rm-hull / jasengo / 30312935917

78%

main: 78%

Build Type

Pull #28

github

Committed by rm-hull

Commit Message

perf(parser): optimize runeReader.Read() hot path

Optimize buffer management in parser/reader.go to reduce per-Read() overhead:

- Pre-compute chunkSize as struct field (avoids float64 on every Read())
- Cache bufferEnd to avoid interface dispatch on hot path
- Inline advanceLocation to eliminate function call overhead
- Use type switches for buffer.Read/Write to enable compiler devirtualization

Benchmark results:
- BenchmarkReaderRead: ~112,000 -> ~82,777 ns/op (~26% faster)
- BenchmarkReaderReadWithLimit: ~167,000 -> ~129,062 ns/op (~23% faster)

All existing functionality preserved: rollback, slice, checkpoint, remaining.
Compatible with both RingBuffer and UnboundedBuffer.

Pull Request Pull Request #28: perf(parser): optimize runeReader.Read() hot path

Coverage Stats

28 of 45 new or added lines in 1 file covered. (62.22%)

19 existing lines in 1 file now uncovered.

482 of 616 relevant lines covered (78.25%)

161379.49 hits per line