Merged
Conversation
RFC-0110 Deterministic BIGINT - Phases 1-3: Phase 1 - ADD/SUB/CMP: - BigInt struct with Vec<u64> limbs + sign - canonicalize(), is_zero(), bit_length() - magnitude_cmp(), compare() for signed comparison - bigint_add(), bigint_sub() with overflow checks Phase 2 - MUL: - bigint_mul() - schoolbook O(n²) multiplication - 128-bit intermediate arithmetic - Post-MUL canonicalization - Overflow check against MAX_BIGINT_BITS Phase 3 - DIV/MOD/SHL/SHR: - bigint_divmod() - binary long division - bigint_div(), bigint_mod() - bigint_shl() - left shift with overflow - bigint_shr() - right shift (arithmetic) Tests: 17 passing Clippy: clean Format: clean
…erification probe - Added TryFrom implementations for i64/u64/i128/u128 with proper MIN value handling - Added bigint_to_i128_bytes for i128 round-trip conversion - Added BigIntEncoding with serialize/deserialize for canonical 24-byte format - Added BigInt verification probe (56 entries) with SHA-256 Merkle tree - Added BIGINT_REFERENCE_MERKLE_ROOT constant for verification - Fixed clippy warnings: manual_div_ceil, needless_borrows - Fixed entry 52 value (MAX_U64 vs Max 4096-bit) - Added implementation fixes documentation in source All 115 tests pass, zero clippy warnings, Merkle root matches reference.
…icalization Critical fixes: - limb_sub: Fix borrow propagation using overflowing_sub chain - limb_mul: Fix carry propagation (replace |= with proper loop) - bigint_divmod: Replace buggy binary search with Knuth Algorithm D High severity fixes: - bigint_shr: Return ZERO for large shifts (not Err) - bigint_shl: Handle shift == 0 correctly - Serialization: Fix byte offsets (num_limbs at byte 4) - Add is_canonical() checks to all arithmetic functions Lower severity: - Add BigIntError::OutOfRange variant - Add debug_assert for non-empty limbs in BigInt::new() - Narrow #[allow(arithmetic_overflow)] scope in probe.rs Regression tests: 69 tests added, 2 previously slow division tests now enabled (Knuth Algorithm D is O(n) not O(quotient)).
Add comprehensive documentation to the BigIntProbeValue enum explaining when to use each variant. This prevents the bug where Int(MAX_U64) and Max encode to different bytes in the compact probe format. Key points: - Max encodes to ff ff ff ff ff ff ff ff (4096-bit sentinel) - Int(MAX_U64) encodes to hash-ref bytes (integer operand) - Trap encodes to de ad de ad de ad de ad (TRAP sentinel) - Int(TRAP) encodes to hash-ref bytes (integer operand) The fix log comment is now visible at the variant definition site, not just in the module docstring.
Critical bug fix: The original i128 cast `q_hat * v[i] as i128` could overflow when q_hat * v[i] > 2^127 (roughly 1 in 2^64 cases). Solution: Two-pass approach using pure u128 arithmetic: 1. Pass 1: Compute q_hat * v[] into temporary qv[] 2. Pass 2: Subtract qv[] from u[] with overflow tracking Also syncs RFC-0110 fixes: - Probe entry count: 64 → 56 - MUL: pre-check → post-check overflow validation
- FIXED: Entry 1 label (2^64 + 1 → 2^64) - matches Python/Rust reference - FIXED: Rule 4 DIV iteration count - now correctly states m+1 where m = dividend.len() - divisor.len() - FIXED: Removed unnecessary j=0 special case - standard D1 formula works with implicit r[-1] = 0 - Updated version to v2.13
- Version: 2.12 → 2.13 - Date: 2026-03-15 → 2026-03-16
bigint.rs: - Update doc comment to match RFC v2.13: "m+1 outer iterations where m = dividend.len() - divisor.len()" instead of "a_norm.limbs.len()" probe.rs: - Fix entry 1 description: "2^64 + 1" → "2^64" - Fix stale python_hashes: entries 52, 54, 55 - Add assert_eq!(mismatches, 0) to fail loudly on future regressions
- Add Display trait implementation for decimal and hex output
- Add LowerHex and UpperHex for {:#x} format
- Add FromStr trait for parsing decimal and hex strings
- Support 0x prefix for hex, -/+ prefix for signed numbers
- Add InvalidString error variant
- Mark slow decimal tests as #[ignore]
All phases complete: - Phase 1: Primitive conversions (i64, i128, u64, u128) - Phase 2: String conversions (Display, FromStr, hex support) - Phase 3: Serialization (BigIntEncoding wire format) - Phase 4: i128 round-trip conversion
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Pull Request Checklist
Type of Change
agent/*branch)Branch Strategy
This PR follows the CipherOcto branch strategy:
feat/*nextagent/*nextresearch/*nexthotfix/*mainnextmainCurrent PR:
<!-- source branch -->→<!-- target branch -->Description
Related Issues
Closes #(issue)
Testing
Performance Impact
Security Considerations
Additional Notes