Originally downloaded from Sourceforge, the 1.13 release with tests.
There is an HTML Manual available as well a the original license file.
RapidXml is an attempt to create the fastest XML parser possible, while retaining useability, portability and reasonable W3C compatibility. It is an in-situ parser with parsing speed approaching speed of strlen function executed on the same data.
Maybe we'll try to get some time with the bugs, look through the patches, and review the feature requests.
Development of new features is going into branches. First thing that needs doing is getting all of
the tests running, then I can consider integrating others' work. I'm going to avoid pointless
changes like wholesale formatting or type changes because this code basically works. I have one
immediate need for it, however, which is my current focus: I need to fix wchar_t handling so that this works with native Windows "Unicode" (aka UTF-16) APIs. For the most part this means properly handling numeric character refs but also it needs to not just blindly copy between narrow- and wide-character strings when the user isn't paying attention. Getting UTF-8 data in wchars isn't terribly helpful, for example.
Once different features reach sufficient maturity I'll merge them into master.