string-width - npm Package Compare versions

Comparing version 7.2.0 to 8.0.0

index.js

 import stripAnsi from 'strip-ansi';
 import {eastAsianWidth} from 'get-east-asian-width';
-import emojiRegex from 'emoji-regex';
 
+/**
+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.
+	2. Emoji clusters are double-width only when VS16 is present, the base has Emoji_Presentation (and not VS15), or the cluster has multiple scalars (flags, ZWJ, keycaps, tags, etc.).
+	3. 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).
+*/
+
 const segmenter = new Intl.Segmenter();
 
-const defaultIgnorableCodePointRegex = /^\p{Default_Ignorable_Code_Point}$/u;
+// Whole-cluster zero-width
+const zeroWidthClusterRegex = /^(?:\p{Default_Ignorable_Code_Point}|\p{Control}|\p{Mark}|\p{Surrogate})+$/v;
 
-export default function stringWidth(string, options = {}) {
-	if (typeof string !== 'string' || string.length === 0) {
+// 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;
+
+// RGI emoji sequences
+const rgiEmojiRegex = /^\p{RGI_Emoji}$/v;
+// Default emoji presentation (single-scalar emoji without VS16)
+const emojiPresentationRegex = /^\p{Emoji_Presentation}$/v;
+
+function baseVisible(segment) {
+	return segment.replace(leadingNonPrintingRegex, '');
+}
+
+function isZeroWidthCluster(segment) {
+	return zeroWidthClusterRegex.test(segment);
+}
+
+function isDoubleWidthEmojiCluster(segment) {
+	const visible = baseVisible(segment);
+	const baseScalar = visible.codePointAt(0);
+	const baseChar = String.fromCodePoint(baseScalar);
+	const baseIsEmojiPresentation = emojiPresentationRegex.test(baseChar);
+	const hasVs16 = segment.includes('\uFE0F');
+	const hasVs15 = segment.includes('\uFE0E');
+	const codePointCount = [...segment].length;
+	const multiScalarMeaningful = codePointCount > 1 && !(codePointCount === 2 && hasVs15 && !hasVs16);
+
+	return hasVs16 || (baseIsEmojiPresentation && !hasVs15) || multiScalarMeaningful;
+}
+
+function trailingHalfwidthWidth(segment, eastAsianWidthOptions) {
+	let extra = 0;
+	if (segment.length > 1) {
+		for (const char of segment.slice(1)) {
+			if (char >= '\uFF00' && char <= '\uFFEF') {
+				extra += eastAsianWidth(char.codePointAt(0), eastAsianWidthOptions);
+			}
+		}
+	}
+
+	return extra;
+}
+
+export default function stringWidth(input, options = {}) {
+	if (typeof input !== 'string' || input.length === 0) {
 		return 0;
@@ -19,2 +70,4 @@ }
 
+	let string = input;
+
 	if (!countAnsiEscapeCodes) {
@@ -31,46 +84,10 @@ string = stripAnsi(string);
 
-	for (const {segment: character} of segmenter.segment(string)) {
-		const codePoint = character.codePointAt(0);
-
-		// Ignore control characters
-		if (codePoint <= 0x1F || (codePoint >= 0x7F && codePoint <= 0x9F)) {
+	for (const {segment} of segmenter.segment(string)) {
+		// Zero-width / non-printing clusters
+		if (isZeroWidthCluster(segment)) {
 			continue;
 		}
 
-		// Ignore zero-width characters
-		if (
-			(codePoint >= 0x20_0B && codePoint <= 0x20_0F) // Zero-width space, non-joiner, joiner, left-to-right mark, right-to-left mark
-			|| codePoint === 0xFE_FF // Zero-width no-break space
-		) {
-			continue;
-		}
-
-		// Ignore combining characters
-		if (
-			(codePoint >= 0x3_00 && codePoint <= 0x3_6F) // Combining diacritical marks
-			|| (codePoint >= 0x1A_B0 && codePoint <= 0x1A_FF) // Combining diacritical marks extended
-			|| (codePoint >= 0x1D_C0 && codePoint <= 0x1D_FF) // Combining diacritical marks supplement
-			|| (codePoint >= 0x20_D0 && codePoint <= 0x20_FF) // Combining diacritical marks for symbols
-			|| (codePoint >= 0xFE_20 && codePoint <= 0xFE_2F) // Combining half marks
-		) {
-			continue;
-		}
-
-		// Ignore surrogate pairs
-		if (codePoint >= 0xD8_00 && codePoint <= 0xDF_FF) {
-			continue;
-		}
-
-		// Ignore variation selectors
-		if (codePoint >= 0xFE_00 && codePoint <= 0xFE_0F) {
-			continue;
-		}
-
-		// This covers some of the above cases, but we still keep them for performance reasons.
-		if (defaultIgnorableCodePointRegex.test(character)) {
-			continue;
-		}
-
-		// TODO: Use `/\p{RGI_Emoji}/v` when targeting Node.js 20.
-		if (emojiRegex().test(character)) {
+		// Emoji width logic
+		if (rgiEmojiRegex.test(segment) && isDoubleWidthEmojiCluster(segment)) {
 			width += 2;
@@ -80,3 +97,8 @@ continue;
 
+		// Everything else: EAW of the cluster’s first visible scalar
+		const codePoint = baseVisible(segment).codePointAt(0);
 		width += eastAsianWidth(codePoint, eastAsianWidthOptions);
+
+		// Add width for trailing Halfwidth and Fullwidth Forms (e.g., ﾞ, ﾟ, ｰ)
+		width += trailingHalfwidthWidth(segment, eastAsianWidthOptions);
 	}
@@ -83,0 +105,0 @@

package.json

 {
 	"name": "string-width",
-	"version": "7.2.0",
+	"version": "8.0.0",
 	"description": "Get the visual width of a string - the number of columns required to display it",
@@ -20,3 +20,3 @@ "license": "MIT",
 	"engines": {
-		"node": ">=18"
+		"node": ">=20"
 	},
@@ -40,2 +40,4 @@ "scripts": {
 		"full-width",
+		"wcwidth",
+		"wcswidth",
 		"full",
@@ -57,11 +59,10 @@ "ansi",
 	"dependencies": {
-		"emoji-regex": "^10.3.0",
-		"get-east-asian-width": "^1.0.0",
+		"get-east-asian-width": "^1.3.0",
 		"strip-ansi": "^7.1.0"
 	},
 	"devDependencies": {
-		"ava": "^5.3.1",
-		"tsd": "^0.29.0",
-		"xo": "^0.56.0"
+		"ava": "^6.4.1",
+		"tsd": "^0.33.0",
+		"xo": "^1.2.2"
 	}
 }

No alert changes

Improved metrics

Total package byte prevSize

8902

14.52%

Dependency count

2

-33.33%

Lines of code

112

17.89%

Dependency changes

• - Removed

  emoji-regex@^10.3.0

• - Removed

  emoji-regex@10.6.0
  (transitive)

• Updated

  get-east-asian-width@^1.3.0