Replace custom protobuf handlers with librespot-protocol

[Pogodaanton]

In #693 we have introduced librespot-protocol and its direct dependency protobuf to psst-core. The code now uses two different protobuf libraries. This PR aims to reduce it back to one.

I have been unsuccessful at porting new protobuf specs to quick-protobuf. I can't get the bash scripts in psst-protocol running, nor could I reproduce the rust code generation for the already existing protobufs. I tried tackling the problem from the other side instead: Use more of the librespot protobuf code.

[SO9010]

Looks great to me, I need @jacksongoode to look over this as I have not been active for a while... Thanks for the help!

[jacksongoode]

Generally this looks fine, does this mean we can remove a lot of the native code from protocol?

[Pogodaanton]

Generally this looks fine, does this mean we can remove a lot of the native code from protocol?

It means that we can remove the entire folder psst-protocol. As I'm writing this, I've noticed that I haven't done so yet in this PR, however, there is no dependence on the psst-protocol package anywhere in the code anymore.

[jacksongoode]

I would say go ahead, it makes sense to if all functionality appears to work.

[jacksongoode]

I've gone ahead and done this. All appears to work. Given that this is a huge chunk of removed code, I'd like to see how it might change the package size.

[Pogodaanton]

I've gone ahead and done this. All appears to work. Given that this is a huge chunk of removed code, I'd like to see how it might change the package size.

Boy, do I have meager news for you 😅

Jokes aside, the size won't change that much given that all the protocols removed are still used, only from a different package now. As far as I can see in the CI builds, the size change is in margin of error, effectively 0.

[jacksongoode]

Let's do it, I think the convergence with librespot is likely the best path towards sustainability anyways.