Merge pull request #1 from prezly/fix/care-1352-pasting-from-word-mis…

…sing-spaces [CARE-1352] Do not consider whitespace-only elements empty
prezly · Apr 4, 2023 · 04bb2ef · 04bb2ef
2 parents 29079fa + d4d48ba
commit 04bb2ef
Show file tree

Hide file tree

Showing 4 changed files with 30 additions and 1 deletion.
diff --git a/src/__tests__/input/whitespaces-4.html b/src/__tests__/input/whitespaces-4.html
@@ -0,0 +1,21 @@
+<!DOCTYPE html>
+<html>
+<head>
+    <meta http-equiv="content-type" content="text/html; charset=utf-8"/>
+    <title></title>
+    <meta name="generator" content="LibreOffice 7.4.6.2 (Linux)"/>
+    <style type="text/css">
+        @page { size: 8.5in 11in; margin: 0.79in }
+        p { line-height: 115%; text-align: left; orphans: 2; widows: 2; margin-bottom: 0.1in; direction: ltr; background: transparent }
+        p.western { font-size: 11pt; so-language: en-US }
+        p.cjk { font-size: 11pt; so-language: en-US }
+        p.ctl { font-size: 11pt }
+        a:visited { color: #800080; text-decoration: underline }
+        a:link { color: #0000ff; text-decoration: underline }
+    </style>
+</head>
+<body lang="nl-BE" link="#0000ff" vlink="#800080" dir="ltr"><p align="justify" style="line-height: 100%; margin-left: 0.5in; margin-bottom: 0in">
+    <font color="#003768"> <font face="Calibri, serif"><font face="Calibri, serif"><span lang="en-GB"><b>euros</b></span></font></font><font color="#000000"><font face="Calibri, serif"><font size="3" style="font-size: 12pt"><span lang="en-GB">
+</span></font></font></font><font color="#000000"><font face="Calibri, serif"><span lang="en-GB">in</span></font></font></font></p>
+</body>
+</html>
diff --git a/src/__tests__/output/whitespaces-4.html b/src/__tests__/output/whitespaces-4.html
@@ -0,0 +1 @@
+<body><p align="justify" style="line-height: 100%; margin-left: 0.5in; margin-bottom: 0in"><span color="#003768"> <font face="Calibri, serif"><font face="Calibri, serif"><span lang="en-GB"><b>euros</b></span></font></font><font color="#000000"><font face="Calibri, serif"><font size="3" style="font-size: 12pt"><span lang="en-GB"> </span></font></font></font><font color="#000000"><font face="Calibri, serif"><span lang="en-GB">in</span></font></font></span></p></body>
diff --git a/src/cleanDocx.test.ts b/src/cleanDocx.test.ts
@@ -45,4 +45,11 @@ describe('cleanDocx', () => {
         const result = cleanDocx(html, MOCK_RTF);
         expect(result).toBe(expected);
     });
+
+    it('Preserves newline-only spans', () => {
+        const html = readTestFile('input/whitespaces-4.html');
+        const expected = readTestFile('output/whitespaces-4.html');
+        const result = cleanDocx(html, MOCK_RTF);
+        expect(result).toBe(expected);
+    });
 });
diff --git a/src/lib/cleanEmptyElements.ts b/src/lib/cleanEmptyElements.ts
@@ -3,7 +3,7 @@ import { traverseElements } from './traverseElements';
 const ALLOWED_EMPTY_ELEMENTS = ['BR', 'IMG'];
 
 function isEmpty(element: Element): boolean {
-    return !ALLOWED_EMPTY_ELEMENTS.includes(element.nodeName) && !element.innerHTML.trim();
+    return !ALLOWED_EMPTY_ELEMENTS.includes(element.nodeName) && !element.innerHTML;
 }
 
 function removeIfEmpty(element: Element): void {