Merge pull request #11732 from keymanapp/fix/web/array-from-polyfill

fix(web): add limited Array.from polyfill for lm-worker use

common/models/wordbreakers/src/default/index.ts

@@ -274,7 +274,7 @@ export class BreakerContext {
  * @param chunk a chunk of text. Starts and ends at word boundaries.
  */
 function isNonSpace(chunk: string, options?: DefaultWordBreakerOptions): boolean {
-  return !Array.from(chunk).map((char) => property(char, options)).every(wb => (
+  return !chunk.split('').map((char) => property(char, options)).every(wb => (
     wb === WordBreakProperty.CR ||
     wb === WordBreakProperty.LF ||
     wb === WordBreakProperty.Newline ||

common/web/lm-worker/build-polyfiller.js

@@ -144,6 +144,9 @@ let sourceFileSet = [
   // Needed for Android / Chromium browser pre-41.
   loadPolyfill('../../../node_modules/string.prototype.codepointat/codepointat.js', 'src/polyfills/string.codepointat.js'),
   // Needed for Android / Chromium browser pre-45.
+  // Not used in this codebase, but used by some compiled model defaults.
+  loadPolyfill('src/polyfills/array.from.js', 'src/polyfills/array.from.js'),
+  // Needed for Android / Chromium browser pre-45.
   loadPolyfill('src/polyfills/array.fill.js', 'src/polyfills/array.fill.js'),
   // Needed for Android / Chromium browser pre-45.
   loadPolyfill('src/polyfills/array.findIndex.js', 'src/polyfills/array.findIndex.js'),

common/web/lm-worker/src/polyfills/array.from.js

@@ -0,0 +1,44 @@
+(function() {
+  if(!Array.from) {
+    function isHighSurrogate(codeUnit) {
+      codeUnit = codeUnit.charCodeAt(0);
+      return codeUnit >= 0xD800 && codeUnit <= 0xDBFF;
+    }
+
+    function isLowSurrogate(codeUnit) {
+      codeUnit = codeUnit.charCodeAt(0);
+      return codeUnit >= 0xDC00 && codeUnit <= 0xDFFF;
+    }
+
+    Array.from = function (obj) {
+      if(Array.isArray(obj)) {
+        // Simple array clone
+        return obj.slice();
+      } else if(typeof obj == 'string') {
+        // Array.from is surrogate-aware and will not split surrogate pairs.
+        // We can start with a full split and then remerge the pairs.
+        var simpleSplit = obj.split('');
+
+        /** @type {string[]} */
+        var finalSplit = [];
+        /** @type {number} */
+        var i;
+
+        while(simpleSplit.length > 0) {
+          // Do we have a surrogate pair?
+          var a = simpleSplit.shift();
+          if(isHighSurrogate(a) && (isLowSurrogate(simpleSplit[0] || ''))) {
+            // yes, so merge them before pushing.
+            a = a + simpleSplit.shift();
+            console.log(a);
+          } // else:  'no', so just push the current char to the array and continue
+
+          finalSplit.push(a);
+        }
+        return finalSplit;
+      } else {
+        throw "Unexpected + nonpolyfilled use of Array.from encountered; aborting";
+      }
+    }
+  }
+}());