And drop the notion of a "multiplex capable" Avocado.  If the plugin is
installed, then we assume the dependencies have been fulfilled.  The plugin
presence is used then for the creation of the test suite, including the
plugin's own tests.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

This drops the notion of a "multiplex capable" Avocado.  If the plugin
is installed, then we assume the dependencies have been fulfilled.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

This change enables plugins tests to also be run in parallel, if
they exist.

Because the "selftests/{unit,functional,doc}" directories are valid
Python modules, that is, they contain `__init__.py` files and can
be imported by the unittest runner, no special handling is necessary.

But, since the complete path for directories which (can) hold plugins
tests are not valid Python modules (and shouldn't be), some special
handling of directories is necessary.

Because of that, it's not currently possible to run *all* of the
tests (from "core" selftests and from plugins) at the same time.
A future enhacement would be to better schedule processes that
can be loaded from the same path at the same time, but all of
them at once with global "test slices".
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: NLukáš Doktor <ldoktor@redhat.com>

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

* https://github.com/avocado-framework/avocado:
  Docs: update location for RPM repo location

We've got a new server, and the RPM repo is already there.  Let's use
that new server in the documentation.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

Some packages are specific to Python 2.7, such as mock which already
exists in the Python 3 standard library.

Also, commit fb98d386 removed the cyclical_deps check from Python 3
environments, because the requirements to run it are not available on
Python 3.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The goal of this change is to inspect how Python 3 compatibility plays
in the Travis-CI environment, while not blocking the supported Python
2 builds.

IIUC, when using "system_site_packages", we're stuck with the Python
version on Ubuntu Trusty (2.7.6 and 3.4.3).  That's the reason behind
the specific "3.4" version added there.

[1] https://docs.travis-ci.com/user/languages/python/#Travis-CI-Uses-Isolated-virtualenvsSigned-off-by: NCleber Rosa <crosa@redhat.com>

The 52lts branch is missing in the configuration file.  Let's add it.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

According to https://docs.travis-ci.com/user/languages/python, the
standard way of setting up that configuration is by using the
"system_site_packages" under "virtualenv".

It looks like the current syntax was a valid fix for a Travis-CI
issue that is now fixed, and it was being used in Avocado to support
the now dropped Python 2.6 version checks.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Recently we renamed the default varianter paths from "mux_path" to
"paths", which results in default "paths" being used when replaying
older results (including 52.0). Let's "upgrade" the results on-load.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

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

This builds a new sub package for this new optional plugin.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

It would be nice to have convinient way to upload results to centralized server.
This plugin superb resultsdb which support external log URL.
Proposed implementation is very simple, but it does it's job.
Signed-off-by: NDmitry Monakhov <dmonakhov@openvz.org>
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

This simple test, after the s390x version, checks the proper parsing
of and x86_64 /proc/cpuinfo.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The output in /proc/cpuinfo is different between s390x and other archs.
This is to support to parse the output on s390x.
Signed-off-by: NDan Zheng <dzheng@redhat.com>
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

Commit 51e46987 changed it from
`mux_debug` to `debug`, but using `debug` is actually dangerous if some
higher level plugin creates the same option, since it can make the
varianter plugin to use TreeNodeDebug instead of TreeNode.

Let's use a more exclusive naming, while still in line with the actual
implementation.
Signed-off-by: NAmador Pahim <apahim@redhat.com>

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

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

Fix the broken algorithm check in the hash_file function, it
should return `None` if the given hash algorithm is unknown.
Signed-off-by: NXu Han <xuhan@redhat.com>

Fix the dead loop caused by a wrong indentation in the
hash_file function.
Signed-off-by: NXu Han <xuhan@redhat.com>

The AvocadoParams, as a database of test parameters, has no need
to know information about the test.  In fact, the test id parameter
has never had any real use so far.

Let's remove it.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

A lot of the core Avocado code still refers to "mux" based terminology.
For instance, what was previously "mux_path" is now "parameter paths",
and what was previously "mux" is now really variants.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

While Avocado is running a test, it constantly requests the test to
send its state, making this operation a good candidate for some
optmization.

This change reduces the amount of attributes that are accessed and
serialized and returned as a test state to the barely minimum
observed, and at the same time, it documents it so that it will
hopefully start to serve as a documentation for the (private) protocol
that runners and tests use.

While at it, as a another smaller optmization, this list is defined at
the module level as a tuple (order doesn't matter here) and prevents
the list from being created on every single state construction.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

The forward reference from the parameter module towards the job log
is not ideal.  The parameter module should be used by test (and
related code) and should not know about the test and job internals.

But, there's the clear requirement and convenience of logging the
access to the parameters.  This makes the user of the parameter
instance define which logger should be used, and not the other
way around.

Logger names are used here, because, the communication between
runner and test (get_state()) currently includes test parameters,
and serializing a logger instance is not possible.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Commit be8a5129 changed the style of open() and then read()/write()
calls.  But, we mistakenly dropped the write method on this one.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Commit ed4dcd1d missed this one attribute that reference the old
ResultProxy attribute.  Let's remove it.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

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

* https://github.com/avocado-framework/avocado:
  Added required methods for ready-to-build source directory

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

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

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

We should only discover tests when we're able to run them. Otherwise we
should let the next available loader to attempt to discover them.
Signed-off-by: NAmador Pahim <apahim@redhat.com>