@formatjs/intl-segmenter - npm Package Compare versions

Comparing version 12.0.6 to 12.0.7

package.json

 {
   "name": "@formatjs/intl-segmenter",
   "description": "Polyfill for Intl.Segmenter",
-  "version": "12.0.6",
+  "version": "12.0.7",
   "license": "MIT",
@@ -15,4 +15,4 @@ "author": "Matija Gaspar <[e-mail]>",
     "tslib": "^2.8.0",
-    "@formatjs/ecma402-abstract": "3.0.6",
-    "@formatjs/intl-localematcher": "0.7.3"
+    "@formatjs/ecma402-abstract": "3.0.7",
+    "@formatjs/intl-localematcher": "0.7.4"
   },
@@ -19,0 +19,0 @@ "bugs": "https://github.com/formatjs/formatjs/issues",

src/segmenter.js

@@ -6,2 +6,8 @@ import { __assign, __spreadArray } from "tslib";
 import { isSurrogate, replaceVariables } from './segmentation-utils.js';
+// Cached regex patterns for word character detection
+// Note: Unicode property escape regex is created at runtime in try-catch
+// to avoid compile-time errors when targeting ES5
+var WORD_CHARACTERS_BASIC_REGEX = /\w/;
+// Lazy-initialized Unicode word character regex (null if not supported)
+var WORD_CHARACTERS_UNICODE_REGEX = undefined;
 /**
@@ -147,2 +153,72 @@ * Adds $ to before rules and ^ to after rules for strictness
 export { Segmenter };
+/**
+ * Determines if a segment is word-like according to Unicode Word Break rules.
+ *
+ * A segment is considered word-like if it contains alphabetic characters,
+ * numbers, or ideographs. Segments containing only whitespace, punctuation,
+ * or symbols are not word-like.
+ *
+ * Per Unicode Word Break (UAX #29) and native Intl.Segmenter implementations,
+ * this matches segments that contain characters from word character classes:
+ * ALetter, Hebrew_Letter, Numeric, Katakana, Hiragana, and Ideographic.
+ *
+ * @param segment - The text segment to check
+ * @param matchingRule - The word break rule that created this segment
+ * @returns true if the segment is word-like
+ */
+function isSegmentWordLike(segment, matchingRule) {
+    // Primary check: Does the segment contain word characters?
+    // Word-like segments contain letters (including ideographs), numbers,
+    // or connecting characters like apostrophes within words
+    //
+    // Regex matches:
+    // - Letters: \p{L} (all Unicode letters)
+    // - Numbers: \p{N} (all Unicode numbers)
+    // - Marks: \p{M} (combining marks, typically part of letters)
+    //
+    // Note: Using Unicode property escapes which work in modern JS engines
+    // and are necessary for proper internationalization
+    // Lazy-initialize Unicode regex on first use
+    if (WORD_CHARACTERS_UNICODE_REGEX === undefined) {
+        try {
+            // Create Unicode property escape regex at runtime to avoid compile-time TS1501 error
+            WORD_CHARACTERS_UNICODE_REGEX = new RegExp('[\\p{L}\\p{N}\\p{M}]', 'u');
+        }
+        catch (_a) {
+            // Environment doesn't support Unicode property escapes
+            WORD_CHARACTERS_UNICODE_REGEX = null;
+        }
+    }
+    var hasWordCharacters;
+    if (WORD_CHARACTERS_UNICODE_REGEX) {
+        // Check if segment contains word characters using Unicode property escapes
+        // This matches the behavior of native Intl.Segmenter in Chrome/Firefox
+        hasWordCharacters = WORD_CHARACTERS_UNICODE_REGEX.test(segment);
+    }
+    else {
+        // Fallback for environments without Unicode property escapes
+        // Match basic word characters: letters, numbers, underscores
+        hasWordCharacters = WORD_CHARACTERS_BASIC_REGEX.test(segment);
+    }
+    // If segment contains word characters, it's word-like
+    if (hasWordCharacters) {
+        return true;
+    }
+    // If no word characters, check if it's definitely not word-like via rules
+    // Non-word-like rules per Unicode Word Break specification (UAX #29):
+    // https://unicode.org/reports/tr29/#Word_Boundaries
+    //
+    // WB3a (3.1): Break before newlines (sot ÷ (Newline | CR | LF))
+    // WB3b (3.2): Break after newlines ((Newline | CR | LF) ÷ eot)
+    // WB3d (3.4): Keep horizontal whitespace together (WSegSpace × WSegSpace)
+    //
+    // These rules specifically identify non-word segments like line breaks and whitespace
+    var definitelyNotWordLikeRules = ['3.1', '3.2', '3.4'];
+    if (definitelyNotWordLikeRules.includes(matchingRule)) {
+        return false;
+    }
+    // For segments without word characters and not matching specific non-word rules,
+    // return false (e.g., punctuation, symbols, whitespace via rule 999)
+    return false;
+}
 var createSegmentDataObject = function (segmenter, segment, index, input, matchingRule) {
@@ -155,3 +231,3 @@ var returnValue = {
     if (getSlot(segmenter, 'granularity') === 'word') {
-        returnValue.isWordLike = matchingRule !== '3.1' && matchingRule !== '3.2';
+        returnValue.isWordLike = isSegmentWordLike(segment, matchingRule);
     }
@@ -158,0 +234,0 @@ return returnValue;

New alerts

Long strings

Supply chain risk

Contains long string literals, which may be a sign of obfuscated or packed code.

Found 1 instance in 1 package

URL strings

Supply chain risk

Package contains fragments of external URLs or IP addresses, which the package may be accessing at runtime.

Found 1 instance in 1 package

Fixed alerts

Long strings

Supply chain risk

Contains long string literals, which may be a sign of obfuscated or packed code.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

957825

0.37%

Worsened metrics

Lines of code

13325

-0.3%

Number of low supply chain risk alerts

112

2.75%

Dependency changes

• + Added

  @formatjs/ecma402-abstract@3.0.7
  (transitive)

• + Added

  @formatjs/fast-memoize@3.0.2
  (transitive)

• + Added

  @formatjs/intl-localematcher@0.7.4
  (transitive)

• - Removed

  @formatjs/ecma402-abstract@3.0.6
  (transitive)

• - Removed

  @formatjs/fast-memoize@3.0.1
  (transitive)

• - Removed

  @formatjs/intl-localematcher@0.7.3
  (transitive)

• Updated

  @formatjs/ecma402-abstract@3.0.7

• Updated

  @formatjs/intl-localematcher@0.7.4