Closed jnewbery closed 2 years ago
Thanks again to @Xekyo and @harding for review. I think the latest version addresses all of your comments.
My overall concern with this PR is that I think the level of technical detail here is good to document, but I worry that a focus on describing what will be rejected by the BIP125 implementation makes it seem like this is solution that's too hard to get right. It might be useful to focus on what someone attempting to create a replacement needs to do in the simple case (just use all of the original inputs for safety, add new confirmed inputs if necessary, and pay a higher feerate by at least 10 nBTC/vByte times a multiplier---Bitcoin Core uses 5x). Then let users know that this might not work due to descendant transactions, but that no node (at least no Bitcoin Core node) will penalize them for trying to relay a consensus-valid replacement that doesn't satisfy all of the BIP125 rules.
This detailed analysis is still quite useful, though, but perhaps it might better belong in an appendix or as a page on the Wiki or in the Bitcoin Core docs/ directory.
Thanks all for the review. I've force-pushed a commit addressing some of the comments.