Add an option to provide module for the custom scalar definitions

[miterst]

Description

Currently, the definitions for the custom scalars have to specified in the parent (super ) module.
For example if you have a custom DateTime scalar it generates for you type DateTime = super::DateTime;

This PR adds the option custom_scalars_module which can be used to specify the path to the module that contains the definitions for the custom scalars that are necessary for the generated query.

For the above example, setting this option to crate::custom_scalars the generated code for the DateTime scalar will be
type DateTime = crate::custom_scalars::DateTime;

How it can be used

• From the graphql-client

   graphql-client generate query.graphql --schema-path schema.graphql --custom-scalars-module='crate::custom_scalars'

• deriving

   #[derive(GraphQLQuery)]
   #[graphql(
       schema_path = "schema.graphql",
       query_path = "query.graphql",
       custom_scalars_module = "crate::custom_scalars",
   )]
   pub struct MyQuery;

Closes #245

[tomhoule]

Amazing, thanks for the PR! I don't have time to review right away, but I will come back to it very soon. Is it already ready for feedback?

[miterst]

Amazing, thanks for the PR! I don't have time to review right away, but I will come back to it very soon. Is it already ready for feedback?

I saw that some tests on the nightly build had failed(worked on my machine 😅 ).
I will try to fix them later.

Anyway it's ready for feedback.

[tomhoule]

I just did a first read-through over my lunch break, and it looks great. Next steps for me is looking into the CI failures in case you need help, and confirm manually that it works, then I think we can merge :) A release is also long overdue, so I'll try to find time for that after this PR is merged.

[miterst]

I just did a first read-through over my lunch break, and it looks great. Next steps for me is looking into the CI failures in case you need help, and confirm manually that it works, then I think we can merge :) A release is also long overdue, so I'll try to find time for that after this PR is merged.

The tests were failing because there was a compiler bug in Rust unstable and was affecting one of the WASM examples.
With the new Rust unstable it seems to be compiling without any problems.

[tomhoule]

I've delayed manual testing for this too long, I'm going to take some time now to do it.