From 2d01f4cd8ba8d0278e7692f3f6641d57a1ece29f Mon Sep 17 00:00:00 2001
From: webzwo0i <webzwo0i@c3d2.de>
Date: Thu, 3 Sep 2020 08:30:30 +0200
Subject: [PATCH] LibreOffice: use filter XHTML Writer File during import
 ImportHandler: remove self-closing title element

---
 src/node/handler/ImportHandler.js    | 10 ++++++++--
 src/node/utils/LibreOffice.js        |  3 ++-
 tests/backend/specs/api/docImport.js |  2 +-
 3 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/src/node/handler/ImportHandler.js b/src/node/handler/ImportHandler.js
index 6ea76cac3..9109853e6 100644
--- a/src/node/handler/ImportHandler.js
+++ b/src/node/handler/ImportHandler.js
@@ -211,9 +211,15 @@ async function doImport(req, res, padId)
      * 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.
+     *
+     * @todo what happens if <title is used somewhere in the pad? need proper XML parser
      */
-    text = text.replace("<title", "<!-- <title");
-    text = text.replace("</title>","</title>-->");
+    if (text.match(/<\/title>/)){
+      text = text.replace("<title", "<!-- <title");
+      text = text.replace("</title>","</title>-->");
+    } else {
+      text = text.replace(/<title[^>]*\/>/,'');
+    }
 
     // node on windows has a delay on releasing of the file lock.
     // We add a 100ms delay to work around this
diff --git a/src/node/utils/LibreOffice.js b/src/node/utils/LibreOffice.js
index dfbee8fa5..3fc80faf0 100644
--- a/src/node/utils/LibreOffice.js
+++ b/src/node/utils/LibreOffice.js
@@ -85,7 +85,8 @@ function doConvertTask(task, callback) {
         '--nologo',
         '--nolockcheck',
         '--writer',
-        '--convert-to', task.type,
+        '--convert-to',
+        `${task.type}:XHTML Writer File:UTF8`,
         task.srcFile,
         '--outdir', tmpDir
       ]);
diff --git a/tests/backend/specs/api/docImport.js b/tests/backend/specs/api/docImport.js
index e2948408e..5c8f73be1 100644
--- a/tests/backend/specs/api/docImport.js
+++ b/tests/backend/specs/api/docImport.js
@@ -108,7 +108,7 @@ describe('getText', function(){
     setTimeout(function(){
     api.get(endPoint('getText')+"&padID="+testPadId)
     .expect(function(res){
-      if(res.body.data.text.indexOf("RichardC. Hay" !== -1)){
+      if(res.body.data.text.indexOf("RichardC. Hay") !== -1) {
         console.warn(res.body.data.text);
         throw new Error("Error with doc keeping spaces", res.body.data.text);
       }