boost.regex:switching between ascii and unicode

Question

Is there a convenient way in boost.regex to switch between ascii and utf?

The only way I see right now is to, for example, switch between boost::u32regex to boost::regex.

Is this the only way to switch between unicode and ascii?
I was hoping to be able to just pass a parameter to boost, specifying my character encoding, thus-by not have to duplicate a lot of code.

Answer 1

Is this the only way to switch between unicode and ascii?

Pretty much. What you think of as boost::regex is really a type alias:

namespace boost{
    template <class charT, class traits = regex_traits<charT>  >
    class basic_regex;

    typedef basic_regex<char>      regex;
    typedef basic_regex<wchar_t>   wregex;
}

Note that the character type is a template parameter - it's not a runtime parameter. Since boost::regex is built on char, it cannot support unicode.

boost::u32regex is the same way:

typedef basic_regex<UChar32,icu_regex_traits> u32regex;

In order to really generalize between them, you would have to write everything as a template too. Instead of taking a boost::regex, you take a boost::basic_regex<charT, traits>. That's one of the downsides of templates - they kind of just permeate everything.