I was trying to use the external runner, in a reasonably simple way:
from within a project (QEMU) build dir, I'd run:

 $ avocado run --external-runner=tests/qemu-iotests/check \
   --external-runner-chdir=runner -- 001 002

But, what should work, did not.  The problem is with the non-absolute
path to the external runner, that is needed because we're changing
directories.

Absolute paths won't hurt otherwise (when no chdir is done), so it's
always done here.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

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

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

Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

The iterating through module statements and lookout for classes is
strictly tied to module status (avocado imports). Let's move it into the
"AvocadoModule" and only treat classes in "find_avocado_tests".
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Some of our inputs were not protected against unwanted input, resulting
in interrupted discovery and no tests reported. Let's fix that and make
sure we keep discovering them.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

This change enables pylint check w0631. Since the only occurrence is on
external spark library, it was disabled specifically.
Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

This check makes sure the excepts in try/except are not duplicated
regarding the exception type.
Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

We'll need to pass those information to other functions, let's bundle
them to make the code clearer. (keep using "path" internally as
accessing local variable is faster)
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The "test_import" and "mod_import" are not really necessary as we store
"test_import_name" and "mod_import_name", which can never be empty
string. Let's depend solely on names.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

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

The build directory of the main Avocado module is already cleaned by
the `clean` rule, but when plugins contain a build directory, they are
not currently cleaned.

Sometimes this leads to annoying `inspekt` failures, because previous
different versions of the plugin may be in the build directory, and
the error message only mentions the name of the module/class/method
violating the check, but not the exact file name.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Besides enabling the check, this change also makes ISO9660PyCDLib
inherit from MixInMntDirMount so it has the implementation of
mount_dir() property.
Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

The W0223 pylint check warns about any abstract method which is not
implemented by subclasses. Currently this warning are being raised for
VMImageHtmlParser (Python 3.7). However it seems it happens because an
open Python bug[1] about HtmlParser not implementing the error()
abstract method of (inner) superclass _markupase.ParserBase().

To enable avocado project checks for cases like that, this change
disables the related check only for the specific case.

[1] - https://bugs.python.org/issue31844Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

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

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

This pylint checks for possible redefining outer names. Some occurrences
was preserved with disable flag because they would broke some public
api, like method/functions names or parameters.
Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

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

The W0613 check for unnused arguments. The most of the current cases
seems legitmate so the proper flag was added in each line so pylint can
ignore it.
Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

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

This pylint check makes sure the log message is not built in log
statement itself, but uses the proper argument from logger to do it
lazy.
Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

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

The dependencies for selftests and CI execution was split in two
requirements files. It causes issues with libs being in different
versions when tests are executed locally and in CI environment. To
reduce theses issues and others the changes bellow was done.

All dependencies listed in requirements.txt are now specified only in
setup.py properly. This way we avoid duplications or missing
requirements that are necessary for avocado.

It also could be done keeping only requirements.txt file and loading it
in setup.py module. However, since avocado-framework have not clear use
case of requirements being installed via `pip install requirements.txt`,
it's better keep the dependencies only in setup module.

It was removed the requirements file only for CI purpose and
updates the requirements for selftest. This way we can have proper
dependencies installed when run selftests both in local or CI
environements.

The requirements-selftest was updated: Setuptools was removed since it's
already a dependency from avocado itself. Pep8 was removes since it's
already a dependecy from inspektor. Since we run lint checks by
inspektor it's better let it manage de pep8 (or pycodestyle) dependency
Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

First of all, a disclaimer: this is hackish.  But, it's the simplest
and thus best fix I could think of.

What happens here is that this module makes no distintcion between
what is a base class "abstract" class, and a concrete class.  With the
goal of reducing code duplication, a goal that will remain valid here
and in other modules, it made sense to create a common Fedora class.

Now, because only being a ImageProviderBase suffices, the
FedoraImageProviderBase class, which is not intended to be used as a
concrete (final) class, is being carried along and its "name" is
checked, causing crashes.

This introduces the ad-hoc, hackish, notion, that a concrete
ImageProvider is one that contains a name.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Based on the distro module API, it should be possible to get the
version and release simply by providing a regex that has either one or
two groups.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

While running avocado in ubuntu system distro.detect().name has output
"unknown".  This is probably due to the removal of StdLibProbe.  Let's
add a proper Ubuntu probe.  Before this:

    >>> distro.detect().version and distro.detect().release
    0
    >>> from avocado.utils import distro
    >>> distro.detect().name
    unknown
    >>> distro.detect().version
    0
    >>> distro.detect().release
    0

After this patch:

    >>> from avocado.utils import distro
    >>> print distro.detect().name
    Ubuntu
    >>> print distro.detect().version
    18
    >>> print distro.detect().release
    04
Reported-by: NKamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Signed-off-by: NPraveen K Pandey <praveen@linux.vnet.ibm.com>
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

The reason for this to exist was supposedly identify (inheritance)
checks, but none of these actually exist.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

Signed-off-by: NCaio Carrara <ccarrara@redhat.com>

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

To decode into the correct encoding and work properly across different
Python versions.

While at it, improve the docstring.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

To the relevant code block that can, for a reason we expect, raise
such exception.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The current sysinfo end_test() hook is badly positioned, in a way that
it only gets called if no exception is raised (that is, tests end with
PASS).
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

Signed-off-by: NCaio Carrara <ccarrara@redhat.com>