chore: move protobuf module from utils to common

[LNSD]

Following previous work with sqlite.nim and envvar_serialziation.nim modules, this PR moves the protobuf module to the common module. This module should, in the future, act as a "facade" for the protobuf codec APIs.

• Move protobuf.nim under common module
• Re-export libp2p's protobuf codec APIs
• Update imports accordingly

[status-im-auto]

Jenkins Builds

❔	Commit	#️⃣	Finished (UTC)	Duration	Platform	Result
✔️	f01590d	#1	2022-11-16 13:38:48	~15 min	macos	📦bin
✔️	f01590d	#1	2022-11-16 13:42:46	~19 min	linux	📦bin

[jm-clius]

Great! Worth adding a line to the Coding Guide re "use common/protobuf and don't import libp2p minprotobuf directly"?

[alrevuelta]

Using @jm-clius guideline "use common/protobuf and don't import libp2p minprotobuf directly"?, perhaps we are missing updating waku_peer_storage.nim as well?

We can leave that for another PR, but do we also want to remove libp2p/protobuf/minprotobuf from the tests?

[LNSD]

Great! Worth adding a line to the Coding Guide re "use common/protobuf and don't import libp2p minprotobuf directly"?

Note that I have only touched the protocol modules. So... I am not 100% sure if we should do it. I have not analyzed the impact of moving everything to import this common module.

I had in mind to do that when we decided the protobuf library (see: #1361)

Using @jm-clius guideline "use common/protobuf and don't import libp2p minprotobuf directly"?, perhaps we are missing updating waku_peer_storage.nim as well?

I am unfamiliar with waku_peer_store.nim; that's why I left it out. To minimize the risk. The waku_peer_store needs a cleanup. Changing that, which will change in the near future, IMO, is out of the scope of this PR.

We can leave that for another PR, but do we also want to remove libp2p/protobuf/minprotobuf from the tests?

Same thing here. I don't know the implication of removing the types. And I am pretty sure that, in some cases, the analysis will lead us to a deeper refactoring of the test suite. I want to avoid this extra work to be part of this PR.

[rymnc]

LGTM!

[LNSD]

The check failure can be ignored. It is not related to the changes.

The offending test case is already tracked in #1357.