Re: [Boost-users] [regex] Working with wchar_t on olderUNIXplatforms

I am intrigued with what you said about converting data from UTF-8 to UTF-32 on the fly. It is absolutely not a problem to convert my Unicode strings to UTF-8 encoded strings. Where could I read about those on the fly conversions and what limitations do they have (e.g. how locale settings are handled)?

What locale settings? UTF-8 is mostly locale-independent (as an encoding), the only locale specific code is in the traits class to handle collation: and it only sees UTF-32 code points. The on-the-fly conversions are performed by iterator adapters in boost/regex/pending/unicode_iterator.hpp and the docs for the Unicode aware code is here: http://www.boost.org/libs/regex/doc/icu_strings.html John.