[cryptography/bls12381] Test Partial Signature Derivation #1082
Conversation
| match namespace { | ||
| Some(namespace) => hm.map(V::MESSAGE, &union_unique(namespace, msg)), | ||
| None => hm.map(V::MESSAGE, msg), | ||
| let hm = match namespace { |
There was a problem hiding this comment.
Switching to hash_message fixes an erroneous linting error.
There was a problem hiding this comment.
Pull Request Overview
Adds a new test for deriving missing BLS partial signatures and refactors message hashing to use hash_message.
- Replaces manual
Signature::zero+mapcalls withhash_messagein three loops. - Introduces
threshold_derive_missing_partialshelper andtest_threshold_signature_derive_partialstest.
Comments suppressed due to low confidence (2)
cryptography/src/bls12381/primitives/ops.rs:2367
- [nitpick] The parameter names
eval_at_xandi_xare ambiguous. Consider renaming them totarget_indexandbasis_indexrespectively to clarify their roles in the Lagrange coefficient calculation.
fn lagrange_coeff(eval_at_x: u32, i_x: u32, x_coords: &[u32]) -> Scalar {
cryptography/src/bls12381/primitives/ops.rs:2366
- [nitpick] Consider expanding this doc comment with the mathematical formula and explaining why
eval_at_x + 1is used before converting toScalar.
// Helper to compute the Lagrange basis polynomial l_i(x) evaluated at a specific point `eval_at_x`.
Codecov ReportAttention: Patch coverage is
@@ Coverage Diff @@
## main #1082 +/- ##
==========================================
+ Coverage 91.10% 91.11% +0.01%
==========================================
Files 191 191
Lines 54839 54919 +80
==========================================
+ Hits 49960 50040 +80
Misses 4879 4879
Continue to review full report in Codecov by Sentry.
🚀 New features to boost your workflow:
|
This test derives
n-tpartial signatures fromtprovided partial signatures (i.e. it tests "generic" signature recovery).Justification
While we never perform this operation in practice (at least for now), it [1] serves as a reminder that using partial signatures for accountability (when
>= texist) is unsound and [2] serves as a conformance test that our share derivation, partial signing, and msm logic is consistent.