> That is easy to change; I'll submit a pull request.

That's certainly a useful improvement. It won't help the existing userbase
though - assuming CHECKLOCKTIMEVERIFY is to go in to the next major
release. If there's going to be an intermediate release (6 months?) which
lays the groundwork for future rule changes, it helps more.

It would be good if getblocktemplate was updated at the same time to serve
errors if the fork warning is active. I'd hope miners have some way to
automatically handle IBD/getting forked off the chain, but I guess some
(newer) pools might not, and refusing to serve work should be the safest
option that shuts them down.

I don't have any opinion on the hard- versus soft- fork debate. I think
> either can work.

P2SH was a soft fork and the sky did not fall, but miners did lose money
and waste electricity mining blocks on the wrong side of the chain:


Presumably they didn't notice for longer because it looked like a run of
unusually bad orphaning luck. It seems safer to have a clean fork, with
alerts telling people during the lockin period before new rule enforcement
starts (and possibly automated termination if there's no upgrade by the
flag day?). Miners who ignore it would still risk losing money, but
merchants who wait for a block at least would not be at risk.

One open question is how can you actually trigger a hard fork? Coinbase
scriptSigs are not executed, so putting some ignored but failing opcode
sequence there wouldn't work. One possibility would be to have a special
invalid tx in the block that marks the start of new rule enforcement. New
nodes would know to ignore it. But this risks corrupting block explorers.
Alternatively the coinbase outpoint structure could have its hash set to 1
instead of 0.
Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer
Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports
Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper
Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer
Bitcoin-development mailing list

Reply via email to