Change format of JsUnicodeProperties for lighter weight

slevithan · Dec 21, 2024 · 0067f62 · 0067f62
1 parent dbc23b4
commit 0067f62
Show file tree

Hide file tree

Showing 2 changed files with 99 additions and 99 deletions.
diff --git a/README.md b/README.md
@@ -1,4 +1,4 @@
-# Oniguruma-To-ES
+# Oniguruma-To-ES (鬼車➜ES)
 
 [![npm version][npm-version-src]][npm-version-href]
 [![npm downloads][npm-downloads-src]][npm-downloads-href]

diff --git a/src/unicode.js b/src/unicode.js
@@ -31,6 +31,11 @@ function getIgnoreCaseMatchChars(char) {
   return [...set];
 }
 
+// The following set includes:
+// - All ES2024 general categories and their aliases (all are supported by Oniguruma). See
+//   <github.com/mathiasbynens/unicode-match-property-value-ecmascript/blob/main/data/mappings.js>
+// - All ES2024 binary properties and their aliases (all are supported by Oniguruma). See
+//   <tc39.es/ecma262/multipage/text-processing.html#table-binary-unicode-properties>
 // Unicode properties must be mapped to property names supported by JS, and must also apply JS's
 // stricter rules for casing, whitespace, and underscores in Unicode property names. In order to
 // remain lightweight, this library assumes properties not in this list are Unicode script names
@@ -39,104 +44,99 @@ function getIgnoreCaseMatchChars(char) {
 // an `In_` prefix). See also:
 // - Properties supported in Oniguruma: <github.com/kkos/oniguruma/blob/master/doc/UNICODE_PROPERTIES>
 // - Properties supported in JS by spec version: <github.com/eslint-community/regexpp/blob/main/src/unicode/properties.ts>
-const JsUnicodeProperties = new Set([
-  // ES2024 general categories and their aliases; all are supported by Oniguruma
-  // See <github.com/mathiasbynens/unicode-match-property-value-ecmascript/blob/main/data/mappings.js>
-  'C', 'Other',
-  'Cc', 'Control', 'cntrl',
-  'Cf', 'Format',
-  'Cn', 'Unassigned',
-  'Co', 'Private_Use',
-  'Cs', 'Surrogate',
-  'L', 'Letter',
-  'LC', 'Cased_Letter',
-  'Ll', 'Lowercase_Letter',
-  'Lm', 'Modifier_Letter',
-  'Lo', 'Other_Letter',
-  'Lt', 'Titlecase_Letter',
-  'Lu', 'Uppercase_Letter',
-  'M', 'Mark', 'Combining_Mark',
-  'Mc', 'Spacing_Mark',
-  'Me', 'Enclosing_Mark',
-  'Mn', 'Nonspacing_Mark',
-  'N', 'Number',
-  'Nd', 'Decimal_Number', 'digit',
-  'Nl', 'Letter_Number',
-  'No', 'Other_Number',
-  'P', 'Punctuation', 'punct',
-  'Pc', 'Connector_Punctuation',
-  'Pd', 'Dash_Punctuation',
-  'Pe', 'Close_Punctuation',
-  'Pf', 'Final_Punctuation',
-  'Pi', 'Initial_Punctuation',
-  'Po', 'Other_Punctuation',
-  'Ps', 'Open_Punctuation',
-  'S', 'Symbol',
-  'Sc', 'Currency_Symbol',
-  'Sk', 'Modifier_Symbol',
-  'Sm', 'Math_Symbol',
-  'So', 'Other_Symbol',
-  'Z', 'Separator',
-  'Zl', 'Line_Separator',
-  'Zp', 'Paragraph_Separator',
-  'Zs', 'Space_Separator',
-
-  // ES2024 binary properties and their aliases; all are supported by Oniguruma
-  // See <tc39.es/ecma262/multipage/text-processing.html#table-binary-unicode-properties>
-  'ASCII',
-  'ASCII_Hex_Digit', 'AHex',
-  'Alphabetic', 'Alpha',
-  'Any',
-  'Assigned',
-  'Bidi_Control', 'Bidi_C',
-  'Bidi_Mirrored', 'Bidi_M',
-  'Case_Ignorable', 'CI',
-  'Cased',
-  'Changes_When_Casefolded', 'CWCF',
-  'Changes_When_Casemapped', 'CWCM',
-  'Changes_When_Lowercased', 'CWL',
-  'Changes_When_NFKC_Casefolded', 'CWKCF',
-  'Changes_When_Titlecased', 'CWT',
-  'Changes_When_Uppercased', 'CWU',
-  'Dash',
-  'Default_Ignorable_Code_Point', 'DI',
-  'Deprecated', 'Dep',
-  'Diacritic', 'Dia',
-  'Emoji',
-  'Emoji_Component', 'EComp',
-  'Emoji_Modifier', 'EMod',
-  'Emoji_Modifier_Base', 'EBase',
-  'Emoji_Presentation', 'EPres',
-  'Extended_Pictographic', 'ExtPict',
-  'Extender', 'Ext',
-  'Grapheme_Base', 'Gr_Base',
-  'Grapheme_Extend', 'Gr_Ext',
-  'Hex_Digit', 'Hex',
-  'IDS_Binary_Operator', 'IDSB',
-  'IDS_Trinary_Operator', 'IDST',
-  'ID_Continue', 'IDC',
-  'ID_Start', 'IDS',
-  'Ideographic', 'Ideo',
-  'Join_Control', 'Join_C',
-  'Logical_Order_Exception', 'LOE',
-  'Lowercase', 'Lower',
-  'Math',
-  'Noncharacter_Code_Point', 'NChar',
-  'Pattern_Syntax', 'Pat_Syn',
-  'Pattern_White_Space', 'Pat_WS',
-  'Quotation_Mark', 'QMark',
-  'Radical',
-  'Regional_Indicator', 'RI',
-  'Sentence_Terminal', 'STerm',
-  'Soft_Dotted', 'SD',
-  'Terminal_Punctuation', 'Term',
-  'Unified_Ideograph', 'UIdeo',
-  'Uppercase', 'Upper',
-  'Variation_Selector', 'VS',
-  'White_Space', 'space',
-  'XID_Continue', 'XIDC',
-  'XID_Start', 'XIDS',
-]);
+const JsUnicodeProperties = new Set(
+`C Other
+Cc Control cntrl
+Cf Format
+Cn Unassigned
+Co Private_Use
+Cs Surrogate
+L Letter
+LC Cased_Letter
+Ll Lowercase_Letter
+Lm Modifier_Letter
+Lo Other_Letter
+Lt Titlecase_Letter
+Lu Uppercase_Letter
+M Mark Combining_Mark
+Mc Spacing_Mark
+Me Enclosing_Mark
+Mn Nonspacing_Mark
+N Number
+Nd Decimal_Number digit
+Nl Letter_Number
+No Other_Number
+P Punctuation punct
+Pc Connector_Punctuation
+Pd Dash_Punctuation
+Pe Close_Punctuation
+Pf Final_Punctuation
+Pi Initial_Punctuation
+Po Other_Punctuation
+Ps Open_Punctuation
+S Symbol
+Sc Currency_Symbol
+Sk Modifier_Symbol
+Sm Math_Symbol
+So Other_Symbol
+Z Separator
+Zl Line_Separator
+Zp Paragraph_Separator
+Zs Space_Separator
+ASCII
+ASCII_Hex_Digit AHex
+Alphabetic Alpha
+Any
+Assigned
+Bidi_Control Bidi_C
+Bidi_Mirrored Bidi_M
+Case_Ignorable CI
+Cased
+Changes_When_Casefolded CWCF
+Changes_When_Casemapped CWCM
+Changes_When_Lowercased CWL
+Changes_When_NFKC_Casefolded CWKCF
+Changes_When_Titlecased CWT
+Changes_When_Uppercased CWU
+Dash
+Default_Ignorable_Code_Point DI
+Deprecated Dep
+Diacritic Dia
+Emoji
+Emoji_Component EComp
+Emoji_Modifier EMod
+Emoji_Modifier_Base EBase
+Emoji_Presentation EPres
+Extended_Pictographic ExtPict
+Extender Ext
+Grapheme_Base Gr_Base
+Grapheme_Extend Gr_Ext
+Hex_Digit Hex
+IDS_Binary_Operator IDSB
+IDS_Trinary_Operator IDST
+ID_Continue IDC
+ID_Start IDS
+Ideographic Ideo
+Join_Control Join_C
+Logical_Order_Exception LOE
+Lowercase Lower
+Math
+Noncharacter_Code_Point NChar
+Pattern_Syntax Pat_Syn
+Pattern_White_Space Pat_WS
+Quotation_Mark QMark
+Radical
+Regional_Indicator RI
+Sentence_Terminal STerm
+Soft_Dotted SD
+Terminal_Punctuation Term
+Unified_Ideograph UIdeo
+Uppercase Upper
+Variation_Selector VS
+White_Space space
+XID_Continue XIDC
+XID_Start XIDS`.split(/\s/)
+);
 
 const JsUnicodePropertiesMap = new Map();
 for (const p of JsUnicodeProperties) {