perf(parser): remove -> Result<T> from all parsing methods

[Boshen]

All credits to @overlookmotel

In parser, optimize for common case that file parses successfully.

Instead of parse_* methods returning Results which are checked with ? in the caller, always return a valid node T.

In case of a fatal parsing error:

• Record the error in fatal_error property of ParserImpl.
• Also record current length of errors.
• Fast forward the lexer to EOF, so parsing will come to an end swiftly.
• Create a dummy node of the required type and return it.
• At end of parsing, if fatal_error.is_some(), then there was a fatal error. In that case:

Truncate errors to the length it was at time of the fatal error (because more errors may occur as the parser finds it's unexpectedly at EOF).
Add fatal_error to errors.
i.e. all parse methods never fail, they just return nonsense if there's a fatal error. This removes all the branches implicit in ?.

---

Wall clock benchmark:

oxc  main ❯ hyperfine './target/release/examples/parser-new ./target/typescript.js' './target/release/examples/parser ./target/typescript.js'
Benchmark 1: ./target/release/examples/parser-new ./target/typescript.js
  Time (mean ± σ):      42.7 ms ±   2.1 ms    [User: 36.7 ms, System: 5.3 ms]
  Range (min … max):    41.6 ms …  59.6 ms    68 runs

  Warning: Statistical outliers were detected. Consider re-running this benchmark on a quiet system without any interferences from other programs. It might help to use the '--warmup' or '--prepare' options.

Benchmark 2: ./target/release/examples/parser ./target/typescript.js
  Time (mean ± σ):      44.2 ms ±   0.7 ms    [User: 38.3 ms, System: 5.4 ms]
  Range (min … max):    43.7 ms …  48.6 ms    64 runs

Summary
  ./target/release/examples/parser-new ./target/typescript.js ran
    1.04 ± 0.05 times faster than ./target/release/examples/parser ./target/typescript.js

/usr/bin/time -al:

old: 517424069  instructions retired
new: 492334561  instructions retired

uses 5% less CPU instructions.

---

Testing:

• passed https://github.com/oxc-project/monitor-oxc
• passed Rolldown
• Did some fuzzing with https://github.com/oxc-project/fuzz-oxc, found no crashes.

[Boshen]

• refactor(coverage): add a visitor for checking ast nodes #10590
• perf(parser): remove -> Result<T> from all parsing methods #10588 👈 (View in Graphite)
• feat(parser): fast forward lexer to EOF if errors are encountered #10579
• main

---

How to use the Graphite Merge Queue

Add either label to this PR to merge it via the merge queue:

• 0-merge - adds this PR to the back of the merge queue
• hotfix - for urgent hot fixes, skip the queue and merge this PR next

You must have a Graphite account in order to use the merge queue. Sign up using this link.

An organization admin has enabled the Graphite Merge Queue in this repository.

Please do not merge from GitHub as this will restart CI on PRs being processed by the merge queue.

This stack of pull requests is managed by Graphite. Learn more about stacking.

[codspeed-hq]

CodSpeed Instrumentation Performance Report

Merging #10588 will improve performances by 7.9%

_{Comparing 04-24-perf_parser_remove_result_ (4f56b2c) with main (e228840)}

Summary

⚡ 5 improvements
✅ 31 untouched benchmarks

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
⚡	parser[RadixUIAdoptionSection.jsx]	93.5 µs	89.3 µs	+4.7%
⚡	parser[antd.js]	114.2 ms	106.3 ms	+7.43%
⚡	parser[cal.com.tsx]	30.3 ms	28.9 ms	+4.93%
⚡	parser[checker.ts]	57.5 ms	54 ms	+6.52%
⚡	parser[pdf.mjs]	18.3 ms	17 ms	+7.9%

[graphite-app]

Merge activity

• Apr 24, 10:28 AM EDT: Boshen added this pull request to the Graphite merge queue.
• Apr 24, 10:38 AM EDT: Merged by the Graphite merge queue.