Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Clarifications to tests(7)



Martin Husemann <martin%duskware.de@localhost> writes:

> On Mon, May 19, 2014 at 12:28:21PM -0400, Greg Troxel wrote:
> [correct statements about root/unprivileged runs]
>
>> +Ideally, tests are self-contained and do not either depend on or
>> +perturb the host environment, aside from skipping tests when optional
>> +facilities are not available.
>
> This would be ideal, but basically impossible to implement. All tests do
> cleanup work after themselves, but for example tests exercising the vnd
> devices need to configure and unconfigure vnds, wich might interfere with
> other uses of vnd on the host.

Do we believe

  tests do a reasonable job e.g. of looking for an unused vnd, so this
  should be mostly ok?

  if tests interfere with a reasonable config, that's a bug?

> Other tests might load and unload kernel modules (just to test this), so there
> is always temporary flux on the test host.
>
> No permanent configuration changes or disturbances are to be expected though.

I can add some caveats.   My real point is: If someone is going to run
the tests, do we mean:

  It's reasonable to run them on your system while it is up multiuser.

or

  It is not safe or reliable to run tests multiuser; boot the system
  into single-user mode before running tests.


It sounds like you are saying it's ok to run multiuser, and there might
be trouble, but if so that's reportable as a bug.

Attachment: pgp9zF2Nv3RL1.pgp
Description: PGP signature



Home | Main Index | Thread Index | Old Index