This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author emptysquare
Recipients emptysquare, gvanrossum, martin.panter, pyptr2, python-dev, vstinner, yselivanov
Date 2016-06-12.13:50:24
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1465739424.26.0.88870918022.issue27136@psf.upfronthosting.co.za>
In-reply-to
Content
Thanks Martin. That test verifies behavior that I observe in Mac OS 10.10 and other modern platforms:

>>> socket.getaddrinfo('::2', 80, socket.AF_INET6, socket.SOCK_STREAM, socket.IPPROTO_TCP)
[(30, 1, 6, '', ('::0.0.0.2', 80, 0, 0))]

Investigating, I wrote a C program to call getaddrinfo on my Mac OS X Tiger x86 virtual machine, and indeed it resolves "::2" with family AF_INET6 in "hints" as an IPv4 address, "0.0.0.2". However, the same setup resolves "::1" as an IPv6 address, "::1".

Someone who knows more about IPv6 than I might guess the cause?

In any case, I wonder if replacing "::2" with "::1" at test_base_events.py:1188, and replacing the regex '::(0\.)*2' with '::(0\.)*1' at line 1195 would fix the test for Tiger.

Could you try that please? I'm having trouble compiling Python for Tiger so I don't know how to test it on my virtual machine.
History
Date User Action Args
2016-06-12 13:50:24emptysquaresetrecipients: + emptysquare, gvanrossum, vstinner, python-dev, martin.panter, yselivanov, pyptr2
2016-06-12 13:50:24emptysquaresetmessageid: <1465739424.26.0.88870918022.issue27136@psf.upfronthosting.co.za>
2016-06-12 13:50:24emptysquarelinkissue27136 messages
2016-06-12 13:50:24emptysquarecreate