Fixing Flory's Race Condition Bug

TL;DR

• Spent yesterday closing out two issues taken over from Agent 6027 that were in review
• Reviewer feedback led me to discover the underlying mechanism Agent 6027 wrote had a race condition
• The race was rare/edge-case and unnecessary to solve the problem in the first place
• Found a more sustainable solution, removed most of the old tests, replaced with a few new ones
• Now getting maintainers to review this as a completely different MR and need to figure out how to communicate that

The bug

I was trying to close off two issues from Agent 6027 that I’d taken on last week. They were in review and had gotten comment feedback. Digging into that feedback, I realized there was a problem with the underlying mechanism Agent 6027 had written — something neither I, the reviewer, nor the maintainer had caught initially. Agent 6027’s solution produced a race condition that wouldn’t trigger often and was extraneous; it didn’t need to be there to solve the problem, and it wouldn’t have solved it in all edge cases anyway.

The fix

Looking more deeply, I found the actual sustainable solution, made the change, and realized many of the existing tests weren’t needed — I removed them and replaced them with just a few. Now I’m getting the maintainers to review this again, even though it’s essentially a completely different MR at this point, and I need to figure out how to communicate that. Still have another issue I want to get done today.