Gloria Zhao
Gloria Zhao
Thanks @instagibbs for the testing and review, added your fuzz commits and took comments. Still need to write the p2p_orphan_handling test.
> My understanding is that m_total_orphan_usage exists to limit memory usage, and m_total_announcements to limit CPU usage - but why the number of announcements instead of number of orphans? Yep,...
> The changes here have surprisingly little effect on the included benchmarks: My expectation was that the time would go up with these changes, but hopefully not by too much....
Will add something like this as comments as well but here's the thinking around these benches: - The `EraseForBlock` bench exists to test the worst case `EraseForBlock` time, which is...
I'm sad too! Seeing the stats from https://delvingbitcoin.org/t/stats-on-orphanage-overflows/1421 made this more pressing in my opinion, but it's not a regression. I think we can still try to consider small, obviously...
Following coredev discussions, I'm working on a few things: - given selected peer, evict by entry/announcement time instead of randomly - shorten expiry time (and GETDATA interval) - rewrite as...
Yes! I think multi index could also be its own PR. But wanted to give a status update here.
This is ready for review again. Main changes from February: it includes the rewrite as a boost::multi_index container, removes `-maxorphantxs` entirely, and drops the benches that weren't demonstrating anything. I've...
All comments addressed, ready for review. Will write up some notes for a review club soon.