@cjser/string-width 8.2.1-cjser.2 → 8.3.0-cjser.2

package/dist-cjser/index.cjs

@@ -26,7 +26,7 @@ var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__ge
 ));
 var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
 
-// packages/@cjser/string-width/index.js
+// packages/@cjser/string-width.tmp-26-1790266896938/index.js
 var index_exports = {};
 __export(index_exports, {
   default: () => stringWidth
@@ -35,8 +35,8 @@ module.exports = __toCommonJS(index_exports);
 var import_strip_ansi = __toESM(require("@cjser/strip-ansi"), 1);
 var import_get_east_asian_width = require("@cjser/get-east-asian-width");
 var segmenter = new Intl.Segmenter();
-var zeroWidthClusterRegex = new RegExp("^(?:\\p{Default_Ignorable_Code_Point}|\\p{Control}|\\p{Format}|\\p{Mark}|\\p{Surrogate})+$", "v");
-var leadingNonPrintingRegex = new RegExp("^[\\p{Default_Ignorable_Code_Point}\\p{Control}\\p{Format}\\p{Mark}\\p{Surrogate}]+", "v");
+var visibleCharacterRegex = new RegExp("[^\\p{Default_Ignorable_Code_Point}\\p{Control}\\p{Format}\\p{Nonspacing_Mark}\\p{Enclosing_Mark}\\p{Surrogate}]", "v");
+var spacingMarkRegex = new RegExp("\\p{Spacing_Mark}", "v");
 var rgiEmojiRegex = new RegExp("^\\p{RGI_Emoji}$", "v");
 var unqualifiedKeycapRegex = /^[\d#*]\u20E3$/;
 var extendedPictographicRegex = new RegExp("\\p{Extended_Pictographic}", "gu");
@@ -54,10 +54,8 @@ function isDoubleWidthNonRgiEmojiSequence(segment) {
   return false;
 }
 function baseVisible(segment) {
-  return segment.replace(leadingNonPrintingRegex, "");
-}
-function isZeroWidthCluster(segment) {
-  return zeroWidthClusterRegex.test(segment);
+  const index = segment.search(visibleCharacterRegex);
+  return index === -1 ? void 0 : segment.slice(index);
 }
 function isHangulLeadingJamo(codePoint) {
   return codePoint >= 4352 && codePoint <= 4447 || codePoint >= 43360 && codePoint <= 43388;
@@ -74,7 +72,7 @@ function isHangulJamo(codePoint) {
 function hangulClusterWidth(visibleSegment, eastAsianWidthOptions) {
   const codePoints = [];
   for (const character of visibleSegment) {
-    if (zeroWidthClusterRegex.test(character)) {
+    if (!visibleCharacterRegex.test(character)) {
       continue;
     }
     codePoints.push(character.codePointAt(0));
@@ -103,7 +101,7 @@ function hangulClusterWidth(visibleSegment, eastAsianWidthOptions) {
   }
   return width;
 }
-function trailingHalfwidthWidth(visibleSegment, eastAsianWidthOptions) {
+function trailingWidth(visibleSegment, eastAsianWidthOptions) {
   let extra = 0;
   let first = true;
   for (const character of visibleSegment) {
@@ -111,7 +109,7 @@ function trailingHalfwidthWidth(visibleSegment, eastAsianWidthOptions) {
       first = false;
       continue;
     }
-    if (character >= "\uFF00" && character <= "\uFFEF") {
+    if (spacingMarkRegex.test(character) || character >= "\uFF00" && character <= "\uFFEF") {
       extra += (0, import_get_east_asian_width.eastAsianWidth)(character.codePointAt(0), eastAsianWidthOptions);
     }
   }
@@ -138,14 +136,14 @@ function stringWidth(input, options = {}) {
   let width = 0;
   const eastAsianWidthOptions = { ambiguousAsWide: !ambiguousIsNarrow };
   for (const { segment } of segmenter.segment(string)) {
-    if (isZeroWidthCluster(segment)) {
+    const visibleSegment = baseVisible(segment);
+    if (visibleSegment === void 0) {
       continue;
     }
     if (rgiEmojiRegex.test(segment) || isDoubleWidthNonRgiEmojiSequence(segment)) {
       width += 2;
       continue;
     }
-    const visibleSegment = baseVisible(segment);
     const hangulWidth = hangulClusterWidth(visibleSegment, eastAsianWidthOptions);
     if (hangulWidth !== void 0) {
       width += hangulWidth;
@@ -153,7 +151,7 @@ function stringWidth(input, options = {}) {
     }
     const codePoint = visibleSegment.codePointAt(0);
     width += (0, import_get_east_asian_width.eastAsianWidth)(codePoint, eastAsianWidthOptions);
-    width += trailingHalfwidthWidth(visibleSegment, eastAsianWidthOptions);
+    width += trailingWidth(visibleSegment, eastAsianWidthOptions);
   }
   return width;
 }

package/index.js

@@ -5,21 +5,20 @@ import {eastAsianWidth} from '@cjser/get-east-asian-width';
 Logic:
 - Segment graphemes to match how terminals render clusters.
 - Width rules:
-	1. Skip non-printing clusters (Default_Ignorable, Control, pure Mark, lone Surrogates). Tabs are ignored by design.
+	1. Skip non-printing clusters (Default_Ignorable, Control, pure nonspacing/enclosing Mark, lone Surrogates). Tabs are ignored by design.
 	2. RGI emoji clusters (\p{RGI_Emoji}) are double-width.
 	3. Minimally-qualified/unqualified emoji clusters (ZWJ sequences with 2+ Extended_Pictographic, or keycap sequences) are double-width.
 	4. Hangul jamo collapse each standard modern Hangul L+V or L+V+T syllable piece to width 2.
 	   Unmatched repeated leading/vowel/trailing jamo stay additive because that matches how the terminals we target render them.
-	5. Otherwise use East Asian Width of the cluster's first visible code point, and add widths for trailing Halfwidth/Fullwidth Forms within the same cluster (e.g., dakuten/handakuten/prolonged sound mark).
+	5. Otherwise use East Asian Width of the cluster's first visible code point, and add widths for trailing spacing marks and Halfwidth/Fullwidth Forms within the same cluster (e.g., dakuten/handakuten/prolonged sound mark).
 */
 
 const segmenter = new Intl.Segmenter();
 
-// Whole-cluster zero-width
-const zeroWidthClusterRegex = /^(?:\p{Default_Ignorable_Code_Point}|\p{Control}|\p{Format}|\p{Mark}|\p{Surrogate})+$/v;
-
-// Pick the base scalar if the cluster starts with Prepend/Format/Marks
-const leadingNonPrintingRegex = /^[\p{Default_Ignorable_Code_Point}\p{Control}\p{Format}\p{Mark}\p{Surrogate}]+/v;
+// Matches a visible (printing) character. Used to find the base scalar of a cluster (skipping leading Prepend/Format/Marks) and to detect whole-cluster zero-width.
+// Intentionally a single negated class without a quantifier: a repeated pattern like `^(?:…|…)+$` backtracks exponentially on overlapping properties, and even `^[…]+$` overflows V8's backtrack stack (RangeError) on clusters with millions of marks.
+const visibleCharacterRegex = /[^\p{Default_Ignorable_Code_Point}\p{Control}\p{Format}\p{Nonspacing_Mark}\p{Enclosing_Mark}\p{Surrogate}]/v;
+const spacingMarkRegex = /\p{Spacing_Mark}/v;
 
 // RGI emoji sequences
 const rgiEmojiRegex = /^\p{RGI_Emoji}$/v;
@@ -47,12 +46,10 @@ function isDoubleWidthNonRgiEmojiSequence(segment) {
 	return false;
 }
 
+// Returns the cluster starting at its first visible character, or `undefined` if the whole cluster is zero-width.
 function baseVisible(segment) {
-	return segment.replace(leadingNonPrintingRegex, '');
-}
-
-function isZeroWidthCluster(segment) {
-	return zeroWidthClusterRegex.test(segment);
+	const index = segment.search(visibleCharacterRegex);
+	return index === -1 ? undefined : segment.slice(index);
 }
 
 function isHangulLeadingJamo(codePoint) {
@@ -80,7 +77,7 @@ function hangulClusterWidth(visibleSegment, eastAsianWidthOptions) {
 	const codePoints = [];
 
 	for (const character of visibleSegment) {
-		if (zeroWidthClusterRegex.test(character)) {
+		if (!visibleCharacterRegex.test(character)) {
 			continue;
 		}
 
@@ -125,7 +122,7 @@ function hangulClusterWidth(visibleSegment, eastAsianWidthOptions) {
 	return width;
 }
 
-function trailingHalfwidthWidth(visibleSegment, eastAsianWidthOptions) {
+function trailingWidth(visibleSegment, eastAsianWidthOptions) {
 	let extra = 0;
 	let first = true;
 
@@ -135,7 +132,10 @@ function trailingHalfwidthWidth(visibleSegment, eastAsianWidthOptions) {
 			continue;
 		}
 
-		if (character >= '\uFF00' && character <= '\uFFEF') {
+		if (
+			spacingMarkRegex.test(character)
+			|| (character >= '\uFF00' && character <= '\uFFEF')
+		) {
 			extra += eastAsianWidth(character.codePointAt(0), eastAsianWidthOptions);
 		}
 	}
@@ -174,7 +174,8 @@ export default function stringWidth(input, options = {}) {
 
 	for (const {segment} of segmenter.segment(string)) {
 		// Zero-width / non-printing clusters
-		if (isZeroWidthCluster(segment)) {
+		const visibleSegment = baseVisible(segment);
+		if (visibleSegment === undefined) {
 			continue;
 		}
 
@@ -184,7 +185,6 @@ export default function stringWidth(input, options = {}) {
 			continue;
 		}
 
-		const visibleSegment = baseVisible(segment);
 		const hangulWidth = hangulClusterWidth(visibleSegment, eastAsianWidthOptions);
 		if (hangulWidth !== undefined) {
 			width += hangulWidth;
@@ -195,8 +195,8 @@ export default function stringWidth(input, options = {}) {
 		const codePoint = visibleSegment.codePointAt(0);
 		width += eastAsianWidth(codePoint, eastAsianWidthOptions);
 
-		// Add width for trailing Halfwidth and Fullwidth Forms (e.g., ﾞ, ﾟ, ｰ)
-		width += trailingHalfwidthWidth(visibleSegment, eastAsianWidthOptions);
+		// Add width for trailing spacing marks and Halfwidth/Fullwidth Forms (e.g., ﾞ, ﾟ, ｰ)
+		width += trailingWidth(visibleSegment, eastAsianWidthOptions);
 	}
 
 	return width;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@cjser/string-width",
-  "version": "8.2.1-cjser.2",
+  "version": "8.3.0-cjser.2",
   "description": "Get the visual width of a string - the number of columns required to display it",
   "license": "MIT",
   "repository": {
@@ -59,7 +59,7 @@
     "east-asian-width"
   ],
   "dependencies": {
-    "@cjser/get-east-asian-width": "1.5.0-cjser.2",
+    "@cjser/get-east-asian-width": "1.7.0-cjser.2",
     "@cjser/strip-ansi": "7.2.0-cjser.2"
   },
   "devDependencies": {
@@ -70,11 +70,11 @@
   "types": "./index.d.ts",
   "main": "./dist-cjser/index.cjs",
   "cjser": {
-    "sourceVersion": "8.2.1",
+    "sourceVersion": "8.3.0",
     "cjserVersion": 2,
     "original": {
       "name": "string-width",
-      "version": "8.2.1",
+      "version": "8.3.0",
       "exports": {
         "types": "./index.d.ts",
         "default": "./index.js"