3 ms·

Exactly my point. Most modern emojis cannot rely on pure codepoints to be extracted.

• arcticbull 7y ago
  Makes sense! Did you end up implementing normalization for your sub-string find, or did you work around it some other way? I couldn't seem to see it when skimming.
  • clauderoux 7y ago
    You can have a look on: s_is_emoji...
    • clauderoux 7y ago
      In https://github.com/naver/tamgu/blob/master/include/conversion.h https://github.com/naver/tamgu/blob/master/include/conversio..., I have implemented a class: agnostring which derives from "std::string". There are some methods to traverse a UTF8 string: begin(): to initialize the traversal end() : is true when the string is fully traversed next(): which goes to the next character and returns the current character. s.begin(); while (!s.end()) { u = s.next(); }
      • arcticbull 7y ago
        Very cool, thanks!