fix: issues 837/838 (#839)

adobe · Aug 30, 2023 · 94c7e93 · 94c7e93
1 parent d94fa01
commit 94c7e93
Show file tree

Hide file tree

Showing 2 changed files with 26 additions and 3 deletions.
diff --git a/packages/helix-shared-indexer/src/index-resource.js b/packages/helix-shared-indexer/src/index-resource.js
@@ -56,23 +56,34 @@ const helpers = {
   innerHTML: (elements) => elements.map((el) => el.children.map((child) => toHtml(child)).join('')),
   match: (elements, re) => {
     // todo: maybe base on function ?
-    const result = [];
+    const results = [];
     const regex = new RegExp(re, 'g');
 
     if (!Array.isArray(elements)) {
       // eslint-disable-next-line no-param-reassign
       elements = [elements];
     }
+    let previousIndex = -1;
     elements.forEach((el) => {
       let m;
       const content = typeof el === 'string' ? el : toText(el);
 
       // eslint-disable-next-line no-cond-assign
       while ((m = regex.exec(content)) !== null) {
-        result.push(m[m.length - 1]);
+        const { index } = m;
+        if (index === previousIndex) {
+          // stop collecting empty matches
+          break;
+        }
+        const result = m.findLast((value) => !!value);
+        if (result) {
+          // only add result if non-empty
+          results.push(result);
+        }
+        previousIndex = index;
       }
     });
-    return result;
+    return results;
   },
   words: (text, start, end) => {
     if (Array.isArray(text)) {

diff --git a/packages/helix-shared-indexer/test/index-resource.test.js b/packages/helix-shared-indexer/test/index-resource.test.js
@@ -104,6 +104,14 @@ indices:
         select: div:foobar([class="embed"])
         value: |
           textContent(el)
+      first-alternate:
+        select: main > div:last-of-type > p:nth-of-type(2)
+        value: |
+          match(innerHTML(el), '(.*?)<br><br>.*|(.*)')
+      second-alternate:
+        select: main > div:last-of-type > p:nth-of-type(3)
+        value: |
+          match(innerHTML(el), '(.*?)<br><br>.*|(.*)')
   `;
 
 const BODY = `
@@ -149,6 +157,8 @@ const BODY = `
   </div>
   <div class="default">
     <p>Topics: A, B, C,</p>
+    <p>before<br><br>after</p>
+    <p>before<br>after</p>
   </div>
 </main>
 <footer></footer>
@@ -166,6 +176,8 @@ describe('Index Resource Tests', () => {
       'call-unknown-function': '',
       'condition-unsupported': '',
       date: 44313,
+      'first-alternate': 'before',
+      'second-alternate': 'before<br>after',
       'last-modified': 1614007680,
       'last-modified-raw': 'Mon, 22 Feb 2021 15:28:00 GMT',
       'match-simple': '',