Introduce a TestAbortError, used for when an avocado test
aborts.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

If the test process is running, the trobber will give feedback about that
to the user.

This requires the process triggering and timeout handling to be completely
different than the previous implementation.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

It's now a known fact that long test runs are can lead to doubts with
regards to the actual test run status (is it frozen or really busy?).

To give users a better experience, the OutputManager now implements
a throbber for progress feedback.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

In some circurnstances, the journal may be accessed simultaneously by
multiple code paths. The database consistency should be enforced in any
case, and only initialized and populated when appropriate.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

Since other more standard plugins have moved to test state in a previous commit
the journal plugin should also follow that.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

This commit implements two very different things, but still very
much connected:

1) Separate test loading to the process that will be actually
running the test. This leads to a better separation of duties of
the process interacting with the user and the process that is
actually doing the (usually) heavy work.

2) Remove the pickling of the test instance and sending it back
and forth.

As an added bonus, it was noted during development, that both
processes load all necessary Python modules (including the test
module) when unmarshalling the pickled instance, resulting in
unecessary work. Actual data on performance improvements were not
collected though.

It has one change in behaviour (regression) though. Test timeout
now comes only from the test parameters. The reasoning is that
since there's now a better separation of roles and the fact that
the first process is not loading the test instance, it can not
look at the timeout defined there.

IMHO this is a sign of good design, since other types of test
(think of shell scripts) can not communicate their intended timeout.
Maybe a 'test.cfg' or something like that should be the solution to
that.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

avocado.test: export environment variables to tests.

Export environment variables to the running test.
Really effective when using drop-in tests, that can make use of shell
variables "AVOCADO_*".

Example:

Export variable AVOCADO_VERSION with the current version of Avocado,
then the drop-in test can use $AVOCADO_VERSION within, and so on.
Signed-off-by: NRuda Moura <rmoura@redhat.com>

Fix vm plugin

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

So the vm plugin can correctly determine the logs dir.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

We missed that on the whiteboard work.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Signed-off-by: NRuda Moura <rmoura@redhat.com>

multiplexer plugin: simple usability fix

Flat tests

Packaging fixes

Make avocado tests to refer to their test ids (paths),
relative to the directory you are in. For example,
running:

avocado run tests/sleeptest.py

will give you

DEBUG LOG: /home/lmr/avocado/logs/run-2014-07-31-12.54.01/debug.log
TOTAL TESTS: 1
(1/1) tests/synctest.py.1: PASS (1.93 s)

Test log and data directories were changed accordingly. Now,
if your test is synctest.py, your data dir will be called
synctest.py.data.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Since the multiplex file is a positional argument, we can not say
it's a required parameter, but we can and should check that at
least one was provided.
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>

Add tests to verify if passing the full test path
works, rather than the test ID, and if an empty
test ID/alias triggers the correct error code.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Give the user a nice message rather than a crash report.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

So we can let the test runner to notify the user of
incorrect options passed in.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

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

Now tests don't need their own directories.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

If tests do want to provide a datadir, it must be in
the same directory as the test source file. We probably
need to review this limitation and provide other
interesting alternatives.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

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

We want avocado tests to not require specific directories,
the user should be able to arrange them in the dir structure
they see fit. Also, if a test requires data, we suppose their
data dir is located in the same directory as the source file,
with the name:

[modulename].data

Re arrange the test dir to reflect this new structure.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Make it possible to load avocado tests from any path,
rather than enforcing the rule '1 test has 1 dir'.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

We're planning to move away from the rule

1 test == 1 directory

In order to do that, we need to be able to tell
the difference between a dropin test (this can
be written in any language as long as it is an
executable file) and an avocado test (a python
file that contains a test class derived from
the base avocado test file).

Introduce in the path library a path inspector,
that has code to figure this out.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Avocado utils work

The timeout there is unnecessary and introduces races.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

It saves quite a lot of code, since this is such a common
operation in avocado.

I'd like to have the reviewer to give a good thought of
whether this is a good solution or not (I think it is,
although API names and such could be better thought off).
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Add a module for path related operations in avocado.utils.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

It is nice to be able to refer to modules using the
dot notation. This fixes some missing imports in the
avocado.utils module, as well as it removes unused
ones.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Having 'misc' libraries tends to encourage people to
dump random functions without much thought as to
their location. Let's rename 'misc' to 'io'.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>

Also, reorder write_file and write_one_line.
Signed-off-by: NLucas Meneghel Rodrigues <lmr@redhat.com>