Feat/adversarial tests#16
Merged
Merged
Conversation
87 passing + 6 ignored bug reproducers (marked B1-B3). Confirmed bugs: 1. B1: set_chunk doesn't mask unused high bits after write File: src/bit_string/impls_for_editing/impls_for_set.rs:26 Fix: add self.words.mask_unused_bits(self.bit_len) at end of set_chunk 2. B2: BitStr::count_ones includes source bits beyond view's end File: src/traits/bits_arith/funcs_for_count_ones.rs:19 Fix: mask bits[full_words] with low_mask(rem) before counting 3. B3: BitStr::find misses occurrences straddling unaligned→aligned word boundary File: src/bit_str/impls_for_matching/impls_for_find.rs:73 Fix: extend unaligned scan range to cover cross-boundary positions Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: DeepSeek AI <service@deepseek.com>
tests/adversarial.rs → entry point with shared helpers + mod declarations tests/adversarial/ tests_for_construction.rs — new, zeros, ones, repeat, from_words, from_str tests_for_access.rs — get, first, last, get_chunk tests_for_editing.rs — push, pop, insert, remove, set, set_chunk, truncate, clear tests_for_concat.rs — push_bit_string, insert_bit_string, split_off tests_for_slice_drain.rs — slice, slice_from, slice_until, drain_interval(_assign) tests_for_replace.rs — replace, replace_interval(_assign) tests_for_retain.rs — retain tests_for_bitwise.rs — not, and, or, xor tests_for_shift.rs — shl, shr tests_for_count.rs — count_ones/zeros, leading_*, trailing_* tests_for_matching.rs — matches_at, starts_with, ends_with, contains, find, rfind, strip tests_for_ord_hash.rs — Ord, PartialOrd, Eq, Hash tests_for_fmt.rs — Display, Debug tests_for_iter.rs — Iterator, DoubleEndedIterator tests_for_predicates.rs — any, all, is_all_zeros, is_all_ones, is_empty tests_for_bitstr.rs — clone invariants, BitStr views, slice chains tests_for_stress.rs — extend, stress, memory, set_chunk precision, invariant chain tests_for_bugs.rs — 7 ignored bug reproducers (B1, B2, B3) Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: DeepSeek AI <service@deepseek.com>
Bug 1 (set_chunk): Add mask_unused_bits() at end of set_chunk to restore the unused-bits-zeroed invariant after writing bits into the last word. - src/bit_string/impls_for_editing/impls_for_set.rs Bug 2 (count_ones): Mask the last partial word with low_mask(rem) before counting. The old code trusted the invariant but BitStr sub-views share the source's words where bits beyond the view's end may be non-zero. - src/traits/bits_arith/funcs_for_count_ones.rs Bug 3 (find/contains/rfind): Extend unaligned scan range to cover needles that straddle the unaligned→aligned word boundary. bits_equal_at already handles cross-boundary reads correctly. - src/bit_str/impls_for_matching/impls_for_find.rs All 93 adversarial tests now pass (was 86+7 ignored). Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: DeepSeek AI <service@deepseek.com>
Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: DeepSeek AI <service@deepseek.com>
Target both-sides-unaligned Eq/Ord/Hash, unaligned find/contains/rfind, structural vs semantic Eq, trailing_zeros/ones on unaligned views, get_chunk on unaligned views, strip on unaligned views, matches_at both misaligned, BitStr iter/predicates/display on unaligned views, and edge cases for retain/set_chunk/push-pop at word boundaries. Discovered B4: rfind misses the last occurrence on unaligned haystacks when the aligned portion has fewer than SMALL_WORDS words. In impls_for_find.rs the find_last_word call is gated behind aligned.len() >= SMALL_WORDS — when this fails the aligned portion is never scanned and rfind falls through to only scanning the first partial word. Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: DeepSeek AI <service@deepseek.com>
The rfind unaligned path gated find_last_word behind aligned.len() >= SMALL_WORDS. When the aligned portion was too small (e.g. 2 words on AVX2 needing 4), find_last_word was never called and only the first partial word was scanned — missing all occurrences in the aligned remainder. Fix: pull the SMALL_WORDS gate up to only control find_any_candidate (quick rejection). find_last_word is always called and handles small inputs via its own internal scalar fallback. Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: DeepSeek AI <service@deepseek.com>
Add cat() helper to adversarial.rs, distribute 34 round-2 tests: - tests_for_ord_hash.rs (+8: both-unaligned Eq/Ord/Hash, structural-vs-semantic) - tests_for_matching.rs (+9: unaligned find/rfind/contains, strip, matches_at) - tests_for_count.rs (+3: trailing_zeros/ones unaligned) - tests_for_access.rs (+2: get_chunk unaligned) - tests_for_retain.rs (+2: retain at word boundaries) - tests_for_editing.rs (+4: set_chunk edge cases, push/pop boundaries) - tests_for_iter.rs (+2: BitStr iter unaligned) - tests_for_predicates.rs (+3: predicates unaligned) - tests_for_fmt.rs (+1: BitStr Display unaligned) Delete tests_for_round2.rs — no separate round file needed. Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: DeepSeek AI <service@deepseek.com>
|
Codecov Report❌ Patch coverage is
📢 Thoughts on this report? Let us know! |
Merging this PR will degrade performance by 12.44%
Warning Please fix the performance issues or acknowledge them on CodSpeed. Performance Changes
Tip Investigate this regression by commenting Comparing 
Footnotes
|
The stdarch_aarch64_prefetch feature gate was copied from the library code but is not needed in the test binary — no prefetch intrinsics are used in tests. On stable Rust this causes a compile error. Co-authored-by: Claude <noreply@anthropic.com> Co-authored-by: DeepSeek AI <service@deepseek.com>
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
2 participants
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.
No description provided.