add Signer interface for Taro virtual transactions #116
Conversation
I don't think they actually need to, given the unit tests can just use a canned private key signer (or maybe that's what you mean?). Similar to this: https://github.com/lightninglabs/taro/blob/main/asset/asset.go#L330-L334
Which check?
I think we'd keep this one to make a mock signer that just uses the raw private key for unit tests. |
jharveyb
force-pushed
the
signer-interface
branch
from
0ca257d
to
3670cc8
Compare
|
Rebased, updated to address all feedback, also added more plumbing for merge / multi-input transfers. |
| // descriptor and TX. | ||
|
|
||
| // NOTE: We currently assume the signature requested is for the | ||
| // BIP 86 spending type. |
There was a problem hiding this comment.
Seems fine for the scope of our current milestone, but isn't this mainly dependent on the SIgnMethod in the sign desc?
So then it's up to the caller to also provide the witness script (for the leaf being revealed/spent).
There was a problem hiding this comment.
Yes, the interface itself doesn't prevent non-BIP86 use. This assumption is really enforced earlier on in SignTaprootKeySpend, there isn't currently a way to provide a witness script at all.
| return nil, err | ||
| } | ||
|
|
||
| validatedAsset.PrevWitnesses[idx].TxWitness = *newWitness |
dstadulis
changed the title
Spun off from send state machine, builds on #115, only last commit is new.
In this PR, we introduce a
Signerinterface responsible for signing over Taro virtual TXs. This is needed to connect the sending state machine to a backing lnd client, and not require direct access to private keys.I've added an equality check between the mock Signer and the old 'direct' signing method, and the existing unit tests pass, so AFAICT this interface + mock logic works. Requesting feedback to make sure that isn't a fluke + general improvements.
Outstanding tasks before pulling out of draft: