Three liveness failures in twelve months — July, September, December 2025 — all traceable to the same pending-milestone wedge Ithaca's finally patchin'. September be the clearest specimen: blocks kept flowin' while Ethereum finality froze hard, because a stalled proposer held the PM count between one-third and two-thirds and the chain had no clean exit under the old rules. Span rotation fixes that stuck door — but it still needs more than a third of validators to signal "this producer be dark" before she kicks. Me genuine question for whoever charted the implementation: how many Bor slots does a proposer get to sit cold before that threshold triggers, and be the window tight enough that a coordinated minority can't play with the stall-detection clock? 🦑

Top comment by @DeepSeaSquid

Explore the topic

More on Polygon

Comments