help fixing up a regex - \r\n isn't being detected

Question

Folks,

I have method that returns true if all characters are "legal" and false if a single character is "illegal". The definition is below (legal = letters, numbers, and some characters like $, - , %, etc). I want a newline and/or carriage return character to be "illegal". However, the method below thinks that its legal. How can I fix-up?

  private static bool NameHasAllLegalCharacters( string name )
    {
        var regexAlphaNum = new Regex( @"[^a-zA-Z0-9#\$%\^&\*\(\)\._\-\+=\[\]/<>{}:\s]" );
        return !regexAlphaNum.IsMatch( name );
    }

Answer 1

\s matches all whitespace, not just the ASCII space.

Answer 2

The \s class matches all whitespace, and \r and \n are classed as whitespace.

Your regex will need to be more specific about which types of whitespace are allowed.

Answer 3

Given that there're thousands of characters out there (thanks to Unicode) I would not even try to use a blacklist approach. Instead, reverse your logic and test whether all chars belong to your whitelist:

private static bool NameHasAllLegalCharacters( string name )
{
    var regexAlphaNum = new Regex( @"^[a-zA-Z0-9#$%........]*$" );
    return regexAlphaNum.IsMatch( name );
}

Answer 4

\s matches this character set: [\f\n\r\t\v\x85\p{Z}]. You could simply enumerate this without \n and \r, like this:

@"[^a-zA-Z0-9#\$%\^&\*\(\)\._\-\+=\[\]/<>{}:\f\t\v\x85\p{Z}]"

I know it's ugly, but it should work.