abci: implement vote extensions

[cmwaters]

This builds on top of #9532 and is part of feature/abci++vef

This PR traces through the following PRs trying to make the appropriate changes on top of main

• abci++: Sync implementation and spec for vote extensions #8141
• abci++: Vote extension cleanup #8402
• abci++: Propagate vote extensions (RFC 017) #8433
• abci++: add consensus parameter logic to control vote extension require height #8547
• abci++: add proto fields for enabling vote extensions #8587

ref: #9396

---

PR checklist

• Tests written/updated, or no tests needed
• CHANGELOG_PENDING.md updated, or no changelog entry needed
• Updated relevant documentation (docs/) and code comments, or no
  documentation updates needed

[sergio-mena]

I just finished doing the diff-of-diffs comparison between the branch with the cherry-picked commits and this PR, which has been rebased to the tip of the feature branch, and now also contains #9532.

This experience was much better than FinalizeBlock. I am glad I could find problems on both sides:

• On this PR, there are some gaps in the production logic, in particular in what has to do with the newly introduced "enable extensions" ConsensusParam. That logic is partially missing, and we would only have found out probably through --painful-- extensive testing, or worse, by our users. So the diff-of-diffs comparison (and therefore this PR) has been instrumental in avoiding that.
• The cherry-pick branch is code complete (no hunks left behind by the very method used) but by comparing it with this PR I found some differences, and even a few bugs!; caused either by the (manual) conflict resolution process during the cherry-picking, or even in the v0.36.x implementation.

The way I am planning to move forward is the following:

• I have extracted all relevant divergences between the two diffs, and added them in a commit here at the tip of the cherry-pick branch including:
  • all bugs, and parts I am not sure about at this time can all be seen in that commit
  • all other relevant divergences (not necessarily bugs), are also there, also useful comments added in this PR
  • all of the above will have their own dedicated PR where we will be able to better examine them
• As it is now up to me to drive this effort, I will go for posting smaller, well defined chunks of this PR for others to review, rather than having someone else review them in one shot as I've done so far.

[sergio-mena]

All the differences between this PR and branch feature/abci++vef are now integrated as tracked by #9887