Luuk Wester

I have created some types that can do some of the lifting when documenting a type (I'm 100% sure that they are not sufficient even for just OpenAPI, but lets...

> The different files for the different impls are unnecessary Gone > we should start with structures that are able to describe the entirety of the serde data model I...

I have added a deep linking feature to make sure that the urls contain the full route that is being viewed. I have also added a feature flag to make...

I did some digging and it seems that the problem could be resolved using [quote_spanned!](https://docs.rs/syn/1.0.14/syn/spanned/index.html). I don't understand the proc_macro API well enough to make the change myself however.

This does indeed fix the error messages! Thank you for this fix.

You are right, I just encountered this issue myself :smile:. It seems that any generic type nested with another generic does not get handled well. Reopening the issue.

It now indeed works for (deeply) nested generic types. Thank you for you update!

In the latest installment of this issue, I have found another instance of `#[openapi]` not producing the right error messages. It happens when I specify in the `#[post]` macro that...

Additionally, it is still possible to create unspanned error message using some functional constructs on `Result`s and `Option`s. Interestingly, the following code produces normal error messages: ```rust #![feature(decl_macro, proc_macro_hygiene)] use...

Hi @GREsau, They all generate sensible errors except for one! The one that remains is that if I do this: ```rust #[openapi] #[get("/", data = "")] fn myroute() -> String...