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

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

Fixing spell error introduced recently.
Signed-off-by: NAmador Pahim <apahim@redhat.com>

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

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

* https://github.com/avocado-framework/avocado:
  YAML Loader: move it into its own plugin and package

The SIGINT/SIGTERM handlers are currently broken. When we execute a
SIMPLE TEST, which ignores SIGINT/SIGTERM, the test process will not
finish on SIGINT/SIGTERM, becoming orphan.

Reproducer::

    $ cat /tmp/test.py
    #!/usr/bin/env python
    import signal
    import time

    if __name__ == "__main__":
        signal.signal(signal.SIGINT, signal.SIG_IGN)
        signal.signal(signal.SIGTERM, signal.SIG_IGN)
        signal.signal(signal.SIGQUIT, signal.SIG_IGN)
        while True:
            time.sleep(0.1)

    $ avocado run /tmp/test.py

Then play with `Ctrl+C` (once and twice), with `kill -2 <pid>` and
`kill <pid>`. Check if Avocado hangs or whether any process is left
behind.

Also, the only test we have for those cases is supposed to catch the
issue with SIGINT (test_interrupt.InterruptTest.test_badly_behaved), but
it's not really working.

This patch fixes the SIGINT/SIGTERM actions, as follows:

- When a SIGINT is sent to the main main process, we send the same SIGINT
  to all the children hope for the best. This reproduces the behaviour
  on Ctrl+C.
- On a second SIGINT (after the 2s ignore period), meaning process is
  still running, we send a SIGKILL to all the children.
- When a SIGTERM is sent to the main process, we now send a SIGKILL to
  all the children (instead of a SIGTERM), so any subprocess ignoring
  SIGTERM will be terminated anyway.

Reference: https://trello.com/c/5RPZPH4FSigned-off-by: NAmador Pahim <apahim@redhat.com>

The newly introduced YAML Loader plugin is currently hosted on the
same plugin (directory and module) as the varianter one.  While they
share a lot of code, they have different purposes.

Let's move the YAML loader into its own plugin, so that its
installation is also optional and the separation is clearer.
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>

* https://github.com/avocado-framework/avocado:
  python-avocado.spec: fix dependency on Python YAML packages
  Makefile: remove section about packages needed to build Avocado
  Makefile: use old chroot implementation for mock

This change is already present on the downstream (Fedora/EPEL)
packages, and for some reason, was missing here.  It causes
the installation to be attempted by pip, which should never
be done during RPM builds.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

All the Avocado deps are now either on Fedora or EPEL, so there's no
need for any extra setup, and consequently, for that documentation
section.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

mock has switched to using systemd-nspawn for its chroots.  It looks
like it's breaking the Avocado workflow for building packages.

Until we can better evaluate what's missing there, let's revert back
to the old implementation.
Signed-off-by: NCleber Rosa <crosa@redhat.com>

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

Use the new "b'' % foo" formats from PEP461.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

At least, selftests.unit.test_xunit passes.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Part of fixing selftests.unit.test_xunit under Python 3.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The precision of float printing has changed.  To provide
the same output for Python 2 and 3, explicitly specify
the number of significant digits.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

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

Writing to sys.stdout|stderr from inside test should be treated as any
other output in test and should be prefixed by [std* ] to emphasize it's
not comming directly from logging, but from direct output. This prefix
should  and is only be prefixed in job.log and not in stdout|stderr
files.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The LoggingFile is used to forward sys.stdout|stderr messages into
expected loggers. Using "paramiko" or "fabric" seems like a bug to me.
Let's only forward them into TEST_LOG, which also propagates to job.log.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

We are about to need to allow different log prefixes per logger in our
LoggingFile helper class. Let's allow it by using list of prefixes,
instead of a single prefix.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The argument logger actually expects multiple loggers in a list. Let's
reflect that to avoid possible confusion.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

shortcommit is supposed to contain the 8 first characters instead of 7.
Signed-off-by: NAmador Pahim <apahim@redhat.com>

Spell fixes and exception for the 53 release.
Signed-off-by: NAmador Pahim <apahim@redhat.com>

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

There are bunch of really short, one-time-used variables defined
globally, which is not really necessary, nor easy to follow. Let's just
define those one/two liners directly to where they are used.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

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

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

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

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

On python3 this naked open reports ResourceWarning. Let's properly close
it:

    avocado/utils/archive.py:191: ResourceWarning: unclosed file
    <_io.TextIOWrapper name='/tmp/avocado___main__rv2ihor0/tmp0drzyrsb/
    link_to_file2' mode='r' encoding='UTF-8'>
      src = open(dst, 'r').read()
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

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

* https://github.com/avocado-framework/avocado:
  Fix byte vs string file io for python 3

Reference: https://github.com/avocado-framework/avocado/issues/2024Signed-off-by: NJan Zerebecki <jan.foss@zerebecki.de>

Most of the LazyProperty variables in Test are actually always read on
"get_state" and other occasions so it does not make sense to set it
Lazily.

Also protect the remaining lazily initialized property from overriding
by using @property instead of our custom LazyProperty, which is not
really a property.

This commit requires some adjustments to nasty example tests which used
to override self.srcdir, which is not allowed (and restricted since this
commit).
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

The fail_reason is the instance of the actual exception. We do store it
only as string in case of old-type exception, but even new style
exceptions might be not available when sys.path is different causing
unpickle on runner part impossible.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

Use the new ability to report full mappings after the discovery and
allow using `test_reference_resolver_class` to specify the resolver
class intended for test discovery.

To allow sufficient flexibility also support
`test_reference_resolver_args` to override resolver arguments and
`test_reference_resolver_extra` to extend the resolver extra_params.

Note the overridden arguments apply only to the resolver, not to whole
Avocado so it's not possible to override wrappers or such just for a
single test execution. Anyway it is possible to set `vt_config` just for
that single loader, which is more-than useful.
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>

When no variants are provided Avocado sets the default variant, but
instead of list of mux paths it created just string "/run", which
processed into ["/", "r", "u", "n"]. Let's fix this by setting list.

Also the proper mux_path should be `/run/*` in order to include all
children nodes of the "/run".
Signed-off-by: NLukáš Doktor <ldoktor@redhat.com>