perf: replace two-comparison bounds check with single unsigned cast

[fcostaoliveira]

Summary

• Replace counts_index < 0 || h->counts_len <= counts_index with (uint32_t)counts_index >= (uint32_t)h->counts_len in hdr_record_values and hdr_record_values_atomic
• For valid non-negative input, counts_index is always >= 0 by construction — the negative check is dead code
• One unsigned comparison vs two signed comparisons, reducing branch count on the hot path

Benchmark (hdr_record_value throughput, ops/sec)

	Baseline	Optimized	Delta
avg last 5 iters	261,584,518	319,067,198	+22.0%

Steps to reproduce

# Build baseline (main)
git clone https://github.com/HdrHistogram/HdrHistogram_c.git hdr_baseline
cd hdr_baseline && mkdir -p build && cd build
cmake .. -DCMAKE_BUILD_TYPE=Release -DHDR_HISTOGRAM_BUILD_PROGRAMS=ON
make hdr_histogram_perf -j$(nproc)
./test/hdr_histogram_perf        # note ops/sec from last 5 iterations

# Build this PR
cd ../../
git clone https://github.com/fcostaoliveira/HdrHistogram_c.git hdr_opt
cd hdr_opt && git checkout perf/opt5-unsigned-bounds-check
mkdir -p build && cd build
cmake .. -DCMAKE_BUILD_TYPE=Release -DHDR_HISTOGRAM_BUILD_PROGRAMS=ON
make hdr_histogram_perf -j$(nproc)
./test/hdr_histogram_perf        # compare ops/sec from last 5 iterations

The benchmark (test/hdr_histogram_perf.c) records 400M values in a tight loop and reports ops/sec per iteration. Use the last 5 of 100 iterations for steady-state throughput.

Notes

• Behavior-preserving: any value rejected before is still rejected
• Full test suite passes (5/5)

🤖 Generated with Claude Code