icu::UnicodeStringで正規表現 icu::RegexMatcher

#include <fstream>
#include <iostream>
#include <vector>

#include <unicode/ucnv.h>
#include <unicode/regex.h>


// 要リンク
#pragma comment(lib, "icuuc.lib")

//regex.hに必要
#pragma comment(lib, "icuin.lib")


// 以下のdllを要求される
// icudt69.dll
// icuuc69.dll

// icuin69.dll regexに必要


int main()
{
  std::fstream w(
    "C:\\dev\\icu-test\\converted.txt",
    std::ios::out | std::ios::binary | std::ios::trunc);


  std::u16string u16s = u"aい山うkえ👨‍👧經お经";

  icu::UnicodeString ustr(u16s.c_str());

  UErrorCode status = U_ZERO_ERROR;
  icu::RegexMatcher matcher(u"[う-お].+[あ-お]", 0, status);

  matcher.reset(ustr);

  if (matcher.find()) {
    icu::UnicodeString ret;
    int start = matcher.start(status); // string index of start of match.
    int end = matcher.end(status);

    ustr.extract(
      start,
      end - start,
      ret
    );
    w.write((char*)ret.getBuffer(), ret.length()*2);
  }
  /////////////////////////////////

}