3 ms·

Unicode regexes is a pretty small edge case. Is it because Perl is awesome, or because Unicode sucks?

• ethelward 9y ago
  > Unicode regexes is a pretty small edge case Except for any other language than english.
• jandrese 9y ago
  Most people who understand what a can of worms those systems are start leaking a little grey matter from their ears when someone mentions the phrase "Unicode Regex". I have mental images of driving a wildly overpowered monster truck with no brake lines over a minefield. And then I start to consider the implications of passing the /i flag in...
  • xeeeeeeeeeeenu 9y ago
    Yes, and that's the reason why it's important to choose a tool that gets Unicode correctly. Perl is an example of such a tool.
    • jandrese 9y ago
      Even if the language handles it correctly the lowly programmer is playing with white hot fire. Unicode adds so many edge cases it's basically a serrated knife in the middle of your code. Will your regex match correctly in the case that the text shifts direction halfway through? Will the case insensitive search work in languages where case is ambiguous (where uc(lc(x)) != x)? Will you be able to match letters that are effectively identical but on different code points? Do you want to? Will it get confused by furigana? Unicode's goal of encoding every language in the world as is means it encompasses every bizarre thing people have ever done to a written language. It is almost impossible to actually do anything with a block of unicode codepoints except treat it like a big binary blob and hope that you never have to personally deal with the strings contained therein. Most everybody gets it wrong in one way or another.