here is an unicode regex example, which I want to get matched:

#include <iostream> #include #include

int main() { std::setlocale(LC_ALL, "");

boost::wregex condition(L"\\p{u}");

std::wstring test_word(L"Ü");

if (boost::regex_match(test_word, condition)) { std::wcout << L"Matches!" << std::endl; }

boost::wregex condition2(L"[[:upper:]]");

if (boost::regex_match(test_word, condition2)) { std::wcout << L"Matches!" << std::endl; }

boost::u32regex condition3 = boost::make_u32regex(L"\\p{u}");

if (boost::u32regex_match(test_word, condition3)) { std::wcout << L"Matches using lib icu!" << std::endl; }

return 0; }

Compiled with -lboost_regex -licuuc.

Result: Only the last regex condition matches.

So I have a few questions:

1. Is u32regex + make_u32regex the *only* way to get my regex condition matched?

Not necessarily, but it's the only way to get *consistent* Unicode support.

2. Why does "upper class" in the second regex condition not match. E.g. when I use:

echo "Ü" | grep '[[:upper:]]'

on command line - it works properly :)

Thanks in advance + regards,

The "Ü" is treated as upper case if std::locale treats it as upper case - I would sort of expect that to be the case - but apparently not :-( HTH, John.