On 3/30/11 8:13 PM, "Philip Black-Knight" <[email protected]> wrote: >I had thought about ID attributes, but IDs could causes other problems. >Basically, we want to be able to take a number of documents all >conforming to the schema and add them into one master document. The >smaller, signed, documents may be generated by different machines or >processes at different times, so there is a possibility that 2 separate >documents assign the same ID to a node. Unless I'm mistaken, this would >causes errors later when attempting to validate the signatures.
All you have to do is require pseudorandom IDs, a common approach being UUIDs (with an underscore in front to address XML syntax constraints on IDs). That's not a problem that would justify XPath in my opinion, though ID attributes are certainly not without other issues. >Me neither. There's XQilla, but AFAIK, it's XPath 2.0 only and there >seems to be enough differences between xpath 1.0 and 2.0 that this XQilla >wouldn't work. Right. It appears that XML Signature 2.0 is going to be XPath 2 only (or at least as good as), so if I ever do a clean room project to support that, XQilla would be an option. I was hoping your question meant you knew of something else. -- Scott
