fix: optimize read-inline-info #61

Job	Run time
benchmarks (bench/testdata/take_strings-aarch64, 3.26.0)	1h 9m 6s
benchmarks (echo Hello, World!, 3.26.0)	2m 42s
benchmarks (tar czf /tmp/news.tar.gz NEWS NEWS.old NEWS.older, 3.26.0)	3m 8s
benchmarks (ls -alh /etc, local)	1h 14m 6s
benchmarks (echo Hello, World!, 3.25.1)	2m 46s
benchmarks (ls -alh /etc, 3.25.1)	5m 32s
benchmarks (ls -alh /etc, 3.26.0)	2m 38s
benchmarks (bench/testdata/take_strings-aarch64, local)	7m 50s
benchmarks (tar czf /tmp/news.tar.gz NEWS NEWS.old NEWS.older, local)	30m 21s
benchmarks (python3 bench/fib.py, local)	29m 30s
benchmarks (python3 bench/fib.py, 3.25.1)	46m 13s
benchmarks (tar czf /tmp/news.tar.gz NEWS NEWS.old NEWS.older, 3.25.1)	3m 1s
benchmarks (echo Hello, World!, local)	6m 35s
benchmarks (python3 bench/fib.py, 3.26.0)	24m 35s
	5h 8m 3s

Provide feedback