (https://github.com/graphql-go/graphql and https://github.com/graph-gophers/graphql-go), I’ve also kept my eyes on graphjin (https://github.com/dosco/graphjin) and thunder (https://github.com/samsarahq/thunder) but to me there is no reason to not use ent with gqlgen - it’s just so nice :)