This revision looks correct and sufficiently tested. (I failed to submit draft earlier)

This revision looks correct and sufficiently tested. (I failed to submit draft earlier)

These changes look good with the race fixed.

This all looks correctly implemented based on existing patterns. I do have a question about the pattern for outbound connection limiting and handling dead peers, but not something that should be addressed in this revision anyways.

The actual change is straightforward and correct. All impacted logic branches appear to be tested.

This change looks correct and well tested. There is one small code-order change I suggest.

Looks good. There is one small nit that might be nice to address.

This split makes sense and looks correctly implemented. Had to brush up on iterator semantics to make sure I was understanding the behavior around erasing the item and assigning to the next value but it looks good.

Verified the removal of the transitive include and that the new one fixes the issue with building.

This revision looks correct and tests look sufficient.

This revision was superseded by D10387.

Update function params documentation.

Undo unnecessary/ineffectual code re-ordering.

Re-work locks to reduce their scopes.

The changes look correct and tests look sufficient. Discussed the intermittent test failure with Fabien and am comfortable with the cause.

These additions and typo-fix look good.

The logic and implementation are good. The negative case added to the tests appears sufficient to check the new behavior.

Looks good; no logical changes and the verify result tagging looks like a smart abstraction for this. The tests for the result tags look correct and comprehensive.

This logic looks good, correctly implemented, and correctly tested.

Now that the assertion is fixed, this looks correct and ready to go.

Logic and tests look correct.

This makes sense; kind of surprised it didn't work this way before. The implementation and tests look correct. I personally dislike using the presence of m_addr_known this way, but certainly not enough to diverge from the upstream.

Fix latest feedback items. Fix failures due to lock checks by taking all required locks in once place.

I don't have enough familiarity here to know if this is something we want to adopt, but it appears to be implemented correctly.

Not a strict change request but something to consider. Other than this issue, this changeset looks good/correct.

Activating the polling looks correct and the test looks correct and comprehensive.

The iterating over proofs instead of peers looks better, and the removal logic appears correctly deduplicated without change in behavior.

One small additional assertion that I think would be ideal, other than that this behavior, implementation, and test look good.

One small additional assertion that I think would be ideal, other than that this behavior, implementation, and test look good.

Remove unnecessary format changes.

Fix arc lint issues.

Get tests working correctly.

I like this change; in particular, I think it's more understandable with the explicit check for the invalid state.

Fix unintentional additions.

Move lock out of loop.

Rebase onto new version of D10686

Addressing some review feedback.

Ok, successfully ran ninja check-functional

Fix stacking against D10387.

Address non-functional feedback items.

Thanks, I had missed that. I also didn't know about .git/info/exclude which is useful and sufficient.

Squashing down patches into single commit.

Remove debug statements from test.

Superseded by https://reviews.bitcoinabc.org/D10387

Squash linting changes.

Update Avalanche stake-proof poll test description and shorten lines.

Add functional tests for Avalanche proof-polling.

Fix build and remove untested code additions.