Count the number of utf8 characters return 0 for invalid/incomplete input byte sequences
| 175 | // Count the number of utf8 characters |
| 176 | // return 0 for invalid/incomplete input byte sequences |
| 177 | FORCE_INLINE |
| 178 | gdv_int32 utf8_length(gdv_int64 context, const char* data, gdv_int32 data_len) { |
| 179 | int char_len = 0; |
| 180 | int count = 0; |
| 181 | for (int i = 0; i < data_len; i += char_len) { |
| 182 | char_len = utf8_char_length(data[i]); |
| 183 | if (char_len == 0 || i + char_len > data_len) { // invalid byte or incomplete glyph |
| 184 | set_error_for_invalid_utf(context, data[i]); |
| 185 | return 0; |
| 186 | } |
| 187 | for (int j = 1; j < char_len; ++j) { |
| 188 | if ((data[i + j] & 0xC0) != 0x80) { // bytes following head-byte of glyph |
| 189 | set_error_for_invalid_utf(context, data[i + j]); |
| 190 | return 0; |
| 191 | } |
| 192 | } |
| 193 | ++count; |
| 194 | } |
| 195 | return count; |
| 196 | } |
| 197 | |
| 198 | // Count the number of utf8 characters, ignoring invalid char, considering size 1 |
| 199 | FORCE_INLINE |
no outgoing calls