On Tuesday 07 March 2006 15:56, Scott Cantor wrote: > I think by not trying to wrap them. It just doesn't pay to transcode > everything because you *might* need to. I stick with XMLCh* until I need > 8-bit strings and then I wrap them in a simple auto_ptr clone that does a > release() instead of a delete.
For what I am currently planning, I will either have to use the XMLCh outside of Xerces, or convert. Performance is not an immediate concern. > In the long run, it just works out to be less work than trying to wrap the > entire DOM, and much more efficient. > > Note that at this time, most compilers again support basic_string<XMLCh> so > you can do string-class stuff without transcoding. I believe, if I try hard enough, I can get C++ to convert from one encoding to another using features from <locale>, but I'm not sure. I'm guessing it would be overly optimistic to expect the assignment operator in std::basic_string<> to provide that conversion. > The worst part of this is not having string literals, I freely admit. w_char doesn't work, eh? I have to wonder what a person would do if - may the High Ones forbid it - they actually wanted to use the DOM to model a document in, say, a word processor. STH --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
