Hi,

up until now the tests here at work regarding openoffice API-functionality were pretty simple. The odt - files were created automatically after a new build and then saved as a txt-file. After that those txt-files were compared to sample txt-files. It worked , but it didn't test everything such as style changes or pictures. I decided to change that testing routine to track down more changes. Therefore I extracted all xml files from the odt documents and tried to compare them with sample files. Worked like a charm ... for a moment. The problem is that sometimes OO handles page-breaks and sometimes even table-definition differently and I don't know why and how to stop it. As an example part of the content.xml from one of the documents:

From the first build :
...
<text:p text:style-name="P9"/>
<text:p text:style-name="P9"/>
<text:p text:style-name="P9"/>
<text:p text:style-name="P9"/>
−<text:p text:style-name="P9">
<text:soft-page-break/>
</text:p>
−<text:p text:style-name="P8">
Es gelten ausschließlich unsere allgemeinen Verkaufs- und Lieferbedingungen.
</text:p>
...

From the second build (same document template, same data, same code):
...
<text:p text:style-name="P9"/>
<text:p text:style-name="P9"/>
<text:p text:style-name="P9"/>
−<text:p text:style-name="P9">
<text:soft-page-break/>
</text:p>
<text:p text:style-name="P9"/>
−<text:p text:style-name="P8">
Es gelten ausschließlich unsere allgemeinen Verkaufs- und Lieferbedingungen.
</text:p>
...

As you can see, the <text:soft-page-break/> is at a different location. A similar change happened once with table definitions. Those changes have no effect in the GUI but are difficult to handle with automated compare routines. Is there anyway to stop those "anomalies"? Or do you have a better idea how to compare documents to check for changes in API-functionality?

Regards,
Steffen

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@api.openoffice.org
For additional commands, e-mail: dev-h...@api.openoffice.org

Reply via email to