feat: Add Input for PSBT #846

ItoroD · 2025-09-03T17:31:20Z

Description

Since signing will soon be done with the PSBT type, I think its a good time to expose some of the other PSBT fields like Input.

This will allow easier access to the input in a PSBT. It might be specifically important because the sign method of PSBT can do interesting stuff like returning a map of input index and keys used to sign it.

I would like to take a pause here and see how this is looking on the bindings side before I go any further.

Changelog notice

Add input type for 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

thunderbiscuit · 2025-09-11T13:18:19Z

Anything else you'd like to add to this PR or should we start reviewing now? No rush, just want to know where it's at.

ItoroD · 2025-09-11T18:30:36Z

Anything else you'd like to add to this PR or should we start reviewing now? No rush, just want to know where it's at.

A few more fields to add before this is good for review

ItoroD · 2025-09-15T16:29:38Z

This is ready for review

bdk-ffi/src/bitcoin.rs

thunderbiscuit

It's looking good!

I think a few of the Vec could be simply HashMap<K, V> as in the rust-bitcoin codebase, unless there is something I'm missing.

It's a lot of conversion between simple types, so I'm wondering if stealing one or two of the rust-bitcoin tests on this type would be good as well? I haven't looked at their tests so don't take that as a requirement. I don't know if they actually do a lot of tests just on this Input type.

thunderbiscuit · 2025-09-25T21:23:00Z

bdk-ffi/src/bitcoin.rs

+pub struct PartialSig {
+    pub pubkey: String,
+    pub signature: Vec<u8>,
+}


Since the rust-bitcoin type for this field is a map, I think we could use the uniffi equivalent here instead of a vector of structs. Let me know if you've tried this and it doesn't work for some reason!

Yes, you are correct. I struggled with trying to use BTreeMap in a record this made me think maps are not just one of the Uniffi compatible type. HashMap is definitely closer and it works. I just checked. I did

pub bip32_derivation: HashMap<String, KeySource>

instead of

pub bip32_derivation: Vec<Bip32Derivation>

where keysource is

#[derive(Clone, Debug, uniffi::Record)] pub struct KeySource { pub fingerprint: String, pub path: String, }

I was thinking about this again, the only thing about collecting these into Hashmap is that we loose the ordering that the BTreeMap gave from the actual rust-bitcoin types. Should we care about the ordering ?

Yeah I wondered that too. But looking at the fields, I do think the keys identify everything you need, and so the ordering doesn't matter. If anything devs would need an extra line of lookup to find the associated data instead of relying on the ordering, but even then I don't know if that's even how people use the BTreeMap for these fields. My hunch would be to use the HashMap.

Alright, I will update. Thanks.

thunderbiscuit · 2025-09-25T21:30:26Z

bdk-ffi/src/bitcoin.rs

+pub struct Bip32Derivation {
+    /// hex-encoded public key (serialized)
+    pub pubkey: String,
+    /// master key fingerprint
+    pub fingerprint: String,
+    /// derivation path segments
+    pub path: String,
+}


Same as above. We might be able to use a map instead of a Vec of this struct.

Thanks for pointing these out. BTreeMap not working threw me off. But I see now

thunderbiscuit · 2025-09-25T21:32:31Z

bdk-ffi/src/bitcoin.rs

+    pub ripemd160_preimages: Vec<PreimagePair>,
+    /// SHA256 hash to preimage map.
+    pub sha256_preimages: Vec<PreimagePair>,
+    /// HASH160 hash to preimage map.
+    pub hash160_preimages: Vec<PreimagePair>,
+    /// HASH256 hash to preimage map.
+    pub hash256_preimages: Vec<PreimagePair>,


Same for these.

ItoroD · 2025-09-26T11:28:40Z

It's looking good!

Thanks

It's a lot of conversion between simple types, so I'm wondering if stealing one or two of the rust-bitcoin tests on this type would be good as well? I haven't looked at their tests so don't take that as a requirement. I don't know if they actually do a lot of tests just on this Input type.

They do have a few test that concern input. Though they are not direct tests. The intention of some of the test are to validate correctness of an action performed, like signing. See here I am sure I can squeeze out a few good test like this from there. Will be a great idea to have them!

An additional thing I will have to perform is ensure whatever Psbt sample I create/use has values for the fields I will be checking/testing. Might take a while to do that for all fields. (Simulating different Psbt states, signed, unsigned, partially signed, finalized ). I could start with a few.

ItoroD · 2025-10-06T12:46:40Z

Just an update on this. I will be pushing new changes today with some tests as well

thunderbiscuit · 2025-10-09T18:16:55Z

Mind rebasing this @ItoroD? I think it's ready to go.

thunderbiscuit · 2025-10-09T18:28:35Z

bdk-ffi/src/bitcoin.rs

+#[derive(Clone, Debug, uniffi::Record)]
+pub struct PartialSig {


Sorry I think this type can now be removed.

This allows easier access to the inputs of our PSBT type

thunderbiscuit

My bad I went over it too quickly earlier.

thunderbiscuit · 2025-10-09T18:30:43Z

bdk-ffi/src/bitcoin.rs

+    pub tap_script_sigs: Vec<TapScriptSig>,
+    /// Map of Control blocks to Script version pair.
+    pub tap_scripts: Vec<TapScriptEntry>,
+    /// Map of tap root x only keys to origin info and leaf hashes contained in it.
+    pub tap_key_origins: Vec<TapKeyOrigin>,


These 3 types can be made HashMaps as well.

yh no worries. I had actually done these locally but i though you wanted them excluded intentionally because of the nature of the key.

tap_script_sigs is a concatenation of xonlypubkey and leafhash
<xonlypubkey>|<leafhash>

Yeah I'd keep the tuple (XOnlyPublicKey, TapLeafHash) as a struct similar to what you have above, but use that as my key on the HashMap, akin to the BTreeMap they have in Rust.

thunderbiscuit · 2025-10-09T18:31:02Z

bdk-ffi/src/bitcoin.rs

+    pub proprietary: Vec<ProprietaryKeyValuePair>,
+    /// Unknown key-value pairs for this input.
+    pub unknown: Vec<KeyValuePair>,


And these last two as well.

github-project-automation bot added this to BDK-Bindings Sep 3, 2025

ItoroD marked this pull request as draft September 3, 2025 18:46

thunderbiscuit self-requested a review September 11, 2025 13:18

ItoroD force-pushed the expose-psbt-input branch from 9e2a1b5 to 420f0ff Compare September 15, 2025 16:15

ItoroD marked this pull request as ready for review September 15, 2025 16:29