There are more plugins on pip than the ones we're listing.  Let's
update the list of the currently available plugins.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

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

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

netifaces is an external library, that is not trivially installable by
pip (requires a C compiler).  Let's make the test conditional to the
presence of the netifaces library.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

In the parallel execution, not honoring the Python version used
elsewhere may cause unexpected issues.  One is an error situation in
which Python 3's unittest.loader can not load an exception definition
that exists only Python 2.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Each test has its own log file within a given job result dir,
located at "test-results/$(test-id)/debug.log".

If the test is an INSTRUMENTED test, using the standard Python logging
module/API, all produced content should be recorded into the test log
file.  This adds a handler that does exactly that to the logging
module root logger.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

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

The fake-vmstat process should run for infinity. Using wait with 1s
timeout will eventually destroy the test after the timeout, but cleaner
solution is to explicitly terminate it and then use the wait to make
sure it finished (or eventually nuke it)
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Let's publicize the COPR repo, both how to use it and the build
status, on our docs.
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>

Since it's not necessary to use proc.terminate() and call directly
process.wait() to get process finished, this change makes the proper
adjustments in test utils partition.
Signed-off-by: NCaio Carrara <ccarrara@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>

Because we're attempting to make Makefile.include reusable accross all
Avocado projects (first one is Avocado-VT), let's move this target
which is not applicable to Avocado-VT to the main Makefile.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The rule that *might* install pip on older Python installs
should be a requirement whenever pip is being used.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Some blocks are putting way too much whitespaces between the
title names and descriptions.  Let's limit that to two spaces.

This is a visual change only.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The recipe for building Debian packages have been removed from
the main Makefile and put into the contrib/ script.  Looks like
we missed this line.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Some pylint fixes and removal of not necessary import.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

With this implementations plugins will be able to modify the path list
for "avocado.core.settings". Mainly it should be used to register extra
configuration, but users should also be able to reorder/remove/modify
the path to default config files.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Originally we only processed files that existed and when they were read
we updated list of config locations and in case of no return list of
failed-to-read paths. Anyway the "config.read" function accepts list of
config files and returns list of successfully parsed files. Missing
files are automatically skipped and broken files raise exception, so
let's just simply create list of to-be-read files and feed it into the
"config.read" function.

As a result the list of not-read files actually works, still in this
version it's only displayed on "avocado config", where users can get
notion of all possible locations and in "avocado run" only parsed files
are displayed as it doesn't seem important enough to log them there to
me.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

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

Safeloader is supposed to keep the order of the classes and methods as
they're present on the source code file.  When not using a ordered
dict as an data structure, we count on pure luck.

This addresses the failure of the test_dont_detect_non_avocado test on
unit.test_loader.TestLoader.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

This docstring is a bit cryptic, let's makes it 100% clear it's only
used by unittests and the "--config" argument uses different approach.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

When the intree config exists, it's the same resource as pkg config.
Let's only process the pkg config and use pkg_resources properly to list
the directory with extra config files.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Currently in-tree is evaluated based on presence of config files, but in
reallity the "settings.intree" used to get tests path. Let's use it to
check for in-tree detection, which should better correspond to how this
value is being used.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

There are way too many warnings being generated by rpmlint, and
judging from other packages (just try a `rpmlint bash`) it seems
to be quite "normal".

Let's just tell the user to not be lenient with errors, but ignore
warnings.

Reference: https://trello.com/c/ihZUtKfc/1347-rpmlint-check-if-which-warnings-errors-can-be-fixedSigned-off-by: NCleber Rosa <crosa@redhat.com>

The local RPM build is now pretty insignificant, given that we have
RPM builds runnings out of Git, from the master branch on COPR.

Then, if we don't have local RPM files, it's not trivial to run
rpmlint on them.  Let's run it on the host we install the packages
into.

Reference: https://trello.com/c/ihZUtKfc/1347-rpmlint-check-if-which-warnings-errors-can-be-fixed
Reference: https://copr.fedorainfracloud.org/coprs/g/avocado/avocado-latest/Signed-off-by: NCleber Rosa <crosa@redhat.com>

We forgot to seek back to beginning of the log file when the size is
lower than the set max-log-size, which resulted in empty log output.
Signed-off-by: NLukáš Doktor <ldoktor@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: NMerlin Mathesius <mmathesi@redhat.com>

This test was always skipped due to different urllib handling in
python3. Let's use our download library to check the connectivity.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

We recently switched from simple to recursive discovery by default,
let's document how the discovery works now.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>