Thread: User defined literal

I don't think you can use unicode characters in identifiers in any version of the C++ standard. The '\u' notation only works inside literal char and string values, i.e. in quotes.

Maybe I was mislead then, and they wrote it here: User-defined literals (since C++11) - cppreference.com

As to show that you are meant to substitute the unicode notation with the actual replacement before trying to compile...? Doing so still gives me an error though:

Code:

error: stray '\200' in program

This just throws me off guard if that is the case. I still don't know why you'd ever use such a character in a literal suffix, but I was just curious. I'd probably use something else anyways to avoid using unicode characters as literals.. (Ex: _pi)

I still don't get as to how they comment this as "OK".

According to this stack overflow question, it's allowed, but if you're using GCC, you have to tell it what encoding you're using. This stack overflow question seems to show the command line options you'll need.

I wrote this in C, but this seems to work:

Code:

#include <stdio.h>

// GCC: -fextended-identifiers
#define assign(ucn, val) ucn = val

int main(void)
{
  int \u0401;
  assign(\u0401, 4);
  printf("%d\n", \u0401);
}

And seems to be compliant from what I've read. I didn't change any of the charset compiler flags or add them to the command line because the default as UTF-8 seems to be fine already. What I needed was commented in the code above. Although I haven't tested it, but those may become relevant when I try to actually use unicode in my source.