API for chaining of tokio_util encoders and decoders

[tomDev5]

In some scenarios, it's desirable to compose multiple encoders or decoders—for example, encoding a struct to JSON, then compressing and encrypting it, while also prepending a length field.

To support this use case, it would be useful to enable chaining decoders via a method like decoder_a.chain_decoder(decoder_b), producing a new decoder that applies both in sequence.

I've put together a small POC here: chained_codec
If this seems like a useful addition, I’d be happy to contribute it as a tokio_util feature. Let me know your thoughts!

[Darksonn]

Encoders turn an AsyncWrite into a Sink, and decoders turn an AsyncRead into a Stream. You can't compose them since once you've converted them once, now you have a Stream/Sink, but to apply a codec you need an AsyncWrite/AsyncRead.

[tomDev5]

Maybe my wording isn't exact, as I have managed to implement what I had in mind.
Specifically this part from the integration tests I added (this if from the PR I created, the POC link is very similar):

    let mut decoder = BitwiseNotDecoder {}
        .compose_decoder(BitwiseNotDecoder {})
        .compose_decoder(BitwiseNotDecoder {})
        .compose_decoder(StringDecoder {});

What this means is - to decode, the data should go through 3 decoders of type BitwiseNotDecoder, and once through the StringDecoder.