🚨 New namespace "w16du" was added in the upcoming office update

[Amer Amayreh]

We faced corrupted .docx file after modified by docx4j
The original file was produced by a beta version of MS Office
After comparing two identical files, one modified with the beta version of office, and the other one modified by the latest stable version, we noticed a new namespace "w16due".

We applied the same fix done in 8.2.9: 1acc819
By adding the mapping for "w16due" and "http://schemas.microsoft.com/office/word/2023/wordml/word16du" to org.docx4j.jaxb.NamespacePrefixMappings class, the issue was fixed.

Thanks

[Jason Harrop]

Thanks for this; the new namespace http://schemas.microsoft.com/office/word/2023/wordml/word16du is documented in https://msopenspecs.azureedge.net/files/MS-DOCX/%5bMS-DOCX%5d-230616.pdf

[Jason Harrop]

Will be in 8.3.10 and 11.4.10 releases.

[Amer Amayreh]

Thanks @plutext , is there a reason why the namespace mapping is not dynamic? I mean read from the xmls and cached instead of hardcoding them?

[Jason Harrop]

To read the namespaces outside of what JAXB does itself, we'd have to parse the XML independently which would affect performance. These changes to @mc:Ignorable are comparatively rare, so trying to handle this dynamically has not been a priority.