oxfmt: Investigate perf regression by #18723 #18839
Description
I've discovered that performance has degraded due to changes made in #18723.
Since the processing itself is necessary, we need to investigate if there are any areas that can be improved.
Target: Storybook repository (mixed with embedded languages)
- 1 warmup runs, 3 benchmark runs
- Git reset before each run
Benchmark 1: oxfmt/dist/cli.js --config ./oxfmtrc.json ./data
Time (mean ± σ): 5.705 s ± 0.098 s [User: 12.973 s, System: 0.879 s]
Range (min … max): 5.610 s … 5.805 s 3 runs
Benchmark 2: oxfmt/dist-nocleanup/cli.js --config ./oxfmtrc.json ./data
Time (mean ± σ): 4.574 s ± 0.087 s [User: 27.031 s, System: 2.562 s]
Range (min … max): 4.489 s … 4.664 s 3 runs
Summary
oxfmt/dist-nocleanup/cli.js --config ./oxfmtrc.json ./data ran
1.25 ± 0.03 times faster than oxfmt/dist/cli.js --config ./oxfmtrc.json ./data
(dist-nocleanup = #18723 reverted version)