juniper-from-schema VS diesel

For my part, I look at something like https://netflix.github.io/dgs/, and look to a project like https://github.com/davidpdrsn/juniper-from-schema and juniper itself (which has async support) as being more of a flexible building block for Rust projects that need GraphQL.

Hello! I'm running into an issue with derives. I'm using juniper-from-schema for a web server, which generates lots of Rust types based on a GraphQL schema file. Trouble is, since the types are generated via macro, I can't apply any derives to the type declaration sites. This is really annoying because I need to implement things like Clone on my types. I've done some googling on this and can't find any way to declare derives away from the type declaration -- is this possible or is it a limitation of the macro system? Is this only possible via some changes to the library to support derive declarations?

There are different ways of creating the GraphQL schema. Either schema first, where you directly define the schema or code first, where your schema is generated based on the code that you have written. Juniper uses the code first schema approach, but there is an alternative schema-to-juniper generator.

7. Diesel

Both seem nice. However, both of them rely very heavily on prepared statements. Unfortunately, using prepared statements is a no-go when you use connection poolers like pgbouncer, or in my case AWS RDS proxy. A discussion in Diesel indicates that disel is not going to provide any support for disabling prepared stements (https://github.com/diesel-rs/diesel/discussions/3575), and a discussion on sqlx hints that disabling prepared statements is possible, but I haven't found any documentation or examples for it.

In addition we are experimenting with prebuild versions of diesel-cli that can be installed directly. We have a set of prebuilt binaries here. We are interested in feedback about how the provided binaries work on your platform.

First of all thanks for making this great tool. As it happens I currently toy around with using it for diesel-cli releases. See the WIP PR here. I think diesel-cli is a good example of a tool that depends on system libraries as it needs to link native database drivers, so this new release is welcome. Defining the dependencies seems to allow easily building things on x86_64-unknown-linux-gnu and x86_64-apple-darwin. It seems to pick up everything in the right way there.

I did look at mysql initially. Then I started checking other crates. Diesel is an Object Relation Model (ORM), I'm not yet keen on taking on the complication of learning ORM, I give this crate a pass in the meantime.

I would recommend people look at diesel from Rust for how nice it could be. https://diesel.rs/ Look at the complex queries example. So much more readable and easier to understand.

Most wanted missing features in diesel

There are two problems with using Rust for web servers:

1. The only production-ready Rust web servers require writing async request handlers. Async Rust is not fun.

2. The only good Postgres client library is async: https://crates.io/crates/sqlx

I'm trying to remedy the first problem with https://crates.io/crates/servlin .

Solving the second problem will be another project. I hope someone else does it. There is https://crates.io/crates/diesel but it has the same problem as async Rust: incomprehensible compiler errors.

Lemmy is written in Rust using Actix Web and Diesel.rs.

https://actix.rs/

https://diesel.rs/