Author: Russell O'Connor 2019-03-13 01:38:44
Published on: 2019-03-13T01:38:44+00:00
In an email exchange between developers, Matt Corallo explains that removing OP_CODESEPARATOR simplifies the theoretical operation of checksig operations, and thus simplifies analysis of future changes. However, another developer disagrees, saying that simplification is not a sufficient reason to make changes to consensus rules nor remove usable op codes. The developer argues that avoiding flushing the sigcache is not critical enough to justify risky consensus changes. They also argue that no one has spoken up about the issue of invalid public keys on this mailing list, so why would people have spoken up about OP_CODESEPARATOR affecting them? The two developers discuss possible solutions, such as counting the maximum number of OP_CODESEPARATORS that can be executed through the various possible OP_IF branches. It is noted that alternating uses of OP_CODESEPARATOR and CheckSig operations are problematic and should be counted as pairs rather than just OP_CODESEPARATORS.
Updated on: 2023-06-13T17:30:01.199518+00:00