I think doug (he can correct me if I am wrong) was more wondering the
question of 'why would you want to use random data' in a test vs using
known fixed data. When a test fails it can be quite hard to identify the
reason if certain data that the test uses is randomized for each run.
Randomness might be useful, but it has always made me feel like
something is off when you see to much of it in tests...
-Josh
TommyLike Hu wrote:
Hmm,it's used to generate the ip address for validation or rule
checking, Meanwhile add some randomness. Of course it's unreasonable the
case you mentioned, please check the existed cases [1] and [2]
[1]
https://github.com/openstack/manila/blob/master/manila_tempest_tests/tests/api/base.py#L828
[2]
https://github.com/openstack/manila/blob/master/manila_tempest_tests/tests/api/test_replication.py#L76
Doug Hellmann <[email protected] <mailto:[email protected]>>于
2016年10月18日周二 上午12:01写道：
Excerpts from TommyLike Hu's message of 2016-10-17 14:46:36 +0000:
> It's used in testcase already, and basic codes is from here:
>
https://github.com/openstack/manila/blob/master/manila_tempest_tests/utils.py#L93
OK, I guess the real question I had is why use *random* addresses.
Because that seems like a way to end up having two tests try to use the
same address at the same time. If that did happen, would it cause
conflicts or race conditions for the manila tests?
>
> Doug Hellmann <[email protected]
<mailto:[email protected]>>于2016年10月17日周一 下午10:13写道：
>
> > Excerpts from TommyLike Hu's message of 2016-10-17 09:56:15 +0000:
> > > When I handle some stuff related to Manila recently, I found
a case which
> > > may be suitable for Oslo, Anyhow I put it in the maillist so
it can be
> > > discussed before I put it in action.
> > > In testcase, we need a function(maybe 2) to generate random
ip address
> > (or
> > > network), also they should in the range of [1](ipv4
documentation range)
> > or
> > > [2](ipv6 documentation range), this is the draft code for ipv4:
> > >
> > > import six
> > > import random
> > >
> > >
> > > def rand_ipv4(network=False):
> > > """This uses the TEST-NET-3 range of reserved IP addresses."""
> > >
> > > test_net_3 = '203.0.113.'
> > > if network:
> > > host_length = random.randint(0, 8)
> > > address_segment = random.randint(0, 8 - host_length)
> > > address_segment <<= host_length
> > > address = test_net_3 + six.text_type(address_segment)
> > > address = '/'.join((address,
> > > six.text_type(32 - host_length)))
> > > else:
> > > address = test_net_3 +
six.text_type(random.randint(0, 255))
> > > return address
> > >
> > > The primary use case here is writing testcases, I am not
sure whether it
> > > is suitable here, or maybe I misunderstood the intention of
TEST-NET-3,
> > > please leave any comment you like.
> > >
> > >
> > > [1] https://tools.ietf.org/html/rfc5737
> > > [2] https://tools.ietf.org/html/rfc5156
> >
> > In what way are you using random addresses in tests?
> >
> > Doug
> >
> >
__________________________________________________________________________
> > OpenStack Development Mailing List (not for usage questions)
> > Unsubscribe:
[email protected]?subject:unsubscribe
<http://[email protected]?subject:unsubscribe>
> > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
> >
__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe:
[email protected]?subject:unsubscribe
<http://[email protected]?subject:unsubscribe>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: [email protected]?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: [email protected]?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
