BIP78: Allow mixed inputs Redux

[DanGould]

This picks up and supersedes #1244 that

• separates concerns into individual commits
• corrects fee verification for a proposal having mixed script inputs while ensuring additional fee indeed paid for input

I know @Kixunil is extremely busy to update these days, so I took the liberty to organize the changes we've discussed according to your wants here. @NicolasDorier I hope these changes satisfy your desire for them to be small and problem-focused while maintaining the spirit of the original.

---

Disallowing mixed inputs was based on incorrect assumption that no
wallet supports mixed inputs and thus mixed inputs imply PayJoin.
However there are at least three wallets supporting mixed inputs.
(Confirmed: Bitcoin Core, LND, Coinomi) Thus it makes sense to enable
mixed inputs to avoid a payjoin-specific fingerptint. To avoid
compatibility issues a grace period is suggested.

[murchandamus]

It seems to me that you are proposing a number of changes to BIP 78 and the biggest point of friction seems to be to get sign-off from that BIP’s champion. Have you considered alternatively proposing your own variant of BIP 78 with the changes you would like to see and building BIP 77 on top of that?

[DanGould]

Are you suggesting I compile the changes into an addendum document, or something else?

[murchandamus]

Sorry, I thought this was a PR with further changes, in addition to your prior attempt to modify BIP 78. At second glance I realize now that this is a second variant with a reduced scope to the prior. Nevermind, please ignore and carry on.

[Kixunil]

I'm slowly getting some time now but feel free to go ahead with this. I'd still like to cover the edge case I mentioned because it affects some software in the wild but that can be done in a separate PR.

And thanks @DanGould for picking this up!

[DanGould]

I'd still like to cover the edge case I mentioned because it affects some software in the wild but that can be done in a separate PR.

To be crystal clear, you're talking about transaction malleability with regard to output substitution. I think that's outside the scope of this pr but also worthy of amendment. That discussion is: https://github.com/bitcoin/bips/pull/1244/files#r1633516491

[NicolasDorier]

It means that only a single input can be covered by the contribution and the rest should be at the cost of the receiver.
Not a big deal, as I welcome that it makes things a little bit simpler while still breaking the heuristic.

1. GetVirtualSize can be removed as it is not used anymore.
2. Can you rename if (output.OriginalTxOut == feeOutput) to if (originalOutput.OriginalTxOut == feeOutput)

[NicolasDorier]

Ok @DanGould can you pickup this commit please: NicolasDorier@1cc129e

This basically just hardcode the maximum amount to pay for any input.

[murchandamus]

@NicolasDorier: It looks like the commit you requested was added. Could you please review this PR and state whether you support or reject these changes?

[NicolasDorier]

@murchandamus yes this is ready to merge

[jonatack]

Closes #1244.

[jonatack]

@DanGould As a follow-up, do you think these dates here and on line 110 above ought to be updated to provide a post-merge grace period?

[DanGould]

A client may still opt to disallow mixed inputs (or any payjoin proposal) by policy rather than by protocol like this one, so I don't think it's so critical that requires an explicit grace period after all, nor does the change in #1396. I think we can leave the grace period as-is and tackle the implementations one by one.

[jonatack]

Thanks! I also see that you are reaching out to implementations at the moment 👍