Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

* https://github.com/avocado-framework/avocado:
  fail_on(): move internal implementation to avocado.core.decorators

The `avocado.core.decorators` module has been recently added with the
`skip*` family of decorators.  Still, we have `fail_on()`, also a
decorator, defined in `avocado.core.exceptions`.

For consistency, let's put them in the same module.  There are no
changes to users of the Avocado Test API.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

Hash files (%s-CHECKSUM) were always written into the last cache
directory specified in `datadir.paths.cache_dirs`. This failed if that
directory does not exist or is not writable.

Fix this by setting self.hash_file again when deciding were to download
the asset to.
Signed-off-by: NLars Karlitski <lars@karlitski.net>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

reduced the number of calls to join/map

Reference: https://trello.com/c/ZcV36ncdSigned-off-by: NMaurice Saldivar <maurice.a.saldivar@hpe.com>

As a matter of fact we lack the sysinfo documentation. Let's add a new
section as it's a standalone plugin with a basic description.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

On broken/unresponsive machines it might happen that some of the sysinfo
commands don't finish. This allows setting "commands_timeout" in
[sysinfo.collect] section to prevent hangs.

Note this is not for all sysinfo collection, only for the commands
execution, which is currently the issue.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

increase read buffer size to avoid deadlock caused
by blocked pipe.
Signed-off-by: NXu Tian <xutian@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

We're not checking if the there's an owner for the STDOUT when
displaying the runner connection information.  This makes
command lines that output, say, xunit to the STDOUT, to become
corrupted with a "LOGIN: " (or similar) line.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Now that Python 2.6 is not supported anymore, we can not officially
support EL6 as a platform for RPM packages.  Let's drop all special
requirements for that platform.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The unittest standard library on Python 2.7 an later has everything
that the unittest2 backport is supposed to have.  Let's then drop all
the conditional imports of unittest2 and stick with unittest.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The main reason for supporting Python 2.6 was the EL6 distros.
Although they are still important, our users are running a combination
of Avocado LTS and Avocado-VT.  That allows us to drop support for
Python 2.6 on master (and further Avocado releases).

Also, with Python 2.6 being dropped as a supported platform, the
version of the stevedore library can be simplified.  Let's reflect
that on both setup.py and requirements*.txt.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The libvirt Python bindings are very very hard to install from source.
On Travis, let's rely on pre-built versions from official distro
packages.

This has not been an issue so far because of the conditional treatment
given on `avocado/core/virt.py`.  When the import fails,
`VIRT_CAPABLE` was set to `False`.

The goal of this change is twofold:

1) Enable the dependency on Travis so that the tests from
`selftests/unit/test_vm.py` are run.

2) Prepare for the split of the runner packages, in which there will
be no conditionals of the libvirt import, since the installation of
the package should mean that the dependencies are also fulfilled.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

There has been intermitent errors on this specific test while running
on Travis.  We should not let this impact our work, so until we
identify the cause of the failure, let's skip this test *on Travis*
only.  This test will still be run as part of `make check-full`.

There's also a Trello card to not let us forget about this test.

Reference: https://trello.com/c/yXtgDRjkSigned-off-by: NCleber Rosa <crosa@redhat.com>

added some code to fetch some more info from vpd
Signed-off-by: NNaresh Bannoth <nbannoth@in.ibm.com>

Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NCleber Rosa <crosa@redhat.com>

The SIMPLE test used here contains SHELL code, so it doesn't make sense
to name with a .py extension.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Avocado does not support self.skip() to be called from inside a test.
This is a design decision that users have to comply with. On the other
hand, users keep requesting alternative ways to skip tests.

This patch creates the decorators skip, skipIf and skipUnless to be
used in test functions. Corresponding exceptions will be raised in
those cases.

Reference: https://trello.com/c/JK5Z5dqlSigned-off-by: NAmador Pahim <apahim@redhat.com>

It happened couple of times that running our selftests produced a new
entries in the default results directory. Let's use this optional check
to avoid such issue.

To enable this check one has to set "AVOCADO_RESULTSDIR_CHECK" env
variable. This check is enabled in travis as I do not expect travis to
run avocado tests while running the selfcheck, unlike on developer
computers.

Note the check is rather simple and only assumes the default results
directory. For selfcheck I think it's enough.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

This PR adjusts our tests to avoid producing results in the default
results directory (usually caused by instantiating Job or Test without a
results dir).

One of the tests still produces test results as it is essential for the
test execution, but it goes through the produced tests and removes the
base directory.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

So far we only executed the first argument, but we might want to execute
complex cmdlines instead just one command. Let's use $* instead of $1.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

There's one functional test that is missing "--sysinfo=off".  I
happened to find this because lspci on my system hangs when running
under X.  That test would timeout at the sysinfo collection.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

The pystache package should only be a requirement for the HTML
result plugin, not for the base avocado package.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NAmador Pahim <apahim@redhat.com>

When we use args:--mux-yaml or --multiplex if the yaml file not existing it
will show more detail error messages

Reference: https://trello.com/c/hYCrS1oI/900-improve-yaml-to-mux-error-messagesSigned-off-by: Nmengalong <alongmeng@gmail.com>

Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

* https://github.com/avocado-framework/avocado:
  selftests: exclude E722 from pep8 check

pep8 v2.3.0 introduced the check for bare except clause (E722). Until
now, we don't explicitly disable the pep8 check we don't want because
inspect has a default exclude list (E501, E265, W601 and E402).

Let's make explicit the exclude list for the checks we don't want in our
selftests, adding to the list the new E722 check.
Signed-off-by: NAmador Pahim <apahim@redhat.com>

Signed-off-by: NCleber Rosa <crosa@redhat.com>

The "make link" feature goes through subdirs trying to either use
`setup.py` or `Makefile` to install all plugins. The problem is that
even when `Makefile` is present it might not support `make link` target
which results in failure. This patch instead of failure in this case
just lists "FAIL" followed by the module name.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>