mld: Initial protocol boilerplate.

[nrybowski]

Very early MLD protocol boilerplate. At this point the YANG lib fails to load the module.

@rwestphal Interestingly, IGMP and MLD are both defined in ietf-igmp-mld. How will we handle that since both protocols are (or will soon be) defined in two different protocol crates?

[nrybowski]

I saw that a2be4c9#diff-5246d5827a28289713a7a2862650266d88b0fb20e7d0f3e76d1c291194ccc9bf already contains some boilerplate to include ietf-igmp-mld and some related YANG modules. Maybe we could extract this addition in a separate commit so that we can progress on both IGMP and MLD in parallel?

[rwestphal]

@nrybowski Makes sense to me. I'm currently traveling but I should handle this once I get back home later today. Happy to see more work on the multicast front!

[rwestphal]

Maybe we could extract this addition in a separate commit so that we can progress on both IGMP and MLD in parallel?

@nrybowski The IGMP code has been merged to master.

@rwestphal Interestingly, IGMP and MLD are both defined in ietf-igmp-mld. How will we handle that since both protocols are (or will soon be) defined in two different protocol crates?

Indeed, this will be the first time a single module is implemented by two separate protocol crates. But since the IGMP and MLD crates will implement disjoint sections of that module, the northbound layer should support this case without issues. The only requirement is that any YANG subtree must be owned by a single crate.

Happy hacking!

[nrybowski]

@rwestphal Thanks for the update! I will rebase on master and continue the implementation from there.