Merge pull request #760 from gijoe0295/gijoe/45635

Fix `<emoji>` is mistaken as `<em>`
Expensify · Jul 26, 2024 · 4cc5e30 · 4cc5e30
2 parents c1c47fd + a018db6
commit 4cc5e30
Show file tree

Hide file tree

Showing 2 changed files with 22 additions and 2 deletions.
diff --git a/__tests__/ExpensiMark-Markdown-test.js b/__tests__/ExpensiMark-Markdown-test.js
@@ -891,3 +891,23 @@ describe('Video tag conversion to markdown', () => {
         expect(cacheVideoAttributes).toHaveBeenCalledWith("https://example.com/video.mp4", ' data-expensify-width="100" data-expensify-height="500" data-expensify-thumbnail-url="https://image.com/img.jpg"')
     })
 })
+
+describe('Tag names starting with common charaters', () => {
+    test('Italic and emoji', () => {
+        const testString = '<emoji>😄</emoji> <em>italic</em>';
+        const resultString = '😄 _italic_';
+        expect(parser.htmlToMarkdown(testString)).toBe(resultString);
+    });
+
+    test('Blockquote and bold', () => {
+        const testString = '<blockquote> quote <b>bold</b></blockquote>';
+        const resultString = '> quote *bold*';
+        expect(parser.htmlToMarkdown(testString)).toBe(resultString);
+    });
+
+    test('Line break and bold', () => {
+        const testString = '<br/><b>bold</b>';
+        const resultString = '\n*bold*';
+        expect(parser.htmlToMarkdown(testString)).toBe(resultString);
+    });
+});
diff --git a/lib/ExpensiMark.ts b/lib/ExpensiMark.ts
@@ -556,12 +556,12 @@ export default class ExpensiMark {
             // Use [\s\S]* instead of .* to match newline
             {
                 name: 'italic',
-                regex: /<(em|i)(?:"[^"]*"|'[^']*'|[^'">])*>([\s\S]*?)<\/\1>(?![^<]*(<\/pre>|<\/code>))/gi,
+                regex: /<(em|i)\b(?:"[^"]*"|'[^']*'|[^'">])*>([\s\S]*?)<\/\1>(?![^<]*(<\/pre>|<\/code>))/gi,
                 replacement: '_$2_',
             },
             {
                 name: 'bold',
-                regex: /<(b|strong)(?:"[^"]*"|'[^']*'|[^'">])*>([\s\S]*?)<\/\1>(?![^<]*(<\/pre>|<\/code>))/gi,
+                regex: /<(b|strong)\b(?:"[^"]*"|'[^']*'|[^'">])*>([\s\S]*?)<\/\1>(?![^<]*(<\/pre>|<\/code>))/gi,
                 replacement: (extras, match, tagName, innerContent) => {
                     // To check if style attribute contains bold font-weight
                     const isBoldFromStyle = (style: string | null) => {