Function uts46_valid_char

email_validator/syntax.py:463–490

(char: str)

def uts46_valid_char(char: str) -> bool:
    # By exhaustively searching for characters rejected by
    # for c in (chr(i) for i in range(0x110000)):
    #   idna.uts46_remap(c, std3_rules=False, transitional=False)
    # I found the following rules are pretty close.
    c = ord(char)
    if 0x80 <= c <= 0x9f:
        # 8-bit ASCII range.
        return False
    elif ((0x2010 <= c <= 0x2060 and not (0x2024 <= c <= 0x2026) and not (0x2028 <= c <= 0x202E))
          or c in (0x00AD, 0x2064, 0xFF0E)
          or 0x200B <= c <= 0x200D
          or 0x1BCA0 <= c <= 0x1BCA3):
        # Characters that are permitted but fall into one of the
        # tests below.
        return True
    elif unicodedata.category(chr(c)) in ("Cf", "Cn", "Co", "Cs", "Zs", "Zl", "Zp"):
        # There are a bunch of Zs characters including regular space
        # that are allowed by UTS46 but are not allowed in domain
        # names anyway.
        #
        # There are some Cn (unassigned) characters that the idna
        # package doesn't reject but we can, I think.
        return False
    elif "002E" in unicodedata.decomposition(chr(c)).split(" "):
        # Characters that decompose into a sequence with a dot.
        return False
    return True


class DomainNameValidationResult(TypedDict):