Patch5

[lovesh]

• Align GBP polynomial degrees with the audit report published by Monero
• Include the size of each vector commitment in the transcript
• Include challenge generators in dlog gadget
• Fix bug in multi_select gadget
• Removed unused code
• Skip unneccesary terms from poly product
• Use JSF for scalar mult. in inner product
• Remove redundant curve check constraint
• Remove unneccessary check divisor creation
• Cache evaluations of curve poly along the interpolator
• Other misc. optimizations like removing allocations
• More tests, including mocking and some docs

[Copilot]

Pull request overview

This PR ("Patch5") is a broad maintenance/refactor pass spanning the relations, bulletproofs, ark-ec-divisors, and ark-dlog-gadget crates. The bulk of the changes turn previously-panicking gadget functions into Result-returning ones (new Error::NeedNonZeroSetSize variant, propagation through multi_select*, select*, single-level and curve-tree prover/verifier paths) and update the test suite accordingly. The bulletproofs prover and verifier are also reworked so the transcript binds the number and sizes of vector commitments, and the polynomial-degree arithmetic is factored into shared helpers (degree, t_poly_degree, transmitted_t_degree_indices). The divisor and barycentric code is reorganized and lightly re-documented; a new to_poly method, compute_modulus relocation, and normalize_x_coefficient move are the substantive moves.

Changes:

• Threaded Result/Error::NeedNonZeroSetSize through select / multi-select / curve-tree gadgets and bumped tests to .unwrap() the new results.
• Bound vector-commitment counts and sizes into the bulletproofs Fiat-Shamir transcript (prover, verifier, and dlog gadget) and refactored VerificationTuple::padded_n to return Result.
• Refactored divisor (compute_modulus/to_poly/normalize_x_coefficient relocation), barycentric weight computation, and commented-out (rather than removed) several utility polynomial types in bulletproofs/src/util.rs.

Reviewed changes

Copilot reviewed 33 out of 33 changed files in this pull request and generated 27 comments.

Show a summary per file

File	Description
relations/tests/curve_tree.rs, ped_comm_group_elems.rs, common/mod.rs, batched_curve_tree.rs	Tests updated for new Result-returning gadgets and added into_group() to get_leaf comparisons
relations/src/error.rs	New NeedNonZeroSetSize variant
relations/src/select.rs	Replaced assert!/panic-based gadgets with Result, added multi_select_public_set, factored construct_* helpers, added new tests
relations/src/single_level_select_and_rerandomize.rs	Functions now return Result<(), Error>
relations/src/{prover,verifier}.rs	? propagation for new Result returns
relations/src/curve_tree_{prover,verifier}.rs	Threaded Result through prover/verifier non-root processing; introduced result collection around rayon::join
relations/src/batched_*.rs	Same Result propagation for batched paths
relations/Cargo.toml, Cargo.toml	Added ark-poly workspace dependency
bulletproofs/src/r1cs/{prover,verifier/mod,verifier/batch,proof,mod}.rs	New transcript bindings, degree-helpers, padded_n returns Result, removed second-phase assertion
bulletproofs/src/util.rs	Commented-out (not deleted) VecPoly1, VecPoly3, Poly2, Poly6, sum_of_powers etc.; coeff renamed to coeff_mut
bulletproofs/tests/r1cs.rs, r1cs_vec.rs	Added randomized shuffle test, padded_n failure tests; removed deref on *n/*x (likely breaks compilation)
bulletproofs/generalized-bulletproofs.md	Replaced \color{} with \textcolor{}, spelling/formatting fixes
ark-ec-divisors/src/{lib,divisor,poly,barycentric,scalar_decomposition,curves/tests}.rs	Refactor: moved compute_modulus/to_poly/normalize_x_coefficient; reworked barycentric weights + evaluation; doc edits
ark-dlog-gadget/src/dlog.rs	Append additional challenge-point components and per-bit beta to transcript (proof-format change)

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.