On 22 Feb 2016 at 12:06:13, Thomas Mortagne ([email protected](mailto:[email protected])) wrote:
As recommended by W3C we are manipulating XML 1.0 in XAR format but what I did not know is that XML 1.0 does not support all possible characters (I tough at worst it would simply be encoded as a UTF8 entity but seems not) and we just got a report of Woodstox failing because of some character it cannot write in XML 1.0 (""Invalid white space character (0x1) in text to output (in xml 1.1, could output as a character entity)").
So I'm wondering if we should move to XML 1.1. It's critical IMO that we are able to export anything that can be stored in XWikiDocument and in the database and it's the case here.
In theory this is most probably a breakage for people using a XML 1.0 parser but not sure how breaking it is in practice.
WDYT ?
I’m not an expert in XML 1.1 vs 1.0 but it seems acceptable to me. Maybe increase the format version of the XAR spec to 1.3 (I think we’re at version 1.2 right now). BTW I’d also list the various version formats on http://extensions.xwiki.org/xwiki/bin/view/Extension/XAR+Module+Specificatio... and explain the differences, i.e. what changes each version brings. Thanks -Vincent
I'm +1 for this but I won't apply it until I get more point of view (ideally from people having more experience that me on XML 1.0 vs 1.1).
-- Thomas Mortagne