Hi, On Thu, Nov 28, 2013 at 1:45 AM, Alex Jitianu <alex_jitianu@sync.ro> wrote:
Hello,
One aspect of using XQuery Update is that it will discard the DOCTYPE and will also expand entities. So you might not end up with what you've expected. We're investigating ways on our part to compensate for this inherent losses.
Yes. The same issue arises of course using XSLT. In the case of XSLT, I can imagine a post-process that would work by performing a (non-XML) parse of the source document to scan for a DOCTYPE declaration and for entity references, find the declarations for the latter, and generate an XSLT near-identity transformation that would serve as a post-process to restore them. It wouldn't be perfect, as it would effectively normalize entity references, and it would only work for entities that expand to single characters. On the other hand it might be good enough for daily wear. (I guess it would have to map all declared entities, not only those used, since the results could conceivably include content not in the source. It would also have to be optional, since there would be cases when switching a document type or cleaning up entities might be the point of the process.) I would be sympathetic to those who felt this sort of complication demonstrates that the functionality isn't a good idea in the architecture. Certainly, it would be a good way to shoot yourself in the foot. Cheers, Wendell