| 67 | } |
| 68 | |
| 69 | static std::string codePointToUTF8(unsigned int cp) |
| 70 | { |
| 71 | std::string result; |
| 72 | |
| 73 | // based on description from http://en.wikipedia.org/wiki/UTF-8 |
| 74 | |
| 75 | if (cp <= 0x7f) |
| 76 | { |
| 77 | result.resize(1); |
| 78 | result[0] = static_cast<char>(cp); |
| 79 | } |
| 80 | else if (cp <= 0x7FF) |
| 81 | { |
| 82 | result.resize(2); |
| 83 | result[1] = static_cast<char>(0x80 | (0x3f & cp)); |
| 84 | result[0] = static_cast<char>(0xC0 | (0x1f & (cp >> 6))); |
| 85 | } |
| 86 | else if (cp <= 0xFFFF) |
| 87 | { |
| 88 | result.resize(3); |
| 89 | result[2] = static_cast<char>(0x80 | (0x3f & cp)); |
| 90 | result[1] = 0x80 | static_cast<char>((0x3f & (cp >> 6))); |
| 91 | result[0] = 0xE0 | static_cast<char>((0xf & (cp >> 12))); |
| 92 | } |
| 93 | else if (cp <= 0x10FFFF) |
| 94 | { |
| 95 | result.resize(4); |
| 96 | result[3] = static_cast<char>(0x80 | (0x3f & cp)); |
| 97 | result[2] = static_cast<char>(0x80 | (0x3f & (cp >> 6))); |
| 98 | result[1] = static_cast<char>(0x80 | (0x3f & (cp >> 12))); |
| 99 | result[0] = static_cast<char>(0xF0 | (0x7 & (cp >> 18))); |
| 100 | } |
| 101 | |
| 102 | return result; |
| 103 | } |
| 104 | |
| 105 | |
| 106 | // Class Reader |