Bound the exceptional bits read to the validated length

[acoustid-bot]

m_exceptional_bits is sized from num_exceptional_bits, but UnpackInt5Array was given the whole remainder of the input as its source range:

m_exceptional_bits.resize(GetUnpackedInt5ArraySize(GetPackedInt5ArraySize(num_exceptional_bits)));
UnpackInt5Array(input.begin() + offset, input.end(), m_exceptional_bits.begin());

UnpackInt5Array writes one output byte per 5 input bits and stops when the source runs out, so the number of writes follows the input length rather than the size of the destination. Any input longer than the fingerprint actually needs therefore writes past the end of the vector.

The check just above only establishes a lower bound:

if (input.size() < offset + GetPackedInt5ArraySize(num_exceptional_bits)) {

so it does not constrain this.

The fix passes the end iterator at the length that check already validated. Trailing data is then ignored, which is what the surrounding code assumes anyway — m_bits is truncated to the number of values the header declares, and everything after the exceptional bits was already unused.

UnpackInt3Array a few lines earlier is not affected: there the destination is sized from the same input length that drives the loop, so the two agree.

Testing

Added OneItemOneBitExceptWithTrailingData — the existing OneItemOneBitExcept case with four trailing bytes, asserting it still decodes to the same value.

On master that test aborts in a Debug build with a past-the-end iterator from the libstdc++ debug checks. With the fix, the full suite passes (98 tests, -DFFT_LIB=kissfft -DCMAKE_BUILD_TYPE=Debug).

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: 2a071bda-bf92-4697-84da-e1a4bd08f1cb

📥 Commits

Reviewing files that changed from the base of the PR and between ab48115 and f7387ee.

📒 Files selected for processing (2)

• src/fingerprint_decompressor.cpp
• tests/test_fingerprint_decompressor.cpp

---

Walkthrough

Changes

Fingerprint decompression

Layer / File(s)	Summary
Bounded exceptional-bit decoding src/fingerprint_decompressor.cpp, tests/test_fingerprint_decompressor.cpp	Exceptional-bit unpacking now uses the exact packed segment size, and a test verifies trailing bytes do not affect the decoded value or algorithm output.

Estimated code review effort: 2 (Simple) | ~10 minutes

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly describes the main fix: limiting exceptional-bit reads to the validated length.
Description check	✅ Passed	The description matches the patch and test changes, explaining the bounds fix and the trailing-data regression test.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch bound-exceptional-bits-read

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}