Document display encoding problem - UTF8

[vassilisx]

Hello everyone,

when I upload a document with greek characters I experience a display problem in the document preview, when in document details pane. I cannot read the characters, as the encoding display is problematic (not UTF-8 compatible?)

Is there something I can do to solve this (except from translating the document to english of course…), or is it a bug?

Thank you in advance,

vassilis

[afaust]

Hello,

if the problem is limited only to the document preview, then it is likely a problem with the 3rd party software that does the transformation. What kind of document is it? If you are converting Office documents with OpenOffice, you have to make sure that OpenOffice has the necessary fonts with Greek characters available / installed, otherwise they will be replaced with a standard "unknown" character. Due to licensing constraints, Alfresco can not ship a bundled OpenOffice in its installer that contains all possible fonts, since quite a lot of the common fonts (i.e. in Microsoft environments) requires you to have a registered copy of Office or some other software.

* OpenOffice Font Troubleshooting

Regards
Axel

[vassilisx]

Thank you for your time, Axel.

Unfortunately I think that what you describe is not the source of the problem. I get the same result when I upload an .odt or a .txt document written in LibreOffice or gedit in an opensuse linux OS. I don't do any transformation from MS Office documents.

[mrogers]

You have posted this in the "Alfresco in the cloud" forum are you actually using the cloud or are you using your own local repo?

[vassilisx]

I have the same problem in my local repo, but this question regards "Alfresco in the cloud".

So, yes, I am actually using the cloud.

[afaust]

Hello,

I didn't notice the forum this was posted in. For the cloud service, this should be handled by Alfresco.
In terms of your local repo: A txt is converted via the PDFBox library, not OpenOffice, so yes, it may very well be that fonts is not the main / only issue or not even an issue at all. For the odt, it could be an issue if the instance of LibreOffice used has different fonts than the OpenOffice instance on the Alfresco server - but I'd assume this as less likely than between MS Office and OpenOffice.
My next major suspect would be PDF2SWF. Are you able to test the PDF2SWF transformation by saving a copy of your odt from LibreOffice as PDF and manually invoking pdf2swf? This would help in verifying or eliminating that component based on the output.

Are you able to provide a sample odt file without sensitive information for a quick test?

Regards
Axel

[vassilisx]

Here is a test .odt file

[afaust]

Hello,

with your test file I am able to verify the stated behavior on cloud while my own local Alfresco renders this just fine. Using content rule with transformation on the cloud, I am able to verify that conversion of ODT to PDF causes the corruption, so PDF2SWF is no longer a suspect and OpenOffice is promoted back to main suspect.

My local Alfresco - which renders this okay - has the following configuration (stating what I take to be significant):
- Ubuntu 10.04 LTS 64bit
- Alfresco Enterprise 4.0.1
- OpenOffice 3.2.0m12 en_US via APT (no special repositories, using only default ones)
- explicitly installed libfreetype6 and libfreetype6-dev via APT (used for building pdf2swf from source, but may impact OpenOffice as well)

Regards
Axel