There's actually at least two different tests that sometimes fail: the other is ( https://people.debian.org/~sanvila/build-logs/execnet/execnet_1.4.1-3_amd64-20170207T152032Z ):

=================================== FAILURES =================================== _______________________ TestMakegateway.test_popen_nice ________________________

self = <test_xspec.TestMakegateway object at 0x7f57cbff97b8>
makegateway = <bound method Group.makegateway of <Group ['gw1']>>

    @pytest.mark.skipif("not hasattr(os, 'nice')")
    def test_popen_nice(self, makegateway):
        gw = makegateway("popen")

        def getnice(channel):
            import os
            if hasattr(os, 'nice'):
                channel.send(os.nice(0))
            else:
                channel.send(None)
        remotenice = gw.remote_exec(getnice).receive()
        gw.exit()
        if remotenice is not None:
            gw = makegateway("popen//nice=5")
            remotenice2 = gw.remote_exec(getnice).receive()
>           assert remotenice2 == remotenice + 5
E           assert 0 == (0 + 5)

testing/test_xspec.py:136: AssertionError
============================ pytest-warning summary ============================

If they are single-CPU-only problems, a potential workaround (warning: untested) is @pytest.mark.skipif(len(os.sched_getaffinity(0))==1,"unreliable on single-CPU systems")

Reply via email to