bazarr/libs/urllib3/util/ssl_match_hostname.py

"""The match_hostname() function from Python 3.5, essential when using SSL."""

# Note: This file is under the PSF license as the code comes from the python
# stdlib.   http://docs.python.org/3/license.html
# It is modified to remove commonName support.

from __future__ import annotations

import ipaddress
import re
import typing
from ipaddress import IPv4Address, IPv6Address

if typing.TYPE_CHECKING:
    from .ssl_ import _TYPE_PEER_CERT_RET_DICT

__version__ = "3.5.0.1"


class CertificateError(ValueError):
    pass


def _dnsname_match(
    dn: typing.Any, hostname: str, max_wildcards: int = 1
) -> typing.Match[str] | None | bool:
    """Matching according to RFC 6125, section 6.4.3

    http://tools.ietf.org/html/rfc6125#section-6.4.3
    """
    pats = []
    if not dn:
        return False

    # Ported from python3-syntax:
    # leftmost, *remainder = dn.split(r'.')
    parts = dn.split(r".")
    leftmost = parts[0]
    remainder = parts[1:]

    wildcards = leftmost.count("*")
    if wildcards > max_wildcards:
        # Issue #17980: avoid denials of service by refusing more
        # than one wildcard per fragment.  A survey of established
        # policy among SSL implementations showed it to be a
        # reasonable choice.
        raise CertificateError(
            "too many wildcards in certificate DNS name: " + repr(dn)
        )

    # speed up common case w/o wildcards
    if not wildcards:
        return bool(dn.lower() == hostname.lower())

    # RFC 6125, section 6.4.3, subitem 1.
    # The client SHOULD NOT attempt to match a presented identifier in which
    # the wildcard character comprises a label other than the left-most label.
    if leftmost == "*":
        # When '*' is a fragment by itself, it matches a non-empty dotless
        # fragment.
        pats.append("[^.]+")
    elif leftmost.startswith("xn--") or hostname.startswith("xn--"):
        # RFC 6125, section 6.4.3, subitem 3.
        # The client SHOULD NOT attempt to match a presented identifier
        # where the wildcard character is embedded within an A-label or
        # U-label of an internationalized domain name.
        pats.append(re.escape(leftmost))
    else:
        # Otherwise, '*' matches any dotless string, e.g. www*
        pats.append(re.escape(leftmost).replace(r"\*", "[^.]*"))

    # add the remaining fragments, ignore any wildcards
    for frag in remainder:
        pats.append(re.escape(frag))

    pat = re.compile(r"\A" + r"\.".join(pats) + r"\Z", re.IGNORECASE)
    return pat.match(hostname)


def _ipaddress_match(ipname: str, host_ip: IPv4Address | IPv6Address) -> bool:
    """Exact matching of IP addresses.

    RFC 9110 section 4.3.5: "A reference identity of IP-ID contains the decoded
    bytes of the IP address. An IP version 4 address is 4 octets, and an IP
    version 6 address is 16 octets. [...] A reference identity of type IP-ID
    matches if the address is identical to an iPAddress value of the
    subjectAltName extension of the certificate."
    """
    # OpenSSL may add a trailing newline to a subjectAltName's IP address
    # Divergence from upstream: ipaddress can't handle byte str
    ip = ipaddress.ip_address(ipname.rstrip())
    return bool(ip.packed == host_ip.packed)


def match_hostname(
    cert: _TYPE_PEER_CERT_RET_DICT | None,
    hostname: str,
    hostname_checks_common_name: bool = False,
) -> None:
    """Verify that *cert* (in decoded format as returned by
    SSLSocket.getpeercert()) matches the *hostname*.  RFC 2818 and RFC 6125
    rules are followed, but IP addresses are not accepted for *hostname*.

    CertificateError is raised on failure. On success, the function
    returns nothing.
    """
    if not cert:
        raise ValueError(
            "empty or no certificate, match_hostname needs a "
            "SSL socket or SSL context with either "
            "CERT_OPTIONAL or CERT_REQUIRED"
        )
    try:
        # Divergence from upstream: ipaddress can't handle byte str
        #
        # The ipaddress module shipped with Python < 3.9 does not support
        # scoped IPv6 addresses so we unconditionally strip the Zone IDs for
        # now. Once we drop support for Python 3.9 we can remove this branch.
        if "%" in hostname:
            host_ip = ipaddress.ip_address(hostname[: hostname.rfind("%")])
        else:
            host_ip = ipaddress.ip_address(hostname)

    except ValueError:
        # Not an IP address (common case)
        host_ip = None
    dnsnames = []
    san: tuple[tuple[str, str], ...] = cert.get("subjectAltName", ())
    key: str
    value: str
    for key, value in san:
        if key == "DNS":
            if host_ip is None and _dnsname_match(value, hostname):
                return
            dnsnames.append(value)
        elif key == "IP Address":
            if host_ip is not None and _ipaddress_match(value, host_ip):
                return
            dnsnames.append(value)

    # We only check 'commonName' if it's enabled and we're not verifying
    # an IP address. IP addresses aren't valid within 'commonName'.
    if hostname_checks_common_name and host_ip is None and not dnsnames:
        for sub in cert.get("subject", ()):
            for key, value in sub:
                if key == "commonName":
                    if _dnsname_match(value, hostname):
                        return
                    dnsnames.append(value)

    if len(dnsnames) > 1:
        raise CertificateError(
            "hostname %r "
            "doesn't match either of %s" % (hostname, ", ".join(map(repr, dnsnames)))
        )
    elif len(dnsnames) == 1:
        raise CertificateError(f"hostname {hostname!r} doesn't match {dnsnames[0]!r}")
    else:
        raise CertificateError("no appropriate subjectAltName fields were found")