Python 3.13 test failure: test_ip_address_ipv6_ipv4_mapped

[suve]

In Fedora Rawhide (the development branch / "eternal alpha" version) we've recently updated to Python 3.13.0b4 pre-release. With this new version, the following test failure can be observed:

FAIL: test_ip_address_ipv6_ipv4_mapped (tests.test_ip_address.MyTests.test_ip_address_ipv6_ipv4_mapped)
Test ipv6 address with ipv4 mapped string
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/builddir/build/BUILD/jc-1.25.3-build/jc-1.25.3/tests/test_ip_address.py", line 95, in test_ip_address_ipv6_ipv4_mapped
    self.assertEqual(jc.parsers.ip_address.parse(data, quiet=True), expected)
    ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError: {'ver[45 chars]ffff:192.168.1.35', 'ip_compressed': '::ffff:1[2371 chars]11'}} != {'ver[45 chars]ffff:c0a8:123', 'ip_compressed': '::ffff:c0a8:[2336 chars]11'}}
  {'bin': {'broadcast': '00000000000000000000000000000000000000000000000000000000000000000000000000000000111111111111111111000000101010000000000100100011',
           'first_host': '00000000000000000000000000000000000000000000000000000000000000000000000000000000111111111111111111000000101010000000000100100011',
           'hostmask': '00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000',
           'ip': '00000000000000000000000000000000000000000000000000000000000000000000000000000000111111111111111111000000101010000000000100100011',
           'last_host': '00000000000000000000000000000000000000000000000000000000000000000000000000000000111111111111111111000000101010000000000100100011',
           'netmask': '11111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111',
           'network': '00000000000000000000000000000000000000000000000000000000000000000000000000000000111111111111111111000000101010000000000100100011'},
-  'broadcast': '::ffff:192.168.1.35',
?                        - ------- -

+  'broadcast': '::ffff:c0a8:123',
?                       +++++

   'cidr_netmask': 128,
-  'dns_ptr': '5.3...1...8.6.1...2.9.1.f.f.f.f.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.ip6.arpa',
?              --  ^   ^   ^ ^ ^^^^^^^

+  'dns_ptr': '3.2.1.0.8.a.0.c.f.f.f.f.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.ip6.arpa',
?                ^   ^   ^ ^ ^

-  'first_host': '::ffff:192.168.1.35',
?                         - ------- -

+  'first_host': '::ffff:c0a8:123',
?                        +++++

   'hex': {'broadcast': '00:00:00:00:00:00:00:00:00:00:ff:ff:c0:a8:01:23',
           'first_host': '00:00:00:00:00:00:00:00:00:00:ff:ff:c0:a8:01:23',
           'hostmask': '00:00:00:00:00:00:00:00:00:00:00:00:00:00:00:00',
           'ip': '00:00:00:00:00:00:00:00:00:00:ff:ff:c0:a8:01:23',
           'last_host': '00:00:00:00:00:00:00:00:00:00:ff:ff:c0:a8:01:23',
           'netmask': 'ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff:ff',
           'network': '00:00:00:00:00:00:00:00:00:00:ff:ff:c0:a8:01:23'},
   'hostmask': '::',
   'hosts': 1,
   'int': {'broadcast': 281473913979171,
           'first_host': 281473913979171,
           'ip': 281473913979171,
           'last_host': 281473913979171,
           'network': 281473913979171},
-  'ip': '::ffff:192.168.1.35',
+  'ip': '::ffff:c0a8:123',
-  'ip_compressed': '::ffff:192.168.1.35',
?                            - ------- -

+  'ip_compressed': '::ffff:c0a8:123',
?                           +++++

-  'ip_exploded': '0000:0000:0000:0000:0000:ffff:192.168.1.35',
?                                                 - ------- -

+  'ip_exploded': '0000:0000:0000:0000:0000:ffff:c0a8:0123',
?                                                ++++++

-  'ip_split': ['0000', '0000', '0000', '0000', '0000', 'ffff', '192.168.1.35'],
?                                                                 - ------- -

+  'ip_split': ['0000', '0000', '0000', '0000', '0000', 'ffff', 'c0a8', '0123'],
?                                                                +++++++++

   'ipv4_mapped': '192.168.1.35',
   'is_global': False,
   'is_link_local': False,
   'is_loopback': False,
   'is_multicast': False,
   'is_private': True,
   'is_reserved': True,
   'is_unspecified': False,
-  'last_host': '::ffff:192.168.1.35',
?                        - ------- -

+  'last_host': '::ffff:c0a8:123',
?                       +++++

   'max_prefix_length': 128,
   'netmask': 'ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff',
-  'network': '::ffff:192.168.1.35',
?                      - ------- -

+  'network': '::ffff:c0a8:123',
?                     +++++

   'scope_id': None,
   'six_to_four': None,
   'teredo_client': None,
   'teredo_server': None,
   'version': 6}

----------------------------------------------------------------------

I'm fully aware that Python 3.13.0b4 is a pre-release, which makes testing difficult, so this issue will probably have to wait until the final 3.13.0 release. You can consider this a heads-up. ;-)

For reference, the Python 3.13 "what's new" page currently lists the following changes:

ipaddress

• Add the ipaddress.IPv4Address.ipv6_mapped property, which returns the IPv4-mapped IPv6 address. (Contributed by Charles Machalow in gh-109466.)

• Fix is_global and is_private behavior in IPv4Address, IPv6Address, IPv4Network and IPv6Network.

Sorry for not contributing a fix, but my Python knowledge doesn't really extend past FizzBuzz level.

[kellyjonbrazil]

Hi there - thanks for reporting this. I believe this is a duplicate of #572 and will be fixed in the next release.

Edit: unless this is a new issue from 3.13 we need to address? It looks like the other issue fixed a change in 6to4 addresses but this is affecting mapped addresses. Will have to investigate.

[suve]

I think this is a new issue. It affects a different test, and applying the patch from PR #573 does not help.

[kellyjonbrazil]

Looks like this is caused by:

python/cpython#87799

python/cpython#29345

[kellyjonbrazil]

I'm working on a fix, but have opened a cpython issue that I think might need to be addressed as well: python/cpython#123409

[kellyjonbrazil]

python/cpython#123409 is in process of being patched: python/cpython#123419

[kellyjonbrazil]

It looks like the CPython issue has been addressed and will be fixed in the next 3.13 release. I have committed some changes that should fix this issue and tests when the latest Python release comes out.

71af0c5

[suve]

So, now that Python 3.13 is out - jc v1.25.3 builds fine and all tests pass, provided one backports the following two commits: 9eb4df3 and 71af0c5.

As such, this issue should be fixed for everyone in the next release.