The "Result" plugin interface was defined a while ago, and we now have
three implementations (JSON, XUnit, HTML).

Let's include them in the plugin listing, that is, when a user calls
`avocado plugins`.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

optional_plugins.html: Minor fixes

The method `logdir` is not used anywhere in the template, nor in the
plugin itself. I tried to find any occurrences and also setting a
breakpoint without any use whatsoever. Let's remove it.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

No actual changes, just style fixes.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

* https://github.com/avocado-framework/avocado:
  HTML Result: host it in a directory outside the avocado package
  HTML Result: move rendering code from core to the plugin itself
  HTML Result: add missing fields to formatted version of test info
  HTML result: use names closer to underlying result instance
  Makefile: smokecheck also needs clean/develop
  Makefile: fix test name used on smokecheck

avocado.core.sysinfo: Force LC_ALL for sysinfo collection [v2]

This commit allows setting custom LC_ALL for sysinfo collection to
simplify `avocado diff` comparison between different machines. One can
disable this by setting no/empty value.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

This step makes the HTML result a proper Python plugin, by living
outside the Avocado package namespace.

The RPM package is also adjusted: it now builds ands install both
Avocado and the HTML result plugin while keeping the two separate
installable packages.

Developers can use the usual `make link` and `make clean` to enable
and disable all plugins (including this one).
Signed-off-by: NCleber Rosa <crosa@redhat.com>

This change removes the knowledge, and dependency, that the current
Avocado core has of the HTML rendering code.

The only user visible change is that the UI message about the
generated HTML file now appears one line after it used to appear.  The
reason is not because of the HTML plugin itself, but because the
results are now generated outside the context of the job, and that is,
chronologically a bit later than it used to be.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Commit 84a8acc0 fixed a problem with 625eefe0, by reverting its
original proposal and copying everything from the inner test result
information to a new dict that is fed to the report template.

It turns out that only two entries were missing (test ID and status)
so we don't really need to copy the whole thing.

Again, this is basically a revert of 84a8acc0 that adds those two
missing fields.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

There's no need to have different names on the ReportModel class from
the Result instance, which is really the source of information.

This change is a step towards using better known names and possibly
making the ReportModel a child of Result itself.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

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

Create a git workflow section based on the devel list
discussion.
Signed-off-by: NAmador Pahim <apahim@redhat.com>

Plugins: add support for optional plugins dir and related Makefile targets

This reverts commit 6cb88049.

The rationale is that we should really use LC_ALL instead of LANG.

The `run` does not accept arguments anymore but stores them in `self`.
Let's get rid of the outdated docstring.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Currently only tree produced from yaml file is filtered. This patch
changes the logic to first produce the full tree and then apply filters.
The difference is that this way one can filter also nodes defined via
`avocado_devault_params` like avocado-vt params or `--mux-inject` ones.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The `MULTIPLEX_CAPABLE` only influences the `yaml` parser. All other
functions are available without it so let's allow users to use them.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The option `--multiplex-files` has been deprecated for several releases,
let's finally remove it.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The multiplexer is our way to define test params. Let's mention the
purpose, rather than our cryptical name in the parser's argument group.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

This commit allows setting custom LANG for sysinfo collection to
simplify `avocado diff` comparison between different machines. One can
disable this by setting no/empty value.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Currently we fail silently when sysinfo configuration files
are not present. This patch adds a log.debug message so the user
will have a chance to understand why sysinfo is not being
collected.

Reference: https://trello.com/c/5rvvdH8lSigned-off-by: NAmador Pahim <apahim@redhat.com>

The current method of IP probe is unreliable, as the IP can change between
reboots and we will still use the previous IP. This patch changes the IP
probe method to use the libvirt's domifaddr feature. We try to use the
AGENT method, which relies on a qemu-guest-agent query.

Reference: https://trello.com/c/y9nKJwfnSigned-off-by: NAmador Pahim <apahim@redhat.com>

Commit 88cd8261 moved the plugin interfaces to `avocado/core` under the
name `plugin_interfaces.py`.  The example plugins were not updated, so
let's fix it now.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The `link` and `clean` targets, very useful under development
environments, have supported directories containing `Makefiles`.

It turns out that for some plugins it may be enough (and more
Pythonic) to just have `setup.py` files, so let's support that.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The goal of the `optional_plugins` directory is to hold all plugins
that ship with Avocado, but are not considered core functionality.
Plugins that are considered to be core still live under
`avocado/plugins`.

This also adds all directories to be found under `optional_plugins` to
the list of candidate plugins when running `make clean` or `make
link`.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

avocado/core/settings.py: allow settings to use the home dir notation [v4]

VIRTUAL_ENV variable might not be always present. That's the
case for ReadTheDocs. On the other hand, sys.real_prefix seems to
be a more reliable way to detect we are under a virtualenv.
Signed-off-by: NAmador Pahim <apahim@redhat.com>

Some settings refer to paths on the filesystem, and the use may want
to set them as `~/something`.  Let's use the type `path` on known
configuration keys.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

It's common to use the tilde notation (`~/foo`) to refer to locations
inside the user directory.  Let's introduce a new settings key type,
called `path`, that will automatically expand the tilde notation
if one is found.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

* https://github.com/avocado-framework/avocado:
  Makefile: limit clean/link to possible plugin directories
  selftests/unit/test_utils_partition.py: also treat raised OSError
  Makefile: add explicit config to all mock commands
  Makefile: RPM repos are not hosted on COPR anymore
  Makefile: move Debian package related rules to contrib
  Makefile: add target that shows general configuration
  Makefile: silence possible setuptools warnings

The current wildcard expansion will look list every file (no matter
what type) as a plugin candidate.  It may be just my setup, but I have
a few files on the same common directory that shouldn't be treated as
plugin dirs.

While at it, also limit to directories that actually have a Makefile.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

When skipping tests due to the lack of sudo permissions, OSError
can be raised when the asked binary does not exist and can not be
executed.

In an ideal world, only one "skip" decorator would have to run, and
this extra treatment wouldn't be necessary since there's already a
`missing_binary` being used.  Unfortunately, that's not how they work.
Signed-off-by: NCleber Rosa <crosa@redhat.com>