Should the documentation have an example for how to use multiple decoders?

In my implementation I need to use the default decoder for byte[] and GsonDecoder to map to objects and I’m doing the following in Java 11:

@Bean
    public Decoder decoder() {
        Decoder decoder = (response, type) -> {
            if (type.getTypeName().equals("java.lang.String") ||
                    type.getTypeName().equals("byte[]")
            ) {
                return new Decoder.Default().decode(response, type);
            } else {
                return new GsonDecoder().decode(response, type);
            }
        };
        return new ResponseEntityDecoder(decoder);
    }

Is this worth having an example in the documentation? I was inspired by this discussion https://github.com/spring-cloud/spring-cloud-netflix/issues/2246

[sshevlyagin]

That sounds way better @velo. My scenario was the API returns image/jpeg or application/json depending on which endpoint you hit.

re: wiki, no need to give me write access you can just copy/paste that gist if you want

[velo]

@sshevlyagin what if we added a new #decoder(MediaType,Decoder) to Feign.Builder

Invoking decoder(Decoder) would overwrite the decoder as is today. Invoking decoder(MediaType,Decoder) would replace decoder with a TypedDecoder, and would assign the existing decoder to media-type */* Subsequent invocations of decoder(MediaType,Decoder) would append a new Decoder to the TypedDecoder. If MediaType already registered on TypedDecoder log warning and replace.

TypedDecoder would be a implementation of Decoder that checks the MediaType first and then delegate the decoding process to Decoder