Color on Windows

[nemequ]

Color on Windows

[nemequ]

I'm okay with b342b1b, but please use _getenv_s or _dupenv_s (I would probably use the former with a small stack based buffer, but whatever) to avoid that CRT warning.

I guess it would be possible to use SetConsoleTextAttribute to set colors on Windows, but TBH I'm not sure it's worth it. I don't think anyone takes the default command prompt seriously, and if conemu supports colors that's good enough for me.

[nemequ]

SetConsoleTextAttribute wouldn't work for people who are using --color always but redirecting the output (like we do to colorize AppVeyor's output), so it's probably better to avoid that.

Also, since you only want to check that ANSICON != NULL, you don't even need a buffer. Something like this should work:

if (isatty(fileno(stream))) {
  size_t ansicon_size;
  getenv_s(&ansicon_size, NULL, 0, "ANSICON");
  return ansicon_size != 0;
}
return false;

[jibsen]

I've updated the branch, but getenv_s is an extension to C11, so now some of the GCC builds break.

Quite oddly, only the mingw builds appear to break.

[nemequ]

Only mingw breaks because the getenv call is in guarded by an #if defined(_WIN32). I think everywhere else it's probably okay to assume that ttys support color. AFAIK Windows is the only platform which implements Annex K, but they're also the only ones who complain about getenv…

I pushed a fix which will just call the standard getenv on mingw, getenv_s on non-mingw windows, and still just use isatty everywhere else.

[jibsen]

Ah, of course -- thanks.