Closed MavenRain closed 6 years ago
Agreed. This would be a good change.
Closing this. People should not use recoverable signatures in basically any application -- their patent story is questionable and they cannot be batch-validated.
@apoelstra doesn't bitcoind
use it?
Yes, but only for historical reasons, in the signmessage API which will be replaced by BIP322 or something similar.
I understand that it may be necessary for Bitcoin-related projects such as libbitcoin. However we still want to ensure that people using the module know they are using something odd.
Maybe note that in the README?
@evoskuil what is your thoughts around this?
@apoelstra after a quick search, this doesn't seem to be a new thought https://bitcointalk.org/index.php?topic=940436.msg10306479#msg10306479 https://bitcoin.stackexchange.com/questions/75322/why-do-we-need-public-key-as-part-of-witness-in-p2wpkh-transactions
Has anyone determined if it is actually a problem? It seems to be guesses at this point.
@laanwj what about bitcoind
? Is it non-issue if the wallet support isn't compiled in?
Why not simply remove it?
Libbitcoin uses it to implement the message signing “standard” as an API and command line utility in bx. It doesn’t matter to us whether it’s enabled by default or otherwise - if it remains conditional we have to deal with it missing. More significantly we’ve always had to deal with lack of versioning on the library itself.
@evoskuil what about the patent story?
no opinion
This is required for building libbitcoin independently, but is currently disabled by default. Turning this default on will provide a better customer experience for developers attempting to quickly building their bitcoin development environment.