The simplest XML parser with XPath and Wide-char support

Question

I need a C ++ XML parser / DOM with a lot of weight, which:

• It can take UTF-8 as input and analyze it in UTF-16. Maybe he does it directly (perfect!), Or maybe he provides a hook for the conversion (for example, to create a custom stream object that performs the conversion before parsing).
• Offers XPath support.

I watched RapidXML , Kranf xmlParser and pugiXML . The first two of them can resolve requirement No. 1 with a hook. The third, pugiXML, supports requirement # 2. But none of these three fulfills both requirements.

What is the smallest (free) library that can handle both requirements?

Answer 1

pugixml has a UNICODE branch . I think UNICODE will be officially supported in the next version (0.6)

Answer 2

I really would go for TinyXML + TinyXPath ... Tiny, fully compiled UTF-8 and zlib / MIT. If you need a more C ++-like interface, also TinyXML ++