Browse Source
We were basing the 2016 block timeout on the blockchain height that we had processed at the time completed the funding, which could lag considerably behind the network-wide blockheight. For example this could happen if we started with `--rescan` from a low height, taking some time to catch up. This means that we could end up forgetting channels even before reaching the network blockheight. This patch instead uses the headercount reported by the backend plugin if we aren't caught up yet. While the chances of this happening are still there, the window it might happen are much reduced, since headers can be synced in a couple of minutes. Reported-by: Riccardo Masutti Changelog-Changed: Funding timeout is now based on the header count reported by the bitcoin backend instead of our current blockheight which might be lower.bump-pyln-proto
Christian Decker
5 years ago
committed by
Rusty Russell
3 changed files with 23 additions and 1 deletions
Loading…
Reference in new issue