Fix integer underflow in BITCOUNT ... BIT byte-boundary mask

[vazois]

Summary

Fixes an integer underflow in BITCOUNT ... BIT (BitmapManagerBitCount.cs) that caused non-zero counts to be returned for unset bits whenever a BIT-mode range ended on a byte boundary (endOffset % 8 == 7).

Root cause

BitIndexCount(byte* value, long startOffset, long endOffset) computed the inclusive right bit index as:

int rightBitIndex = (int)((endOffset + 1) & 7);   // endOffset=7 -> 0 (wrong)

When this wraps to 0, the inner BitIndexCount(byte, start, end) builds mask = (byte)((1<<0) - (1<<start)), a wrapping byte-mask underflow (e.g. 0x81). AND-ed with the bit-reversed payload this reports set bits that are not set. For byte 0x80, BITCOUNT key 7 7 BIT returned 1 even though GETBIT key 7 == 0.

Fix

Use the same formula already used (correctly) in BitmapManagerBitPos.cs:

int rightBitIndex = (int)(endOffset & 7) + 1;      // endOffset=7 -> 8 (correct)

Tests

• Added BitmapBitCountBitBoundaryUnderflowTest covering single-bit and boundary-crossing BIT ranges on byte 0x80.
• Updated BitmapBitCountLongBitOffsetParsingTest: its previous expectation of 8 was an artifact of the underflow. Under Garnet''s modulo-wrapping negative-offset semantics the clamped start wraps to bit index 1, so bits 1..7 of 0xFF = 7.
• All BITCOUNT/BITPOS tests pass.

[Copilot]

🟢 Ready to approve

The fix is narrowly scoped, matches the described root cause, and is covered by a focused regression test plus an updated expectation aligned to the documented offset semantics.

This review doesn't count toward merge requirements. Sign up for the private preview to control whether Copilot approvals count.

Pull request overview

This PR fixes a correctness bug in Garnet’s bitmap BITCOUNT ... BIT implementation (BitmapManagerBitCount.cs) where ranges ending exactly on a byte boundary could return a non-zero count for bits that are actually unset (due to an integer underflow in the boundary mask calculation).

Changes:

• Fixes the inclusive “right bit index” calculation for BIT-mode ranges that end on a byte boundary.
• Adds a regression test validating single-bit and boundary-ending BIT ranges on a 0x80 payload.
• Updates a long negative-offset parsing expectation that previously passed only because of the underflow behavior.

File summaries

File	Description
test/standalone/Garnet.test.complexstring/GarnetBitmapTests.cs	Adds a targeted regression test for the byte-boundary underflow and updates the long negative-offset BITCOUNT expectation.
libs/server/Resp/Bitmap/BitmapManagerBitCount.cs	Corrects rightBitIndex computation to avoid wrapping to 0 on endOffset % 8 == 7, preventing mask underflow.

Review details

• Files reviewed: 2/2 changed files
• Comments generated: 0
• Review effort level: Low

We're testing this review assessment. Please use 👍 or 👎 to tell us if it's correct.