perf(regular_expression): skip capturing-group pre-parse when pattern has no `(` by linyiru · Pull Request #23908 · oxc-project/oxc

linyiru · 2026-06-28T20:57:37Z

What

PatternParser::parse always runs a full pre-pass over the whole pattern to count
capturing groups and collect named-group names before the real parse. The existing
comment already flags this:

// NOTE: It means that this perform 2 loops for every cases.
// - Cons: 1st pass is completely useless if the pattern does not contain any capturing groups
// We may re-consider this if we need more performance rather than simplicity.

A pattern with no ( can have no capturing group, no named group and no duplicate
name, so the pre-parse results are exactly State::new's defaults — only
named_capture_groups differs, and for a group-free pattern it depends solely on the
u/v flag. This detects the absence of ( with a cheap scan over the
already-decoded units (Reader::contains) and skips the pre-parse and its
allocations (the AlternativeTracker vec, the per-named-group vecs, the group-name
FxHashSet) in that case.

A ( that turns out to be escaped (\() or non-capturing ((?:…), lookarounds)
still takes the existing slow path — contains('(') is a conservative
over-approximation, so it is only ever skipped when the pattern is provably
group-free.

Correctness

Output is byte-identical. I verified this with a differential harness that parses 60
patterns with both the old and new code and compares the resulting AST / error
exactly. The corpus deliberately covers the tricky equivalences:

backreferences in group-free patterns: \1, \1\2\3, a\1b (legacy octal in
Annex B) — both with and without the u flag;
\k<a> in a group-free pattern with "", "u", "v" flags (the only case where
named_capture_groups is non-default);
a ( that must still take the slow path: \(, [(], (?:…), lookarounds,
named groups, duplicate-name errors ((?<a>)(?<a>)), unterminated (.

All 60 cases produce identical output, and cargo test -p oxc_regular_expression
passes.

Measurement

Local microbench over 15 realistic paren-free regexes (release):

	ns/parse
before	~258 ns
after	~230 ns

~11% faster for paren-free patterns, which are the large majority of real-world
regexes. The change removes a whole code-point pass plus its allocations for those
patterns.

The allocation snapshot (cargo allocs) confirms the reduction on real files — the
allocation count drops while total bytes are unchanged:

file	allocs before → after
antd.js	3720 → 3661
kitchen-sink.tsx	2108 → 2058
pdf.mjs	356 → 337
checker.ts	1825 → 1818
App.tsx	364 → 360

Honest note on CodSpeed: the bench corpus is regex-light (regex parsing is well under
2% of parse time on the heaviest bench file and ~0% on the rest), so the CodSpeed
comment will likely show no change. The win shows up on regex-heavy inputs (e.g.
moment.js / vue.js, where regex parsing is ~7–10% of parse). This PR is offered in
response to the maintainer note quoted above, framed as removing documented useless
work rather than as a benchmark mover.

Disclosure: developed with AI assistance (Claude), reviewed and verified by the author.

codspeed-hq · 2026-06-28T21:06:30Z

Merging this PR will not alter performance

✅ 62 untouched benchmarks
⏩ 9 skipped benchmarks¹

_{Comparing linyiru:perf/regex-skip-prescan (aaf5745) with main (7cb85c4)}

9 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

… has no `(` `PatternParser::parse` always ran a full pre-pass over the pattern to count capturing groups and collect named-group names. As the existing comment notes, that pass "is completely useless if the pattern does not contain any capturing groups". A pattern with no `(` has no capturing group, no named group and no duplicate name, so the pre-parse results are exactly `State::new`'s defaults — only `named_capture_groups` differs, and for a group-free pattern it depends solely on the `u`/`v` flag. Detect the absence of `(` with a cheap scan over the already-decoded units (`Reader::contains`) and skip the pre-parse and its allocations in that case. A `(` that turns out to be escaped or non-capturing still takes the existing slow path, so behavior is preserved. ~11% faster for paren-free regexes in a local microbench; output is byte-identical (verified by a differential AST/error check over 60 patterns covering backreferences, `\k<name>`, escaped parens, character classes, named/duplicate groups and `u`/`v` flags).

leaysgur

Thanks!

### 💥 BREAKING CHANGES - 94fbacb ast: [**BREAKING**] Only export `AstBuilder` and `NONE` in `builder` module (#23876) (overlookmotel) - 8de5122 ecmascript: [**BREAKING**] Switch to new `AstBuilder` (#23834) (overlookmotel) - dc0ef38 transformer: [**BREAKING**] Switch to new `AstBuilder` (#23831) (overlookmotel) - 88f4455 str: [**BREAKING**] `Str` and `Ident` methods take `&GetAllocator` (#23781) (overlookmotel) - 36009dd allocator: [**BREAKING**] `GetAllocator::allocator` take `&self` (#23676) (overlookmotel) - bd74f9d allocator: [**BREAKING**] Rename `AllocatorAccessor` trait to `GetAllocator` (#23675) (overlookmotel) ### 🚀 Features - 326fe25 transformer_plugins: Support `typeof` `define` keys (#23605) (Alexander Lichter) - f2091b3 ast: Unify old and new `AstBuilder`s (#23875) (overlookmotel) - cd1fd12 codegen: Expose `Codegen::print_string` API (#23785) (camc314) - 785461b ast: Add custom builder methods to AST types (#23651) (overlookmotel) - 05d1357 ast: Add AST creation methods to AST types (#23650) (overlookmotel) - 2580eda str: Add `Str::from_str_in` and `Ident::from_str_in` methods (#23767) (overlookmotel) - 6883fcf minifier: Fold write-once falsy var to false in boolean context (#23540) (Dunqing) - fcbf993 allocator: Add `Vec::from_value_in` method (#23718) (overlookmotel) - 989ddb7 allocator: Add `Vec::from_box_in` method (#23717) (overlookmotel) - 9d1aa7f allocator: Improve `PartialEq` for `Vec` (#23716) (overlookmotel) ### 🐛 Bug Fixes - da0e5bf minifier: Don't reorder a closed-over TDZ read when inlining a var (#23771) (Dunqing) - 0b3021f allocator: Remove `Vec::from_box_in` (#23873) (overlookmotel) - 0ab64ec ast: Silence deprecation warnings within files defining deprecated `AstBuilder` methods (#23889) (overlookmotel) - 8c07cad all: Enable `disable_old_builder` Cargo feature for `oxc_ast` crate in tests (#23888) (overlookmotel) - 3800f01 ast: Legacy `AstBuilder` methods take `self` not `&self` (#23891) (overlookmotel) - 869ac20 semantic/cfg: Connect for update exit to loop test (#23791) (camc314) - d3e92d5 semantic/cfg: Connect while branches from condition exit (#23790) (camc314) - 025045d ast: `ExportNamedDeclaration` plain builder methods return boxed nodes (#23783) (overlookmotel) - 7537c58 ast: Fix name of `AstBuilder` method for `Expression::V8IntrinsicExpression` (#23766) (overlookmotel) - 3f574f5 traverse: Fix unsoundness in `Traverse` walk functions (#23745) (overlookmotel) - 585760f parser: String in AST reference arena (#23721) (overlookmotel) - 7231d55 allocator: Fix unsound lifetime extension in `Box::new_in` (#23685) (overlookmotel) ### ⚡ Performance - d5c916a semantic: Flatten hoisting_variables to avoid per-scope map allocation (#23927) (Lawrence Lin) - e71609d minifier: Bail member-expr folding before the side-effect walk (#23924) (Lawrence Lin) - e1f89ab minifier: Reduce string allocations folding addition (#23846) (overlookmotel) - 9f6ee3b isolated-declarations: Pool scope maps to avoid per-scope alloc/rehash (#23761) (Boshen) - 0b07c4c semantic: Avoid heap alloc for catch-clause binding ids (#23911) (Lawrence Lin) - c5eef8b regular_expression: Skip capturing-group pre-parse when pattern has no `(` (#23908) (Lawrence Lin) - b4f5b4b isolated_declarations: Remove redundant clone of formal parameter pattern (#23912) (Lawrence Lin) - 53d083f isolated_declarations: Use `TakeIn` not `CloneIn` (#23847) (overlookmotel) - 3ea9304 react_compiler: Use faster API to arena allocate strings (#23849) (overlookmotel) - a6d8e45 parser: Avoid span lookup for arrow expression body (#23788) (camc314) - e1886a0 transformer, minifier: Use `static_ident!` macro to create static `Ident`s (#23727) (overlookmotel) - 5527bef transformer/object-rest-spread: Reduce iteration (#23720) (overlookmotel) - 680ffbc transformer: Allocate AST nodes in arena directly (#23711) (overlookmotel) - 1c63c66 parser: Allocate AST nodes in arena directly (#23712) (overlookmotel) - 3855f0c minifier: Allocate AST nodes in arena directly (#23710) (overlookmotel) - d025887 isolated_declarations: Allocate AST nodes in arena directly (#23709) (overlookmotel) - 10b96c6 parser: Remove string search from parsing JSX element name (#23713) (overlookmotel) ### 📚 Documentation - 3d61dea all: Correct capitalization in comments (#23887) (overlookmotel) - aa1ad74 ast: Add `#[deprecated]` to legacy `AstBuilder` methods (#23877) (overlookmotel) - a4676db ast: Correct doc comment for `NONE` (#23765) (overlookmotel) - 419ec80 syntax: Fix typo in doc comment (#23674) (overlookmotel) ### 🛡️ Security - 3cdd18f deps: Update npm packages (#23690) (renovate[bot]) Co-authored-by: Boshen <1430279+Boshen@users.noreply.github.com> Co-authored-by: Cameron <cameron.clark@hey.com>

… has no `(` (#23908) ## What `PatternParser::parse` always runs a full pre-pass over the whole pattern to count capturing groups and collect named-group names before the real parse. The existing comment already flags this: ```rust // NOTE: It means that this perform 2 loops for every cases. // - Cons: 1st pass is completely useless if the pattern does not contain any capturing groups // We may re-consider this if we need more performance rather than simplicity. ``` A pattern with no `(` can have no capturing group, no named group and no duplicate name, so the pre-parse results are exactly `State::new`'s defaults — only `named_capture_groups` differs, and for a group-free pattern it depends solely on the `u`/`v` flag. This detects the absence of `(` with a cheap scan over the already-decoded units (`Reader::contains`) and skips the pre-parse and its allocations (the `AlternativeTracker` vec, the per-named-group vecs, the group-name `FxHashSet`) in that case. A `(` that turns out to be escaped (`\(`) or non-capturing (`(?:…)`, lookarounds) still takes the existing slow path — `contains('(')` is a conservative over-approximation, so it is only ever *skipped* when the pattern is provably group-free. ## Correctness Output is byte-identical. I verified this with a differential harness that parses 60 patterns with both the old and new code and compares the resulting AST / error exactly. The corpus deliberately covers the tricky equivalences: - backreferences in group-free patterns: `\1`, `\1\2\3`, `a\1b` (legacy octal in Annex B) — both with and without the `u` flag; - `\k<a>` in a group-free pattern with `""`, `"u"`, `"v"` flags (the only case where `named_capture_groups` is non-default); - a `(` that must still take the slow path: `\(`, `[(]`, `(?:…)`, lookarounds, named groups, duplicate-name errors (`(?<a>)(?<a>)`), unterminated `(`. All 60 cases produce identical output, and `cargo test -p oxc_regular_expression` passes. ## Measurement Local microbench over 15 realistic paren-free regexes (release): | | ns/parse | | --- | --- | | before | ~258 ns | | after | ~230 ns | ~11% faster for paren-free patterns, which are the large majority of real-world regexes. The change removes a whole code-point pass plus its allocations for those patterns. The allocation snapshot (`cargo allocs`) confirms the reduction on real files — the allocation *count* drops while total bytes are unchanged: | file | allocs before → after | | --- | --- | | antd.js | 3720 → 3661 | | kitchen-sink.tsx | 2108 → 2058 | | pdf.mjs | 356 → 337 | | checker.ts | 1825 → 1818 | | App.tsx | 364 → 360 | Honest note on CodSpeed: the bench corpus is regex-light (regex parsing is well under 2% of parse time on the heaviest bench file and ~0% on the rest), so the CodSpeed comment will likely show no change. The win shows up on regex-heavy inputs (e.g. moment.js / vue.js, where regex parsing is ~7–10% of parse). This PR is offered in response to the maintainer note quoted above, framed as removing documented useless work rather than as a benchmark mover. --- _Disclosure: developed with AI assistance (Claude), reviewed and verified by the author._ --------- Co-authored-by: Yuji Sugiura <y.sugiura.0316@gmail.com>

### 💥 BREAKING CHANGES - 94fbacb ast: [**BREAKING**] Only export `AstBuilder` and `NONE` in `builder` module (#23876) (overlookmotel) - 8de5122 ecmascript: [**BREAKING**] Switch to new `AstBuilder` (#23834) (overlookmotel) - dc0ef38 transformer: [**BREAKING**] Switch to new `AstBuilder` (#23831) (overlookmotel) - 88f4455 str: [**BREAKING**] `Str` and `Ident` methods take `&GetAllocator` (#23781) (overlookmotel) - 36009dd allocator: [**BREAKING**] `GetAllocator::allocator` take `&self` (#23676) (overlookmotel) - bd74f9d allocator: [**BREAKING**] Rename `AllocatorAccessor` trait to `GetAllocator` (#23675) (overlookmotel) ### 🚀 Features - 326fe25 transformer_plugins: Support `typeof` `define` keys (#23605) (Alexander Lichter) - f2091b3 ast: Unify old and new `AstBuilder`s (#23875) (overlookmotel) - cd1fd12 codegen: Expose `Codegen::print_string` API (#23785) (camc314) - 785461b ast: Add custom builder methods to AST types (#23651) (overlookmotel) - 05d1357 ast: Add AST creation methods to AST types (#23650) (overlookmotel) - 2580eda str: Add `Str::from_str_in` and `Ident::from_str_in` methods (#23767) (overlookmotel) - 6883fcf minifier: Fold write-once falsy var to false in boolean context (#23540) (Dunqing) - fcbf993 allocator: Add `Vec::from_value_in` method (#23718) (overlookmotel) - 989ddb7 allocator: Add `Vec::from_box_in` method (#23717) (overlookmotel) - 9d1aa7f allocator: Improve `PartialEq` for `Vec` (#23716) (overlookmotel) ### 🐛 Bug Fixes - da0e5bf minifier: Don't reorder a closed-over TDZ read when inlining a var (#23771) (Dunqing) - 0b3021f allocator: Remove `Vec::from_box_in` (#23873) (overlookmotel) - 0ab64ec ast: Silence deprecation warnings within files defining deprecated `AstBuilder` methods (#23889) (overlookmotel) - 8c07cad all: Enable `disable_old_builder` Cargo feature for `oxc_ast` crate in tests (#23888) (overlookmotel) - 3800f01 ast: Legacy `AstBuilder` methods take `self` not `&self` (#23891) (overlookmotel) - 869ac20 semantic/cfg: Connect for update exit to loop test (#23791) (camc314) - d3e92d5 semantic/cfg: Connect while branches from condition exit (#23790) (camc314) - 025045d ast: `ExportNamedDeclaration` plain builder methods return boxed nodes (#23783) (overlookmotel) - 7537c58 ast: Fix name of `AstBuilder` method for `Expression::V8IntrinsicExpression` (#23766) (overlookmotel) - 3f574f5 traverse: Fix unsoundness in `Traverse` walk functions (#23745) (overlookmotel) - 585760f parser: String in AST reference arena (#23721) (overlookmotel) - 7231d55 allocator: Fix unsound lifetime extension in `Box::new_in` (#23685) (overlookmotel) ### ⚡ Performance - d5c916a semantic: Flatten hoisting_variables to avoid per-scope map allocation (#23927) (Lawrence Lin) - e71609d minifier: Bail member-expr folding before the side-effect walk (#23924) (Lawrence Lin) - e1f89ab minifier: Reduce string allocations folding addition (#23846) (overlookmotel) - 9f6ee3b isolated-declarations: Pool scope maps to avoid per-scope alloc/rehash (#23761) (Boshen) - 0b07c4c semantic: Avoid heap alloc for catch-clause binding ids (#23911) (Lawrence Lin) - c5eef8b regular_expression: Skip capturing-group pre-parse when pattern has no `(` (#23908) (Lawrence Lin) - b4f5b4b isolated_declarations: Remove redundant clone of formal parameter pattern (#23912) (Lawrence Lin) - 53d083f isolated_declarations: Use `TakeIn` not `CloneIn` (#23847) (overlookmotel) - 3ea9304 react_compiler: Use faster API to arena allocate strings (#23849) (overlookmotel) - a6d8e45 parser: Avoid span lookup for arrow expression body (#23788) (camc314) - e1886a0 transformer, minifier: Use `static_ident!` macro to create static `Ident`s (#23727) (overlookmotel) - 5527bef transformer/object-rest-spread: Reduce iteration (#23720) (overlookmotel) - 680ffbc transformer: Allocate AST nodes in arena directly (#23711) (overlookmotel) - 1c63c66 parser: Allocate AST nodes in arena directly (#23712) (overlookmotel) - 3855f0c minifier: Allocate AST nodes in arena directly (#23710) (overlookmotel) - d025887 isolated_declarations: Allocate AST nodes in arena directly (#23709) (overlookmotel) - 10b96c6 parser: Remove string search from parsing JSX element name (#23713) (overlookmotel) ### 📚 Documentation - 3d61dea all: Correct capitalization in comments (#23887) (overlookmotel) - aa1ad74 ast: Add `#[deprecated]` to legacy `AstBuilder` methods (#23877) (overlookmotel) - a4676db ast: Correct doc comment for `NONE` (#23765) (overlookmotel) - 419ec80 syntax: Fix typo in doc comment (#23674) (overlookmotel) ### 🛡️ Security - 3cdd18f deps: Update npm packages (#23690) (renovate[bot]) Co-authored-by: Boshen <1430279+Boshen@users.noreply.github.com> Co-authored-by: Cameron <cameron.clark@hey.com>

linyiru requested a review from leaysgur as a code owner June 28, 2026 20:57

camc314 added the A-regular-expression Area - Regular Expression label Jun 28, 2026

linyiru force-pushed the perf/regex-skip-prescan branch from 8504dc4 to 3a5fd7b Compare June 28, 2026 21:06

linyiru requested a review from overlookmotel as a code owner June 28, 2026 21:06

linyiru force-pushed the perf/regex-skip-prescan branch from 3a5fd7b to 3f5b66f Compare June 28, 2026 21:14

leaysgur added 2 commits June 29, 2026 10:31

fix

d2a23c5

fix

aaf5745

leaysgur approved these changes Jun 29, 2026

View reviewed changes

leaysgur merged commit c5eef8b into oxc-project:main Jun 29, 2026
38 checks passed

oxc-guard Bot mentioned this pull request Jun 29, 2026

release(crates): oxc v0.138.0 #23934

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

perf(regular_expression): skip capturing-group pre-parse when pattern has no `(`#23908

perf(regular_expression): skip capturing-group pre-parse when pattern has no `(`#23908
leaysgur merged 3 commits into
oxc-project:mainfrom
linyiru:perf/regex-skip-prescan

linyiru commented Jun 28, 2026 •

edited

Loading

Uh oh!

codspeed-hq Bot commented Jun 28, 2026 •

edited

Loading

Uh oh!

leaysgur left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Uh oh!

Uh oh!

Conversation

linyiru commented Jun 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What

Correctness

Measurement

Uh oh!

codspeed-hq Bot commented Jun 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will not alter performance

Footnotes

Uh oh!

leaysgur left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

linyiru commented Jun 28, 2026 •

edited

Loading

codspeed-hq Bot commented Jun 28, 2026 •

edited

Loading