This should work for ASCII:
constexpr bool IsDigit(char c) { return c >= '0' && c <= '9'; } // ASCII only
If you need to keep your prototype:
template<char C>
constexpr bool IsDigit() { return C >= '0' && C <= '9'; } // ASCII only
For wchar_t
support you could try something like:
constexpr bool IsDigit(wchar_t c)
{
return (c >= L'0' && c <= L'9') ||
(c >= L'\u0660' && c <= L'\u0669') || // Arabic-Indic
(c >= L'\u06F0' && c <= L'\u06F9') || // Extended Arabic-Indic
(c >= L'\u07C0' && c <= L'\u07C9') || // NKO
(c >= L'\u0966' && c <= L'\u096F') || // Devanagari
(c >= L'\u09E6' && c <= L'\u09EF') || // Bengali
(c >= L'\u0A66' && c <= L'\u0A6F') || // Gurmukhi
(c >= L'\u0AE6' && c <= L'\u0AEF') || // Gujarati
(c >= L'\u0B66' && c <= L'\u0B6F') || // Oriya
(c >= L'\u0BE6' && c <= L'\u0BEF') || // Tamil
(c >= L'\u0C66' && c <= L'\u0C6F') || // Telugu
(c >= L'\u0CE6' && c <= L'\u0CEF') || // Kannada
(c >= L'\u0D66' && c <= L'\u0D6F') || // Malayalam
(c >= L'\u0E50' && c <= L'\u0E59') || // Thai
(c >= L'\u0ED0' && c <= L'\u0ED9') || // Lao
(c >= L'\u0F20' && c <= L'\u0F29'); // Tibetan
// Missing check for Myanmar, Khmer, Mongolian, Limbu, New Tai Lue,
// Tai Tham Hora, Tai Tham Tham, Balinese, Sundanese, Lepcha, Ol Chiki,
// Vai, Surashtra, Kayah, Javanese, Cham, Meetei Mayek, Osmanya, Brahmi,
// Sora, Chakma, Sharada, Takri, Mathematical.
// For codes see http://www.unicode.org/ucd/
}