This reverts commit 71bafe56.

The original commit wasn't well thought through.

Signed-off-by: Andrew Shadura <andrew.shadura@collabora.co.uk>
Reviewed-by: Luis Araujo <luis.araujo@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D5272

Purely rely on systemctl to stop and start the ofono service.
This is much cleaner as tests don't have to bother killing the process
manually any more.

The LD_PRELOAD is now handled by generating a temporary systemd service
config file containing the 'Environment' key.

Apertis: https://phabricator.apertis.org/T1296



Signed-off-by: Guillaume Desmottes <guillaume.desmottes@collabora.co.uk>
Reviewed-by: Emanuele Aina <emanuele.aina@collabora.com>
Differential Revision: https://phabricator.apertis.org/D5310

Add to option to enable more debugging infos

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.com>
Differential Revision: https://phabricator.apertis.org/D5054

In test_opp_client, the exception is not handled. When this happens,
the test stalls. Handle the exception by printing it, failing the test
and processing the next task.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.com>
Differential Revision: https://phabricator.apertis.org/D5053

ubt from git is not executable. It must be chmoded frequently.
Let's change that!

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.com>
Differential Revision: https://phabricator.apertis.org/D5052

Not doing so breaks binary rebuilds

Apertis: T3162

Signed-off-by: Andrew Shadura <andrew.shadura@collabora.co.uk>

The folks packages were dropped from the images by default and moved
into another component, so they all were added as dependencies for most
of the automated tests, nevertheless not all of them should be
explicitly required as deps and could cause problems later as explained
by Simon at https://phabricator.apertis.org/D1307#25238

.

Apertis: T789

Signed-off-by: Andrew Shadura <andrew.shadura@collabora.co.uk>
Reviewed-by: Héctor Orón Martínez <hector.oron@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4919

This test is deprecated as a new one needs to be developed for
wayland images and it is not valid for SDK images.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4946

The agent in bluez test has the possibility to accept everything.
It saves the tester from typing yes to continue the test everytime a
bluetooth event occurs. Thus reducing error and test duration.
This feature is already in use in bluez-hfp test.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.com>
Differential Revision: https://phabricator.apertis.org/D4907

MAP profile test was using a deprecated API dating back from 2012.
Update it according to version of bluez present in Apertis.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.com>
Differential Revision: https://phabricator.apertis.org/D4906

Before testing, install the necessary media files. This allows the
automated tracker tests to work in LAVA.

Apertis: https://phabricator.apertis.org/T2051
Differential Revision: https://phabricator.apertis.org/D4904


Reviewed-by: Luis Araujo <luis.araujo@collabora.co.uk>
Signed-off-by: Marc Ordinas i Llopis <marcoil@collabora.co.uk>

Bluez-hfp test creates an agent but it is not made default.
Because of this, when running in the Apertis SDK, an incoming connection
from bluetooth gets authorized by blueman agent which requires some
actions with the mouse to click the right button.
This is not appropriate as the agent from bluez-hfp is automated to
reduce the number of interactions from the tester.

Apertis: https://phabricator.apertis.org/T2880



Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.com>
Differential Revision: https://phabricator.apertis.org/D4898

This commit groups most of the tests in a single boot session for
the apparmor, common and clutter LAVA jobs.

This greatly improves the time it takes to run these jobs, since
the image won't reboot for every test anymore.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4767

Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>

Signed-off-by: Andrew Shadura <andrew.shadura@collabora.co.uk>

When the profile list are parsed process_next is called. It is also
caled when the pairing is terminated. When only one profile was
discovered, the only task in process_q would be popped at first
process_next() call. The second call would find an empty process_q
and quit the mainloop without executing the test.

This patch arrange the pairing and non pairing tests so they are
symmetrical. The call to tester.get_device is synchronous.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4561

In dun and pan tests, a file is retrieve from http to verify connection,
it is saved at the location defined in variable path. But path do not
contain a file path. Instead it's a dbus object path in /net/connman
namespace. As such, it is not suitable for using as a filename and any
tentative to open this file will end with an "File Not Found" error.

Instead, create a temporary filename that will be automatically deleted.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4447

The result of an urlrequest.read is a binary stream of data.
The input for file.write() is a string.
Make sure the write call is fed with the proper type of data.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4446

Reviewed-by: Luis Araujo <luis.araujo@collabora.co.uk>
Signed-off-by: Héctor Orón Martínez <hector.oron@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4435

Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>

ofono runs as root, so having the "malicious" LD_PRELOAD module read
/home/user/.bash_history requires it to exercise CAP_DAC_OVERRIDE,
leading to an AppArmor denial that doesn't match what we expect. Try
to read /etc/shadow instead, and update the expected denials
accordingly.

With that change, there's no real need for the "malicious" code to use
GLib, so use stdio instead.

Apertis: https://phabricator.apertis.org/T1848



Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.co.uk>
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4388

The ofonod wrapper script uses pkill 'ofonod$' to terminate the
malicious ofonod process. This call has the bad idea of killing every
process which name ends in ofonod. This include "test-ofonod" which
is the master script that started ofonod wrapper.

Remove pkill since we have the pid of the process to kill.
Also it's not needed to change the process name.

Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.co.uk>
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4386

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4383

Auto-discovering ${top directory}/resources instead of
/usr/share/chaiwala-tests/resources means we don't have to depend
on the legacy chaiwala-tests package.

We still need to hard-code /usr/lib/apertis-tests to load the
LD_PRELOAD hack that we use to simulate a compromised pulseaudio
process, but we can reduce the test/fix cycle for the rest by
running it directly from a git checkout.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4382

This package is no longer required as the test is deprecated.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4345

This test is deprecated as a new test need to be developed to be run
from Canterbury (https://phabricator.apertis.org/T2119

).

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4344

Signed-off-by: Héctor Orón Martínez <hector.oron@collabora.co.uk>

Hopefully this should help debugging T2504 by telling us what's going on
in the http server.

Need to run the test with DEBUG=2 to have these extra debug info.

Signed-off-by: Guillaume Desmottes <guillaume.desmottes@collabora.co.uk>
Reviewed-by: Justin Kim <justin.kim@collabora.com>
Differential Revision: https://phabricator.apertis.org/D4354

We reported in https://bugzilla.gnome.org/show_bug.cgi?id=755181


a "weird" behaviour that has now been fixed upstream.

Follow the recommendation given in the comment to remove the
calls to get_parent(), and remove the comment itself.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>
Differential Revision: https://phabricator.apertis.org/D4160

Reviewed-by: Héctor Orón Martínez <hector.oron@collabora.co.uk>
Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4350

Signed-off-by: Héctor Orón Martínez <hector.oron@collabora.co.uk>

This syntax is wrong, it generates the following error:
File "/usr/lib/python3/dist-packages/apertis_tests_lib/bluez.py", line
  557, in __show_menu
      name = str(i['org.bluez.Device1'].get(['Name'], address))
      TypeError: unhashable type: 'list'

Signed-off-by: Frédéric Dalleau <frederic.dalleau@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4340

Reviewed-by: Héctor Orón Martínez <hector.oron@collabora.co.uk>
Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>
Differential Revision: https://phabricator.apertis.org/D4310

In a fresh VM, following the instructions in
https://wiki.apertis.org/QA/Test_Cases/bluez-phone


means that the device object has not yet been created, fiddling
with d-feet showed me that `StartDiscovery` was necessary for this
to happen, which is called by scan_for_device.

The result was an enigmatic D-Bus error message:

Create device failed org.freedesktop.DBus.Error.UnknownObject: Method
"Pair" with signature "" on interface "org.bluez.Device1" doesn't exist

which is solved with this patch.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>
Differential Revision: https://phabricator.apertis.org/D4264

Previously, this test didn't provide an obvious indication that it
had all succeeded, and assumed that it was run as root with $(pwd) in
/usr/lib/apertis-tests or a git clone of apertis-tests.

To use this as a manual test:

    Prerequisites:
        - apertis-tests installed (sudo apt install apertis-tests)
    Steps:
        sudo /usr/lib/apertis-tests/common/sanity-check
    Successful result:
        Output ends with "# Sanity check successful"
    Failing result:
        Output ends with "Failed, see above for details" or ends with
        neither of the expected messages
    Diagnosing failures:
        Look for "TEST_RESULT:fail:" in the output (it may appear more
        than once), and look for warnings/errors shortly before each
        occurrence of that message

Reviewed-by: Luis Araujo <luis.araujo@collabora.co.uk>
Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4284

os-release pretended we were Ubuntu for a while, to make packages
with different build steps for Debian and Ubuntu follow Ubuntu-like
code paths. This has now been fixed.

Reviewed-by: Simon McVittie <smcv@collabora.co.uk>
Signed-off-by: Héctor Orón Martínez <hector.oron@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4272

As the SDK image is supposed to run in a virtual machine,
there is no need to run its tests in a board.

This commit changes the LAVA profile configuration so the
SDK AMD64 image tests will be executed using the i386 device
(virtual machine).

Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4175

This test is being deprecated and will be replaced by a new
webkit2gtk layout test.

Reviewed-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D4149

Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>