On Sun, Apr 26, 2015 at 02:20:04PM +0200, Jorge Timón wrote: > On Sun, Apr 26, 2015 at 1:35 PM, Jorge Timón wrote: > > And a new softfork rule could enforce that all new CTxIn set nHeight > > to the correct height in which its corresponding prevout got into the > > chain. > > That would remove the need for the TxOutputGetter param in > > bitcoinconsensus_verify_script, but unfortunately it is not reorg safe > > (apart from other ugly implementation details). > > Wait, wait, this can be made reorg-safe and more backards compatible. > The new validation rule at the tx validation level (currently in > main::CheckInputs()) would be So, seems to me that RCLTV opens up a whole rats nest of design decisions and compromises that CLTV doesn't. Yet CLTV itself is a big step forward, it's been implemented on Viacoin for the past few months with no issues found, and has an extremely simple and easy to audit implementation. I think I'm going to argue we implement it as-is in a soft-fork. Pieter Wuille's been working on a new way to handle soft-fork upgrades in the block nVersion field, so this would be a good opportunity to add something simple and well tested, and also make sure the new nVersion soft-fork mechanism works. Equally, doing both at the same time ensures we don't burn yet another version bit. -- 'peter'[:-1]@petertodd.org 00000000000000000e7980aab9c096c46e7f34c43a661c5cb2ea71525ebb8af7