Fixes #3019 - Better import from LO / OO / soffice

src/node/handler/ImportHandler.js

@@ -181,8 +181,15 @@ async function doImport(req, res, padId)
   if (!req.directDatabaseAccess) {
     text = await fsp_readFile(destFile, "utf8");
 
-    // Title needs to be stripped out else it appends it to the pad..
-    text = text.replace("<title>", "<!-- <title>");
+    /*
+     * The <title> tag needs to be stripped out, otherwise it is appended to the
+     * pad.
+     *
+     * Moreover, when using LibreOffice to convert the file, some classes are
+     * added to the <title> tag. This is a quick & dirty way of matching the
+     * title and comment it out independently on the classes that are set on it.
+     */
+    text = text.replace("<title", "<!-- <title");
     text = text.replace("</title>","</title>-->");
 
     // node on windows has a delay on releasing of the file lock.

src/node/utils/LibreOffice.js

@@ -38,6 +38,14 @@ var libreOfficeLogger = log4js.getLogger('LibreOffice');
  * @param  {Function}   callback    Standard callback function
  */
 exports.convertFile = function(srcFile, destFile, type, callback) {
+  // Used for the moving of the file, not the conversion
+  var fileExtension = type;
+
+  if (type === "html") {
+    // "html:XHTML Writer File:UTF8" does a better job than normal html exports
+    type = "html:XHTML Writer File:UTF8";
+  }
+
   // soffice can't convert from html to doc directly (verified with LO 5 and 6)
   // we need to convert to odt first, then to doc
   // to avoid `Error: no export filter for /tmp/xxxx.doc` error
@@ -47,11 +55,11 @@ exports.convertFile = function(srcFile, destFile, type, callback) {
       "destFile": destFile.replace(/\.doc$/, '.odt'),
       "type": 'odt',
       "callback": function () {
-        queue.push({"srcFile": srcFile.replace(/\.html$/, '.odt'), "destFile": destFile, "type": type, "callback": callback});
+        queue.push({"srcFile": srcFile.replace(/\.html$/, '.odt'), "destFile": destFile, "type": type, "callback": callback, "fileExtension": fileExtension });
       }
     });
   } else {
-    queue.push({"srcFile": srcFile, "destFile": destFile, "type": type, "callback": callback});
+    queue.push({"srcFile": srcFile, "destFile": destFile, "type": type, "callback": callback, "fileExtension": fileExtension});
   }
 };
 
@@ -102,7 +110,7 @@ function doConvertTask(task, callback) {
     // Move the converted file to the correct place
     function(callback) {
       var filename = path.basename(task.srcFile);
-      var sourceFilename = filename.substr(0, filename.lastIndexOf('.')) + '.' + task.type;
+      var sourceFilename = filename.substr(0, filename.lastIndexOf('.')) + '.' + task.fileExtension;
       var sourcePath = path.join(tmpDir, sourceFilename);
       libreOfficeLogger.debug(`Renaming ${sourcePath} to ${task.destFile}`);
       fs.rename(sourcePath, task.destFile, callback);

src/static/js/contentcollector.js

@@ -526,6 +526,16 @@ function makeContentCollector(collectStyles, abrowser, apool, domInterface, clas
         if (isPre) cc.incrementFlag(state, 'preMode');
         var oldListTypeOrNull = null;
         var oldAuthorOrNull = null;
+
+        // LibreOffice Writer puts in weird items during import or copy/paste, we should drop them.
+        if (cls === "Numbering_20_Symbols" || cls === "Bullet_20_Symbols") {
+          styl = null;
+          cls = null;
+
+          // We have to return here but this could break things in the future, for now it shows how to fix the problem
+          return;
+        }
+
         if (collectStyles)
         {
           hooks.callAll('collectContentPre', {