Failed to extract `Data<awc::client::Client>` for `proxy`

Well, the samples from the README are just snippets showing the relevant parts of how to use the traits of this crate. What they don't show is that you need to add a awc::client::Client to the app data of your actix-web webserver in order for it to work. Like this, for example:

use awc::Client;

use actix_web::{get, web, HttpResponse};

use actix_proxy::{IntoHttpResponse, SendRequestError};

#[get("/{url:.*}")]
async fn proxy(
    path: web::Path<(String,)>,
    client: web::Data<Client>,
) -> Result<HttpResponse, SendRequestError> {
    let (url,) = path.into_inner();

    let url = format!("https://duckduckgo.com/{url}");

    // here we use `IntoHttpResponse` to return the request to 
    // duckduckgo back to the client that called this endpoint
    Ok(client.get(&url).send().await?.into_http_response())
}

#[actix_web::main]
async fn main() -> std::io::Result<()> {
    HttpServer::new(|| {
        App::new()
            .app_data(web::Data::new(Client::default())) // This is the important line
            .service(proxy)
    })
    .bind(("127.0.0.1", 8080))?
    .run()
    .await
}

How can I improve documentation to make it clear that you are required to add an instance of awc::client::Client to the app data in the examples? Do you want me to add the main function to the snippets as well? I didn't add it, because I thought it'd make the snippets too complicated as it is not immediately clear what the benefit of this crate might be. Alternatively, I could reduce the examples and create the client inside the endpoint, but that's very wasteful and I don't necessarily want to promote code that is clearly suboptimal. My preferred enhancement to the docs would be to add a fully working example (#3) to the repo (which would basically be the code I listed above), would that have been helpful for you? Then we could add a link to the example in the docs.

jofas / actix_proxy

Failed to extract `Data<awc::client::Client>` for `proxy` #17