I'm a little at a loss at these CI failures but will keep digging.

Edit: Turned out to be nondeterminism from -tags invariants.

Reviewable status: 0 of 3 files reviewed, all discussions resolved

What was the underlying cause of the non-determinism?

Reviewed 5 of 5 files at r1.
Reviewable status: all files reviewed (commit messages unreviewed), 2 unresolved discussions (waiting on @jbowens)

iterator.go line 139 at r1 (raw file):

	Count int
	// ContainedPoints records the number of point keys encountered within the
	// bounds of a range key.

this is just about the bounds and nothing about masking, so the point keys could be more recent?

iterator.go line 141 at r1 (raw file):

	// bounds of a range key.
	ContainedPoints int
	// SkippedPoints records the count of the subset of CoveredPoints point keys

... subset of ContainedPoints ...?

Has this really been sitting unreviewed since August?

Ha, no, it's been sitting with CI test failures since August. I had a hard time debugging the CI test failures and then got caught up in the 22.2 rush.

What was the underlying cause of the non-determinism?

it was the disabling of the noop try-seek-using-next optimization, which uses the iterator pointer. a noop would result in fewer encountered point keys

TFTR!

Reviewable status: all files reviewed (commit messages unreviewed), 2 unresolved discussions (waiting on @sumeerbhola)

iterator.go line 139 at r1 (raw file):

yeah, added another sentence to clarify that

iterator.go line 141 at r1 (raw file):

Done.