oxc-project/oxc - CodSpeed

oxc

Blog Docs Changelog

Performance History

Latest Results

fix(minifier): validate RegExp patterns before marking as pure Invalid RegExp patterns like `RegExp("[")` or invalid flags like `RegExp("a", "xyz")` throw SyntaxError at runtime, so they should not be removed by the minifier. This change uses `oxc_regular_expression` to validate patterns at compile time: - Valid patterns are marked as pure and can be removed when unused - Invalid patterns or non-literal arguments are kept Fixes #18050 Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

fix/minifier-regexp-validation

6 minutes ago

fix(minifier): validate RegExp patterns before marking as pure Invalid RegExp patterns like `RegExp("[")` or invalid flags like `RegExp("a", "xyz")` throw SyntaxError at runtime, so they should not be removed by the minifier. This change uses `oxc_regular_expression` to validate patterns at compile time: - Valid patterns are marked as pure and can be removed when unused - Invalid patterns or non-literal arguments are kept Fixes #18050 Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

fix/minifier-regexp-validation

11 minutes ago

[autofix.ci] apply automated fixes

feat/unambiguous-module-parsing

11 minutes ago

docs(linter): prefer enum docs over struct field docs for no-unused-vars Remove duplicated documentation from struct field comments that was already present on enum variants. This reduces duplication in the generated website docs. Closes #17937 Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

docs/no-unused-vars-dedupe-enum-docs

24 minutes ago

fix(codegen): always escape BOM (U+FEFF) in strings BOM (U+FEFF), U+2028 (Line Separator), and U+2029 (Paragraph Separator) must always be escaped in string literals, regardless of ascii_only setting. Confirmed with esbuild: even with --charset=utf8, BOM is escaped as \uFEFF. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

feat/codegen-ascii-only

1 hour ago

fix(codegen): always escape BOM (U+FEFF) in strings BOM (U+FEFF), U+2028 (Line Separator), and U+2029 (Paragraph Separator) must always be escaped in string literals, regardless of ascii_only setting. Confirmed with esbuild: even with --charset=utf8, BOM is escaped as \uFEFF. Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

feat/codegen-ascii-only

1 hour ago

feat(ast): add escape_raw parameter to template_element builders Add `escape_raw: bool` parameter to `template_element` and `template_element_with_lone_surrogates` builder methods. When building template literals programmatically from cooked values, special characters (backticks, dollar-braces, backslashes, carriage returns) need to be escaped in the raw field. Setting `escape_raw: true` automatically handles this escaping. The parser passes `false` since it receives already-escaped values from source. Programmatic users should pass `true` when creating template elements from runtime string values. Closes #18101 Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

feat/template-literal-helper

2 hours ago

perf(mangler): turn off scope_tree_child_ids for SemanticBuilder

01-17-perf_mangler_turn_off_scope_tree_child_ids_for_semanticbuilder

2 hours ago

Active Branches

fix(minifier): validate RegExp patterns before marking as pure

last run

6 minutes ago

0%

feat(parser): implement unambiguous module parsing for JS/TS files

last run

11 minutes ago

0%

docs(linter): prefer enum docs over struct field docs for no-unused-vars

last run

24 minutes ago

0%

© 2026 CodSpeed Technology

Home Terms Privacy Docs