decide for _one_ protobuf compiler, use (and enforce) it everywhere, run it on CI

[marten-seemann]

We're currently using both gogofast and gogofaster. Both of them are deprecated: https://github.com/gogo/protobuf#deprecated-protocol-buffers-for-go-with-gadgets

p2p/protocol/internal/circuitv1-deprecated/pb/Makefile
7:              protoc --proto_path=$(GOPATH)/src:. --gogofast_out=. $<

p2p/protocol/circuitv2/pb/Makefile
7:              protoc  --gogofast_out=. $<

p2p/protocol/holepunch/pb/Makefile
7:              protoc --proto_path=$(GOPATH)/src:. --gogofast_out=. $<

p2p/protocol/circuitv1/pb/Makefile
7:              protoc --gogofast_out=. $<

p2p/security/noise/pb/Makefile
7:              protoc --proto_path=$(PWD):$(PWD)/../.. --gogofaster_out=. $<

p2p/protocol/identify/pb/Makefile
7:              protoc --proto_path=$(GOPATH)/src:. --gogofast_out=. $<

p2p/host/autonat/pb/Makefile
6:      protoc --gogofast_out=. --proto_path=$(GOPATH)/src:. $<

p2p/host/peerstore/pb/Makefile
7:              protoc --proto_path=$(GOPATH)/src:. --gogofaster_out=. $<

core/peer/pb/Makefile
7:              protoc --proto_path=$(PWD):$(PWD)/../.. --gogofaster_out=. $<

core/record/pb/Makefile
7:              protoc --proto_path=$(PWD):$(PWD)/../.. --gogofaster_out=. $<

core/crypto/pb/Makefile
7:              protoc --proto_path=$(PWD)/../..:. --gogofaster_out=. $<

core/introspection/pb/Makefile
7:      protoc --proto_path=$(PWD):$(PWD)/../..:$(GOPATH)/src --gogofaster_out=Mgoogle/protobuf/timestamp.proto=github.com/gogo/protobuf/types:. $<

core/sec/insecure/pb/Makefile
7:              protoc --proto_path=$(GOPATH)/src:../../../crypto/pb:. --gogofaster_out=. $<

We should decide for one compiler to use, to be consistent. We should also run go generate / make on CI, to make sure that our Protobuf code is consistent with the proto definitions.

[marten-seemann]

I added a benchmark test in a branch here: https://github.com/libp2p/go-libp2p/compare/benchmark-identify-protobuf?expand=1. Note that this is benchmarking a proto2.

Not sure if we want to merge this into master, since it's actually just a benchmark of the protobuf library.

Using GoGo protobuf:

goos: darwin
goarch: arm64
pkg: github.com/libp2p/go-libp2p/p2p/protocol/identify
BenchmarkIdentifyProtobufMarshal-10              2770962               404.3 ns/op           896 B/op          1 allocs/op
BenchmarkIdentifyProtobufUnmarshal-10             741806              1473 ns/op            2328 B/op         45 allocs/op

Using Google protobuf:

goos: darwin
goarch: arm64
pkg: github.com/libp2p/go-libp2p/p2p/protocol/identify
BenchmarkIdentifyProtobufMarshal-10              2404834               496.0 ns/op           896 B/op          1 allocs/op
BenchmarkIdentifyProtobufUnmarshal-10             643617              1675 ns/op            2312 B/op         45 allocs/op

This confirms that Google protobuf is slower than GoGo (about 20%). I would have expected the difference to be much bigger. I get similar results for proto3.

[MarcoPolo]

Given that gogo protobuf is deprecated, and the slow down is small relative to network overhead. I think it makes sense to use google protobuf.

If users need that 20% they can always make sure to use gogo for their own protocols. We don’t send that many protobufs in core. Thoughts?