Open kitterma opened 2 months ago
I can delay the release a week no problem, no need to work on stuff on your vacation! I agree it's suboptimal that there are live tests at all, and that they can be flaky in some environments. Hotel networks are notoriously bad for DNS testing!
See your other ticket for info about DDR. The dangling CNAME test failing is usually caused by some intermediate DNS software mangling the answer or not doing the usual thing. Dnspython tries to disable the test when it knows the system is broken, but I'm sure our list of broken things is not comprehensive. This test should also be made "non-live" in the future to solve this problem.
Describe the bug I would expect tests that are dependent on external environment factors to either pass or be skipped when they cannot be executed. I'm currently working on packaging dnspython 2.7.0 release candidate for Debian and am having problems with a few tests. The unusual part is that I'm traveling right now, so I'm limited to either hotel WiFi or tethering to my phone.
If I am connected to the hotel WiFi, the ddr tests fail:
If am tethered via my phone the dangling target tests fail:
Since all tests pass in one environment or the other, I'm confident there's not an actual dnspython bug, but it would be nice to have the skip conditions work so that these are just skipped when they can't be run. Help!
To Reproduce Not sure. Looking for guidance.
Context (please complete the following information):