Fix/chunk test suite stale num pubs - #20

…#280)

* chore: fix array out-of-bounds error with index validation

* Update mmr_guest.rs

* Update header-chain/src/mmr_guest.rs

Co-authored-by: Aaron Feickert <66188213+AaronFeickert@users.noreply.github.com>

---------

Co-authored-by: Aaron Feickert <66188213+AaronFeickert@users.noreply.github.com>

* WIP: Bigint configuration

* feat: bigint transforms fuzzing

* test: bigint limb transforms fuzzing
-add: macro for cleaning, adjusted test parameter, and idiomatic cleanup

* clean: separate files and remove unncessary struct param

* chore: cargo fmt

* Revert "chore: cargo fmt"

This reverts commit 0ea384a.

* chore: cargo formatting error, remove single line enforcement

* chore: move stack element verify script to fuzz library

BitVM#282)

* Fix connector_b_leaf_2_script test for proper Taproot script verification

* Update connector_b.rs

* Update connector_b.rs

* fix cargofmt

* chore: remove paste macro

I simply transform `paste! { X }` into `X`. If the diff for this commit
is large, then git is having an issue. The `paste!` macro is used
elsewhere in the codebase, so I won't remove the dependency.

* chore: clippy

cargo clippy --fix --lib -p bitvm

* chore: replace static with const

* chore: Inline

* refactor: Replace macros with functions

Functions are easier to read and tools like `rustfmt` and `clippy` have
an easier time parsing them.

* refactor: clarify endianness

* doc: winternitz

* refactor: Rename "blocks" to "digits"

This commit renames "blocks" to "digits" in function signatures and
struct members. I used the opportunity to clearly separate lengths
in units of digits from lengths in units of bits.

This commit touches a lot of code. The Parameter struct is technically
public and changed by this commit. However, I argue that no one should
manually set up his parameters. Instead, everyone should use the
high-level methods for wots-hash, wots-blake and wots256, which are not
affected by this commit.

I plan to make the Parameter struct private in a follow-up PR.

* doc: address comments

Only changes to docstrings and parameter names.
No changes to public API.

* fix: documentation

* chore: Mark potential footgun

* feat: New WOTS API

* test: Add WOTS test vectors

The test vectors are generated directly in Rust, so we can add new
vectors or modify existing vectors in the future.

* refactor: Adapt bitvm crate to new WOTS API

Removes the dependency on the modules wots_api and signing_winternitz.

* refactor: adapt bridge to new wots api

Removes the dependency on wots_api.
Removing signing_winternitz turned out to be too much work, as it
touches a lot of code. In particular, the connector outputs handle
public keys for different message lengths in the same vector.
The bridge code can be refactored in the future if needed.

* refactor: Move winternitz_hash into signing_winternitz

signing_winternitz is the only place that calls winternitz_hash,
so why not combine the modules?

BLAKE3 switched from 20-byte WOTS signatures to Wots16 a while ago,
so I use the opportunity to remove the code for 20-byte WOTS.

* chore: Remove wots_api.rs

No more code depends on this, so it is safe to remove.

* fix: Address comments

* refactor: WinternitzSecret constructors

Deprecate from_string constructors. Add from_bytes as a safer
alternative.

* fix: Address comments

* fix: address comments

* fix: cargo fmt

* feat: add tmul for fp2 with 4 window stack len

* feat: add fp2_w4 in ss_mul to reduce script size

* feat: add w4 version of lc4_keep_elements

* chore: specify vec capactiy

* doc: add docs to new functions

* removed unnecessary log asserts

* removed unnecessary if branch in tonelli shanks

* removed unused imports

* fix fq6_hinted_mul_keep_elements

* removed unused import

* fmt

…put if NUM_PUBS > 1 (BitVM#369)

* fixed num_pub>1 case

* fixed num_pubs>1 case

* add empty stack check

* fix comment

…-1 (BitVM#367)

…itVM#371)

* drop c left in the altstack

* remove if branch in utils_fq12_square since it doesnt occur

…BitVM#383)

* fix BitVM#365 Blake3 script inscorrect

fix BitVM#365 Blake3 script inscorrect for more than 1,024 bytes of input or more than 32 bytes of output

* add assert to blake3 compress to check final_rounds is 8

---------

Co-authored-by: lynndell | bitlayer <lynndell.zhong@bitlayer.ltd>

…nit_table (BitVM#380)

…, and add note in tmul about a case about hint (BitVM#386)

…VM#318)

* fix checksum length

* add warning for repeated hashes

* fix pushing digits

* fmt

* add hash length checks and comment verify_test_vector's temporarily since its not compatible with the new checksum length

* fix secret key generation

* remove BinarysearchVerifier bound check and add test to verify that it doesn't cause a vulnerability

* fix raw_witness_to_signature digit conversion

* add validation for hint being a correctly formed T type value to tmul

* check that hint y is valid in Fp254Impl::hinted_inv

* add check_validity to places where hints are used

* add Fq::check_validity and use it in necessary place

* add hint validity checks to chunk::taps_points:ops::utils_point_add_eval

* chore: rename check validity functions for cherry-picking optimized version

* feat: add check_validity function to Fp254

* chore: incorperate cherry-picked version

* opt: optimize copies for hint verifications with the cherry-picked verification

* fmt

---------

Co-authored-by: just-erray <c.e.aslan37@gmail.com>

* fix fp254impl issues

* fmt

* fmt

* add back necessary conflicted functions

* clear part1 report

* add assertion to push_u32_le

* fmt

* remove debug prints and improve test for limb_shr1_carry

* fmt

* limb_to_bits functions assert num_bits>0

* change G1Affine::roll, G2Affine::roll, G2Affine::copy parameters to in terms of Fq, and fix tests for bigint change

* fmt

* assert LIMB_SIZE < 31 in BigIntImpl

* restore bigint/mul.rs because it is fixed in another pr

* fmt

* apply parameter change to double_prevent_overflow_keep_element too

* feat: add check_validity function to Fp254

* feat: add validity checks to some functions in taps_ext_miller.rs as samples

* add validity check to taps_mul

* feat: add is_valid function to Fp254

* fix-feat: fix validity checks for precompute_p and complete validity checks for taps_ext_miller.rs

* feat: fix and add validity checks in taps_msm.rs

* fix: a validity check in taps_points_ops.rs

* fix-feat: correct and add validity checks

* fix: correct tests for test_point_ops_and_multiply_line_evals_step_1 tests according to the changed form of the input

* fix: add validity checks for G1Acc in chunk_msm

* opt: remove unnecessary q4 validity check

* fix: validity checks of taps_ext_miller::chunk_precompute_p_from_hash

* feat: add functions to analyze max stack usage of chunks

* fix: some typos and delete leftovers

---------

Co-authored-by: Hakkush-07 <hakan.karakus@std.bogazici.edu.tr>

* create a version of G1Affine::hinted_check_add that the caller must use with points that are not t==q nor t==-q

* optimize G1Affine::hinted_check_add and add TODO comments for later

* rearrange wrong if checks in G1Affine::hinted_check_add

* add validity checks to hints of new G1Affine::hinted_check_add

* fix hint validity check and roll in G1Affine::hinted_check_add

Sync BitVM updates

…etwork#12)

* remove length-prefix when get groth16 pubin

* fix guest inputs verify script

generate_segments_using_mock_vk_and_mock_proof moved to
api_compiletime_utils in an earlier refactor; this test module's import
was never updated, so `cargo test -p bitvm chunk::` has been failing to
compile outright.

With this fix applied, the suite compiles and reveals 7 further test
failures (chunk::api::test::full_e2e_execution, test_largest_chunks,
chunk::api_runtime_utils::test::test_runtime_execution_looped,
chunk::g16_runner_core::test::{test_groth16,test_verify_pairing},
chunk::taps_msm::test::{test_tap_msm_valid_inputs,test_tap_msm_invalid_inputs_scalar_not_fr}),
all tracing to one shared hardcoded test fixture (vk_bytes/proof_bytes/
scalar) built for a 1-public-input verifying key, while NUM_PUBS has been
2 since the Nov 21 2025 merge-conflict-resolution commit 7b31dfd. That
mismatch is a separate, still-open issue (not fixed here — needs a
regenerated 2-public-input fixture) filed alongside this PR.

Not believed to be exploitable in production: the live Ziren integration
test in goat::disprove_scripts (test_verify_guest_pubin_ziren, using a
real 2-element ziren/public_inputs.bin) passes cleanly. The risk is pure
test-suite/CI coverage: this module's regression suite for the chunked
Groth16 pairing verifier has evidently not produced a passing run in
~9 months.