#4450 introduced a check for non-equivalent double spends and relaying, but for equivalent double spends it changed existing behaviour by still calling mempool.addUnchecked() on them. This breaks mempool consistency (detected using -checkmempool after a few days of running), which could be exploited to crash miner's nodes. #4484 seems to fix this.
-
sipa commented at 8:28 PM on July 12, 2014: member
- sipa cross-referenced this on Jul 18, 2014 from issue Reverted double spend relaying (for now) by jgarzik
-
laanwj commented at 10:49 AM on September 25, 2014: member
This problem no longer exists as the change has been reverted.
- laanwj closed this on Sep 25, 2014
- bitcoin locked this on Sep 8, 2021