This way commands that do use the paginator will correctly
redirect error messages to stderr:

scripts/avocado list sbrubles > /dev/null
Cannot access 'sbrubles': File not found
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Make sure the paginator is cleaned up after displaying
an error.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Following the [ObjectSingular][List] convention.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Let's standardize the class names for plugins that
list things in [ObjectSingular][List].
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Following the [object_singular]_[list] convention.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Let's standardize the file names for plugins that
list things in [object_singular]_[list].
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

More checks to make sure we are not outputting errors to
the wrong loggers.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Create an additional logging handler for the avocado application,
that outputs to stderr, and two filters, that filter messages
per logging level. This way, normal information (app output)
goes to stdout and error messages go to stderr. The unittests
were updated accordingly.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

fiotest: fix async engine name and update to latest upstream version

This simply updates fiotest to use the latest release of fio.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

It looks like the name of the loadable engine on fio itself has changed
from simpler "aio" to either a Linux or POSIX version.

Based on my tests on build attempts, it looks like POSIX is more dependant
when it comes to building out of the box. Let's default on that version.

This fixes the following error when running fio:

 fio: engine aio not loadable
 fio: failed to load engine aio
 ...
 fio: file:ioengines.c:99, func=dlopen, error=aio: cannot open shared object file: No such file or directory

That resulted in a test failure.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Documentation: simple fixes/improvements to the Get Started Guide

doublefree test: fix the temporary directories being used

This test is using the location of the source file as the build
location, instead of the `srcdir` provided by the Test class.

Also, since `srcdir` is a temporary directory, I believe we should
not have to manually clean it up.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Even though internally we have some more descendants of avocado.test.Test, to
a user, AFAICT, only instrumented and simple tests are available.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

It's obviously not true that Avocado only allows to run tests on "your laptop".
Also let's keep the running of tests and the outcome all in the present tense.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

Add a new config section 'runner.output', with a key 'enable' to
enable/disable colored output in the avocado ui. With this, users
troubled with terminal color schemas may turn off colors to make
things more readable.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Make lines fit into 80 columns. No functional changes here.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

HTML result fixes

Modify the VM test info in order to correct paths of
test directories. This will fix a long standing bug with
regards to broken HTML report links.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Add the fields 'logdir', 'logfile' and 'fail_reason'.
Those fields will be necessary to expand the HTML
reports to correctly exhibit failure information
from remotely run tests.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Looking in retrospect, we used quite cryptic names for
the logfile and logdir template dict keys. Let's fix that
so that the code is more understandable.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Make sure all brower terminal output is suppressed. In order
to do that, eschew the use of the python module web browser
and use subprocess, executing the utility xdg-open in a subprocess
with stdout/err redirected to /dev/null.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

* Make code lines to fit under 80 columns
* Slight reorganization of commits

No functional changes were made to the code.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Drop features HTML Report and Multiplexer gracefully.

Argument parsing job result [V4]

The current Job object initialization is doing many setups at once,
 for example, it creates a new job results dir with an "id" file inside.
 It handles the sysinfo start, update the latest
 link, etc. Things that can be done later.

This changes allows split the initialization in subparts:
* _setup_job_results() - creates a new job results directory with an id inside.
* _update_latest_link() - update the "latest (symbolic) link" to the job results.
* _start_sysinfo() - start the system information module.
* _remove_job_results() - remove the current job results from the directory.

So when running tests that are not found, say `avocado run quuuux`
then the job results will be removed and the link for the previous
result (latest) is preserved. Everything else should work as it was before.
Signed-off-by: NRudá Moura <rmoura@redhat.com>

Instead of delaying the parsing of the test list to code that handles
the jobs, let's handle that early, and in the usual Python way, simply
by relying on argparse.

For other, possibly more complex command line requirements, custom code
can be written, but it should still be placed at the argparse level.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Some obvious, but AFAICT not yet existing tests, that check if when
using bad command line arguments, avocado exists with a proper status
code.

Also, a more complex command line behavior test that checks if avocado
errors out early, before it creates the job result directory.
Signed-off-by: NCleber Rosa <crosa@redhat.com>
Signed-off-by: NRudá Moura <rmoura@redhat.com>

Documentation: Add section about installing from git repository.

Test for feature MULTIPLEX_CAPABLE to graful disable the multiplexer
if PyYAML is not installed.
Signed-off-by: NRudá Moura <rmoura@redhat.com>

Test for features HTML_REPORT_CAPABLE and HTML_REPORT_SUPPORT to
graceful disable the htmlreport if pystache is not installed.
Signed-off-by: NRudá Moura <rmoura@redhat.com>

Add a new section inside the "Getting started guide", about installing Avocado
from the git repository with confidence that the dependences will be installed.
Signed-off-by: NRudá Moura <rmoura@redhat.com>

Cleanups sprint 21 - stage 1

docs: Add workshop mindmap to Introduction docs

Add link to the DevConf mindmap to introduction documentation.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>