boost::regex with case-insensitive match with UTF-8 (e.g. uppercase versus lowercase umlauts)

3
votes

After building boost::regex version 1.52 libraries with International Components for Unicode (ICU) support, a regular expression with a case-insensitive match doesn't appear to handle uppercase and lowercase German umlaut characters as expected.

static const std::string pattern("^.*" "\303\226" ".*$");
static const std::string   test1("SCH" "\303\226" "NE");
static const std::string   test2("sch" "\303\266" "ne");
static const boost::regex exp(pattern, boost::regex::icase);
const char *result = (boost::regex_match(test1, exp)) ? "Match" : "NoMatch";
std::cout << "Testing \"" << test1 << "\" against pattern \"" << pattern 
    << "\" : " << result << std::endl;
result = (boost::regex_match(test2, exp)) ? "Match" : "NoMatch";
std::cout << "Testing \"" << test2 << "\" against pattern \"" << pattern 
    << "\" : " << result << std::endl;

Yields:

Testing "SCHÖNE" against pattern "^.*Ö.*$" : Match
Testing "schöne" against pattern "^.*Ö.*$" : NoMatch
1
c++regexboostutf-8

1 Answers

2
votes

Working with Unicode and ICU string types.

Example on LWS.

#include <iostream>
#include <boost/regex.hpp>
#include <boost/regex/icu.hpp>
int main()
{
   static const std::string pattern("^.*" "\303\226" ".*$");
   static const std::string   test1("SCH" "\303\226" "NE");
   static const std::string   test2("sch" "\303\266" "ne");
   static const boost::u32regex exp=boost::make_u32regex(pattern, boost::regex::icase);
   const char *result = (boost::u32regex_match(test1, exp)) ? "Match" : "NoMatch";
   std::cout << "Testing \"" << test1 << "\" against pattern \"" << pattern 
      << "\" : " << result << std::endl;
   result = (boost::u32regex_match(test2, exp)) ? "Match" : "NoMatch";
   std::cout << "Testing \"" << test2 << "\" against pattern \"" << pattern 
      << "\" : " << result << std::endl;
}