Office viewer issues

Hi Daniel, is there any exception logged on the server when you view the page with the:

{{office attachment="haaswin.ppt" filterStyles="false"/}}

What happens when you import the haaswin.ppt file using the "Add > Page from Office" menu? If you open the haaswin.ppt file with Open/LibreOffice does "File > Preview in web browser" work?

Regarding the xsl files, what kind of links do you have at the top? Are they absolute or relative URLs? Can you give an example?

Can you attach to this issue a sample ppt and xls file that I can use to reproduce the issue locally?

I can't see any exceptions or anything in the log.

I can open the document fine on the server with open office (3.1.1). If I import the ppt it gives me a new page with the same black view with a spinner. There's also 11 jpgs attached with 0 size that show nothing.

Excel sheet broken link example
http://haaseesoft:8081/xwiki/bin/create/Machine+Knowledge/%23table0?parent=Machine+Knowledge.CNC+IO+map

So does "File > Preview in web browser" menu work when you open your ppt file in OpenOffice? The fact that slide images generated by OpenOffice have 0 size clearly indicates that something wrong happens during the import. In order to find if the problem is on our side or on OpenOffice side I would need a sample ppt file that reproduces the problem. Do you have this problem with any ppt file?

Regarding the excel sheet,

http://haaseesoft:8081/xwiki/bin/create/Machine+Knowledge/%23table0?parent=Machine+Knowledge.CNC+IO+map

is the URL you see in the wiki syntax after you import the xls file, right? What is the URL in the excel file?

I have tried several ppt files, they all do the same thing. The web preview seems to be working.

The links are autogenerated,they don't exist in the excel sheet. The seem to be shortcuts down to the generated sheets, sort of like a TOC.

Please follow this steps:

1. Create a new page with this content:

   {{groovy}}
   import java.io.InputStream;
   import java.util.Map;
   import java.util.Collections;
   
   import org.apache.commons.lang.StringUtils;
   import org.xwiki.officeimporter.openoffice.OpenOfficeManager;
   
   import com.xpn.xwiki.web.Utils;
   
   if (request.convert) {
     String officeFileName = request.getParameter("attachment");
     InputStream officeFileStream = doc.getAttachment(officeFileName).getAttachment().getContentInputStream(xcontext.getContext());
     Map<String, InputStream> inputStreams = Collections.singletonMap(officeFileName, officeFileStream);
     String outputFileName = StringUtils.substringBeforeLast(officeFileName, ".") + ".html";
     Map<String, byte[]> artifacts = Utils.getComponent(OpenOfficeManager.class).getConverter().convert(inputStreams, officeFileName, outputFileName);
     print "= Artifact / Size =\n\n";
     for(Map.Entry<String, byte[]> entry : artifacts) {
       print "* " + entry.getKey() + " : " + entry.getValue().length + "\n";
     }
   } else {
     print "Add convert=1&attachment=your.ppt to the query string to trigger the conversion";
   }
   {{/groovy}}
   

2. Attach your ppt file to this new page
3. Add convert=1&attachment=your.ppt to the query string of the view URL. E.g.:

   http://<server>/xwiki/bin/view/Spage/Page?convert=1&attachment=test.ppt
   

Do all *.jpg artifacts have 0 size? Can you try with a newer version of OpenOffice?

Ok, the script showed 0 for all jpgs. I then upgraded to the latest Open office and it showed a non zero size.
But all inline ppts were still not working. But if I deleted the attached ppt an uploaded it again it started working, I assume that was a caching issue and the main issue were old oo version. Thanks for the help.

Now that it is working I noticed that the inline viewer is skipping the last slide of all the ppts. The test page you provided seems to be creating 30 jpgs (0-29) but the viewer only shows 29.

Thanks Daniel for catching this. I'm surprised no one else noticed it before.. I'm going to fix it for 3.1 final and 3.0.1. What about the excel issue, do you still have it after the Office upgrade?

That is still there after the upgrade. I attached a minimal example which duplicates the issue on my setup.

Thanks Daniel. The attached excel file helped me spot XRENDERING-117. I'll close this issue. You can follow XRENDERING-117 for the excel import fix.