The regexp is not bound to either end of the process name, so despite
the fact that the test script changed its effective process name to
‘ofonod_’, the regexp ‘ofonod’ still matches that, and hence the script
was killing itself, which was causing the systemd unit it was running as
to fail, and hence the overall test to fail.

Tighten the pkill regexp to match at the end of the process name to
avoid this.

Bug-Apertis: https://bugs.apertis.org/show_bug.cgi?id=681


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

Differential Revision: https://phabricator.apertis.org/D2284

If an AppArmor malicious test is run as a systemd system job (using
`run-test-in-systemd --system`), $HOME will explicitly not be set, which
results in the program trying to read (null)/.bash_history, rather than
the expected /home/user/.bash_history.

Fix that by hard-coding it to use /home/user/.bash_history if $HOME is
not set. If the username changes in future, the tests should start
failing, which will allow us to update it again.

Bug-Apertis: https://bugs.apertis.org/show_bug.cgi?id=681


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

Differential Revision: https://phabricator.apertis.org/D2283

This matches the command used in ofono.service (minus the --nodetach,
because we want the process to detach so we can do other things in the
test). This eliminates some spurious AppArmor failures caused by the RIL
code, and allows ofono to work (the RIL plugin does not work on the
i.MX6).

Bug-Apertis: https://bugs.apertis.org/show_bug.cgi?id=681


Signed-off-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2278

As explained in the new comment in the Tumbler mixin, it is possible for
the Error and Ready signals from tumblerd to race if one thumbnailer
fails fast while another succeeds (for the same URI) slowly. This
happened for me with the cover and gst thumbnailers on
big_buck_bunny_smaller.ogv — the cover thumbnailer failed fast because
the search terms ‘big buck bunny smaller’ return nothing on omdbapi.com,
while the gst thumbnailer eventually succeeded by grabbing a frame from
the video.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D1299

We can't to anything with bluetooth until its enabled via connman
(otherwise everything else will blocked/soft rfkilled).

Enable it explicitely in the agent so the user doesn't have to first use
connmanctl.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2209

object_set_properties sets the properties one by one, but the bluez
interface doesn't allow setting Discoverable or Pairable unless the
adaptor is powered. So before setting those, power it on first.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2208

The simple-agent command line arguments were a bit weird, it can just
act as an agent by giving no arugments or the adaptor to use. However to
pair a specific device a adaptor was mandatory (which is pointless on
most systems with just one adaptor). And for extra fun adding a third
argument which (could be whatever).

Sanitize things by making the device an optional argument rather then a
positional one. Drop the recgonition of the third positional argument
(bluetoothctl can be used instead to remove pairings).

On top of that switch to argparse (The recommended argument parser in
python 3), fix outgoing pairings (with bluez5 you can only trigger a
pairing with devices found during scanning it seems) and add a --yes
argument to accept all incoming connection making it more convenient to
use in testing.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2207

This variable can be used to specify a base url for the
images url link, which will allow to set http credentials
for those profile requiring so.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Set the main adaptor to be discoverable when starting the simple agent
for incoming pairing requests.

Also set the agent to be the default to prevent an already running
system agent interfering with the test.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2203

Sometimes just setting the adaptor to be pairable and/or discoverable
as a one-shot action without change notifications is enough. Support
this.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2202

Add a symlink in the python 3 system path to maek apertis_tests_lib
available to the system.

Note that this is a somewhat ugly solution. To solve it properly the
build system needs to be a bit revamped to install the python libraries
(and really binaries) in the right spots in the system.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2201

Philip Withnall authored 9 years ago and Sjoerd Simons committed 9 years ago

This is effectively a complete rewrite of the tests. Some of the common
functionality has been split out to apertis_tests_lib. The external
interfaces and behaviour of the tests should have remained unchanged.

The tests needed to change considerably in structure because of the
difference between how BlueZ 4 and BlueZ 5 handle device objects. BlueZ
4 only creates a device object if an adapter is paired with that device.
BlueZ 5 creates it as soon as a device is discovered.

Apart from that, the main differences between the two APIs are in the
shift to standard D-Bus interfaces (Properties and ObjectManager) in
BlueZ 5.

While rewriting the tests, a lot of multi-second sleep() calls have been
removed in favour of waiting for the relevant completion signals from
the D-Bus API. This means some potentially unbounded waits if the
Bluetooth daemon misbehaves or ends up in a state the unit test does not
expect. But otherwise it means faster test execution and no failures due
to races between the sleep() and the Bluetooth daemon completing an
operation.

[ Small updates for review  comments ]

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D971

Reviewed-by: Mathieu Duponchelle <mathieu.duponchelle@opencreed.com>
Signed-off-by: Emilio Pozuelo Monfort <emilio.pozuelo@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2194

Of the schedulers we compile, only cfq supports the block I/O weighting
feature that is exercised by test_blkio_weights_random_read and
test_blkio_weights_sequential_read.

Related to <https://bugs.apertis.org/show_bug.cgi?id=635

>

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/D2165

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/D2164

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/D2163

Leave a symlink behind for compatibility.

The test still has a dependency on chaiwala-tests for now, so that it
can be run against older versions of apertis-tests.

While I'm editing dependencies, add python, since the test runs it
explicitly.

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/D2162

This makes it much easier to try new versions.

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/D2161

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>

Simon McVittie authored 9 years ago and Sjoerd Simons committed 9 years ago

These are ostensibly JSON, but what they actually are is Jinja
templates whose *output* is JSON. In particular, they can contain
{# comments in this syntax #}, causing syntax-check to fail.

lqa will check that their output is valid JSON anyway, so I think
the ability to have comments is worth more than the syntax check.

Reviewed-by: Luis Araujo <luis.araujo@collabora.co.uk>
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2156

Mathieu Duponchelle authored 9 years ago and Sjoerd Simons committed 9 years ago

We no longer ship chaiwala-btrfs.

Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2158

Mathieu Duponchelle authored 9 years ago and Sjoerd Simons committed 9 years ago

mc-tool (among other things), wasn't installed on the LAVA instance.

This package is provided by telepathy-mission-control-5 , which
is correctly listed here:

https://wiki.apertis.org/QA/Test_Cases/apparmor-folks



As such, this commit adds it, and other dependencies listed on
the wiki, but absent in the yaml file.

Reviewed-by: Luis Araujo <luis.araujo@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2132

Emilio Pozuelo Monfort authored 9 years ago and Sjoerd Simons committed 9 years ago

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Signed-off-by: Emilio Pozuelo Monfort <emilio.pozuelo@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2131

Simon McVittie authored 9 years ago and Sjoerd Simons committed 9 years ago

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/D2117

Simon McVittie authored 9 years ago and Sjoerd Simons committed 9 years ago

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/D2116

Mathieu Duponchelle authored 9 years ago and Sjoerd Simons committed 9 years ago

And check the id could actually be looked up.

Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Differential Revision: https://phabricator.apertis.org/D2115

Luis Araujo authored 9 years ago and Sjoerd Simons committed 9 years ago

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Luis Araujo authored 9 years ago and Sjoerd Simons committed 9 years ago

This commit removes the {{arch}} variable from the image
names in the job template files.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Luis Araujo authored 9 years ago and Sjoerd Simons committed 9 years ago

This commit removes the changes of an optional board
and image type for the minimal template.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Luis Araujo authored 9 years ago and Sjoerd Simons committed 9 years ago

Added the following attributes to the job templates: image.date,
image.version, image.board and image.type.

These attributes can be used to filter LAVA tests results based
on their values.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Luis Araujo authored 9 years ago and Sjoerd Simons committed 9 years ago

This commit reverts a previous change about testing for an optional
image type and board to work around the target wayland image url not
matching the image name. The issue is now properly fixed in the image
generation scripts.

Signed-off-by: Luis Araujo <luis.araujo@collabora.co.uk>

Luis Araujo authored 9 years ago and Sjoerd Simons committed 9 years ago

This commit changes the template-dir variable to the relative path
apertis-tests/templates/ , so the templates files can be found from
the same directory this repository is checked out.

Luis Araujo authored 9 years ago and Sjoerd Simons committed 9 years ago

This commit updates all the test templates and profiles for the new
images urls. It includes the following changes:

- Add two new optional variables opt_image_type/opt_board and check
if they are defined before using them as a work-around for the amd64
wayland image url.
- Move some common profiles variables to the parent profile.
- Include the new minnowboard-max-minimal profile.