Add finalize to Psbt

[rustaceanrob]

Description

#580 does the initial work to adding finalize to Psbt. I would like to get this in for @andreasgriffin and revisit this later in bitcoin-ffi. I think #580 has the right idea, but I would like a more detailed system of why finalizing might have failed. Unfortunately, the finalize function using a Vec<E> of errors, which is not expressible by UniFFI types, so I created a new FinalizedPsbtResult type that contains either the original Psbt or the finalized Psbt, and any errors that may be associated with the finalize step.

I have thought about this for a while and I think this is the best approach, but open to any ideas.

Notes to the reviewers

Please confirm this works with your example Psbt @andreasgriffin

Changelog notice

• Add finalize method to Psbt

Checklists

All Submissions:

• I've signed all my commits
• I followed the contribution guidelines
• I ran cargo fmt and cargo clippy before committing

New Features:

• I've added tests for the new feature
• I've added docs for the new feature

Bugfixes:

• This pull request breaks the existing API
• I've added tests to reproduce the issue which are now passing
• I'm linking the issue being fixed by this PR

[andreasgriffin]

Thanks! Yes I can confirm that it finalizes the PSBT in #469 correctly.

And the could_finalize is super helpful!!!

[reez]

but I would like a more detailed system of why finalizing might have failed

Same, more details are great for things like this

Unfortunately, the finalize function using a Vec of errors, which is not expressible by UniFFI types, so I created a new FinalizedPsbtResult type that contains either the original Psbt or the finalized Psbt, and any errors that may be associated with the finalize step.

This makes sense to me.

Concept ACK c432e2d

Built and tested the swift bindings locally but didn't get to finish putting it thru a few different scenarios to test it out more (not a blocker on the PR, just ran out of time while taking a quick look at it tonight), did get this as expected when testing out a scenario where I expected couldFinalize to be false 👍:

finalized: FinalizedPsbtResult(psbt: BitcoinDevKit.Psbt, couldFinalize: false, errors: Optional([BitcoinDevKit.PsbtFinalizeError.InputError(reason: "Could not satisfy Tr descriptor", index: 0)]))

[thunderbiscuit]

I think this is a decent way to deal with our inability to offer the Rust API directly.

It reminds me that I meant to keep track of those small reshapes of the API somewhere. Not sure if it should just be a personal file on my local system or part of a doc/readme somewhere, but I'm starting one today!

Thanks for taking the time to think through this one @rustaceanrob.

ACK c432e2d.

[thunderbiscuit]

One thought I just had is that if this is not really agreed upon by other users of bitcoin-ffi, we'll end up with a breaking change if we migrate our Psbt type to the bitcoin-ffi type, or we'll have to just keep supporting our custom Psbt type.

[thunderbiscuit]

@rustaceanrob do you mind rebasing this? Ready to merge.

[rustaceanrob]

My branch is showing I am up to date