@contentful/field-editor-rich-text 3.4.14 → 3.4.15

package/CHANGELOG.md

@@ -3,6 +3,12 @@
 All notable changes to this project will be documented in this file.
 See [Conventional Commits](https://conventionalcommits.org) for commit guidelines.
 
+## [3.4.15](https://github.com/contentful/field-editors/compare/@contentful/field-editor-rich-text@3.4.14...@contentful/field-editor-rich-text@3.4.15) (2023-03-20)
+
+### Bug Fixes
+
+- **RTE:** split up sanitize regex to prevent invalid markup ([#1377](https://github.com/contentful/field-editors/issues/1377)) ([6591252](https://github.com/contentful/field-editors/commit/659125216f52ac249b06168d4a7651a1b2dde0b7))
+
 ## [3.4.14](https://github.com/contentful/field-editors/compare/@contentful/field-editor-rich-text@3.4.13...@contentful/field-editor-rich-text@3.4.14) (2023-03-16)
 
 ### Bug Fixes

package/dist/field-editor-rich-text.cjs.development.js

@@ -6249,23 +6249,38 @@ var sanitizeHTML = function sanitizeHTML(html) {
   var doc = transformers.reduce(function (value, cb) {
     return cb(value);
   }, new DOMParser().parseFromString(html, 'text/html'));
-  var previous;
-
-  do {
-    // save previous first before doing modifications
-    previous = doc.body.innerHTML; // Update the body with the cleaned up content
-
-    doc.body.innerHTML = doc.body.innerHTML // remove div container from tables
+  var replacers = [// remove whitespaces between some tags, as this can lead to unwanted behaviour:
+  // - table -> empty table cells
+  // - list -> leading whitespaces
+  function (innerHtml) {
+    return innerHtml.replace(/<(\/)?(table|thead|tbody|tr|td|th|caption|col|colgroup|ol|ul|li)(.*)>\s+<(\/)?(table|thead|tbody|tr|td|th|caption|col|colgroup|ol|ul|li)/g, '<$1$2$3><$4$5');
+  }, // remove empty elements before the ending block element tag
+  function (innerHtml) {
+    return innerHtml.replace(/(?:<[^>^/]*>)\s*(?:<\/[^>]*>)<\/(div|p|table|thead|tbody|tr|td|th|caption|col|colgroup|ol|ul|li)/g, '</$1');
+  }, // remove whitespaces before the ending block element tag
+  function (innerHTML) {
+    return innerHTML.replace(/\s*<\/(div|p|table|thead|tbody|tr|td|th|caption|col|colgroup|ol|ul|li)/g, '</$1');
+  }]; // Call this replacer only if there are tables inside, as it can have a larger performance impact
+
+  if (doc.querySelector('table')) {
+    // remove div container from tables
     // capture groups explained:
     // 1. and 3. every content/linebreaks before and after the div container
     // 2. the table inside the container, including content and linebreaks
     // The div container including attributes and possible linebreaks inside wil be removed
-    .replace(/(.*\s)?<div.*>\s*(<table(?:.|\s)*<\/table>)\s*<\/div>(.*\s)?/g, '$1$2$3') // remove whitespaces between some tags, as this can lead to unwanted behaviour:
-    // - table -> empty table cells
-    // - list -> leading whitespaces
-    .replace(/<(\/)?(table|thead|tbody|tr|td|th|caption|col|colgroup|ol|ul|li)(.*)>\s+<(\/)?(table|thead|tbody|tr|td|th|caption|col|colgroup|ol|ul|li)/g, '<$1$2$3><$4$5') // Removes empty elements before the closing tag of the capture group
-    // <p><span> </span></p> -> <p></p>
-    .replace(/(?:<[^>^/]*>)?\s*(?:<\/[^>]*>)?<\/(div|p|table|thead|tbody|tr|td|th|caption|col|colgroup|ol|ul|li)/g, '</$1');
+    replacers.unshift(function (innerHtml) {
+      return innerHtml.replace(/(.*\s)?<div.*>\s*(<table(?:.|\s)*<\/table>)\s*<\/div>(.*\s)?/g, '$1$2$3');
+    });
+  }
+
+  var previous;
+
+  do {
+    // save previous first before doing modifications
+    previous = doc.body.innerHTML;
+    doc.body.innerHTML = replacers.reduce(function (innerHTML, replacer) {
+      return replacer(innerHTML);
+    }, doc.body.innerHTML);
   } while (doc.body.innerHTML !== previous);
 
   return doc.body.innerHTML;