Xml Diffing Best Practices

Question

There's a lot of xml out there, and diffing it against canned data has helped me in the past.

But there's a few things I find I need to do:

• Delete segments of the xml I don't care about (timestamps)
• Sort certain xpaths by some key (a relative child xpath)
• Indent and 'Normalise' the xml =

I've written this a few times at various companies. Is there something out there in the open source world we can use to avoid me writing it again at the next company?

Also, what are your experiences with testing xml?

Answer 1

One example of software that could be used for xml testing:

http://xmlstar.sourceforge.net/

I think the only thing missing is being able to sort an xpath based on a child xpath.

E.g.

<xml>
      <b key="bee"></b>
      <a key="apple"></a>
<xml>

I'd like to sort /xml/* using ./@key

Answer 2

If you've written it yourself, why don't you keep a copy of it for yourself and just add it to your toolkit wherever you go? It would probably be more robust and extensible than any other solution.

Answer 3

If you are willing to use Lisp, I suggest looking at the XML parsing and writing functions in Clojure.