Streamlined RTL text detection

1ec5 · Aug 22, 2024 · 91e523c · 91e523c
1 parent 5209ec8
commit 91e523c
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 23 deletions.
diff --git a/src/util/script_detection.test.ts b/src/util/script_detection.test.ts
@@ -1,4 +1,4 @@
-import {charAllowsIdeographicBreaking, charAllowsLetterSpacing, charHasUprightVerticalOrientation, charInComplexShapingScript, charInRTLScript} from './script_detection';
+import {charAllowsIdeographicBreaking, charAllowsLetterSpacing, charHasUprightVerticalOrientation, charInComplexShapingScript, stringContainsRTLText} from './script_detection';
 
 describe('charAllowsIdeographicBreaking', () => {
     test('disallows ideographic breaking of Latin text', () => {
@@ -107,35 +107,35 @@ describe('charInComplexShapingScript', () => {
     });
 });
 
-describe('charInRTLScript', () => {
+describe('stringContainsRTLText', () => {
     test('does not identify direction-neutral text as right-to-left', () => {
-        expect(charInRTLScript('3'.codePointAt(0))).toBe(false);
+        expect(stringContainsRTLText('3')).toBe(false);
     });
 
     test('identifies Arabic text as right-to-left', () => {
         // Arabic
-        expect(charInRTLScript('۳'.codePointAt(0))).toBe(true);
+        expect(stringContainsRTLText('۳')).toBe(true);
         // Arabic Supplement
-        expect(charInRTLScript('ݣ'.codePointAt(0))).toBe(true);
+        expect(stringContainsRTLText('ݣ')).toBe(true);
         // Arabic Extended-A
-        expect(charInRTLScript('ࢳ'.codePointAt(0))).toBe(true);
+        expect(stringContainsRTLText('ࢳ')).toBe(true);
         // Arabic Extended-B
-        expect(charInRTLScript('࢐'.codePointAt(0))).toBe(true);
+        expect(stringContainsRTLText('࢐')).toBe(true);
         // Arabic Presentation Forms-A
-        expect(charInRTLScript('ﰤ'.codePointAt(0))).toBe(true);
+        expect(stringContainsRTLText('ﰤ')).toBe(true);
         // Arabic Presentation Forms-B
-        expect(charInRTLScript('ﺽ'.codePointAt(0))).toBe(true);
+        expect(stringContainsRTLText('ﺽ')).toBe(true);
     });
 
     test('identifies Hebrew text as right-to-left', () => {
         // Hebrew
-        expect(charInRTLScript('ה'.codePointAt(0))).toBe(true);
+        expect(stringContainsRTLText('ה')).toBe(true);
         // Alphabetic Presentation Forms
-        expect(charInRTLScript('ﬡ'.codePointAt(0))).toBe(true);
+        expect(stringContainsRTLText('ﬡ')).toBe(true);
     });
 
     test('identifies Thaana text as right-to-left', () => {
         // Thaana
-        expect(charInRTLScript('ޘ'.codePointAt(0))).toBe(true);
+        expect(stringContainsRTLText('ޘ')).toBe(true);
     });
 });
diff --git a/src/util/script_detection.ts b/src/util/script_detection.ts
@@ -389,10 +389,6 @@ const rtlScriptCodes = [
 
 const rtlScriptRegExp = sanitizedRegExpFromScriptCodes(rtlScriptCodes);
 
-export function charInRTLScript(char: number) {
-    return rtlScriptRegExp.test(String.fromCodePoint(char));
-}
-
 export function charInSupportedScript(char: number, canRenderRTL: boolean) {
     // This is a rough heuristic: whether we "can render" a script
     // actually depends on the properties of the font being used
@@ -401,7 +397,7 @@ export function charInSupportedScript(char: number, canRenderRTL: boolean) {
 
     // Even in Latin script, we "can't render" combinations such as the fi
     // ligature, but we don't consider that semantically significant.
-    if (!canRenderRTL && charInRTLScript(char)) {
+    if (!canRenderRTL && rtlScriptRegExp.test(String.fromCodePoint(char))) {
         return false;
     }
     if ((char >= 0x0900 && char <= 0x0DFF) ||
@@ -419,12 +415,7 @@ export function charInSupportedScript(char: number, canRenderRTL: boolean) {
 }
 
 export function stringContainsRTLText(chars: string): boolean {
-    for (const char of chars) {
-        if (charInRTLScript(char.codePointAt(0))) {
-            return true;
-        }
-    }
-    return false;
+    return rtlScriptRegExp.test(chars);
 }
 
 export function isStringInSupportedScript(chars: string, canRenderRTL: boolean) {