- Mar 03, 2016
-
-
-
Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D480
-
-
-
Don’t hard-code the traffic controlled interface to eth0, or the traffic-control-tcdemo test will fail on devices where eth0 is not the main network interface. Bug: https://bugs.apertis.org/show_bug.cgi?id=328 Differential Revision: https://phabricator.apertis.org/D431 Signed-off-by:
Philip Withnall <philip.withnall@collabora.co.uk> Reviewed-by:
Sjoerd Simons <sjoerd.simons@collabora.co.uk>
-
When cleaning up, ignore failure to kill tcdemo processes, caused by them already exiting or being killed earlier on. This fixes all test cases after the first one in the traffic-control-tcdemo unit test running. Bug: https://bugs.apertis.org/show_bug.cgi?id=328 Differential Revision: https://phabricator.apertis.org/D430 Signed-off-by:
Philip Withnall <philip.withnall@collabora.co.uk> Reviewed-by:
Sjoerd Simons <sjoerd.simons@collabora.com>
-
On an i.MX6, this test can take around 15 minutes under normal conditions, so any network delays are going to tip it over the edge and cause test failures due to timeouts. Double the timeout to 30 minutes to hopefully avoid this. Bug: https://bugs.apertis.org/show_bug.cgi?id=476 Signed-off-by:
Philip Withnall <philip.withnall@collabora.co.uk> Reviewed-by:
Luis Araujo <luis.araujo@collabora.com> Differential Revision: https://phabricator.apertis.org/D429
-
If the manual run-tcmmd-test fails because the user replies ‘no’ to any of the questions, it doesn’t clean up its tcmmd process. This causes an immediate failure of the test if it’s run again. Fix that by always cleaning up tcmmd. Differential Revision: https://phabricator.apertis.org/D427 Signed-off-by:
Philip Withnall <philip.withnall@collabora.co.uk>
-
There’s a race condition between killing a process and waiting for it to terminate; since this test is run with `set -e`, this can cause the test to fail if the process dies before `wait` is invoked. Fix that by ignoring failure from it. Bug: https://bugs.apertis.org/show_bug.cgi?id=327 Differential Revision: https://phabricator.apertis.org/D426 Signed-off-by:
Philip Withnall <philip.withnall@collabora.co.uk>
-
Differential Revision: https://phabricator.apertis.org/D411 Reviewed-by: xclaesse
-
Differential Revision: https://phabricator.apertis.org/D412 Reviewed-by: xclaesse
-
Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D414
-
Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D413
-
-
Differential Revision: https://phabricator.apertis.org/D410 Reviewed-by: xclaesse
-
`lava-test-case --shell` waits for LAVA_ACK from the dispatcher after each of these[1], but run-test-in-systemd does not. This is a point of incompatibility that could potentially cause or exacerbate issues in LAVA's "screen-scraping" on the serial console, such as <https://bugs.apertis.org/show_bug.cgi?id=396>. [1] https://git.linaro.org/lava/lava-dispatcher.git/blob/HEAD:/lava_dispatcher/lava_test_shell/lava-test-case Differential Revision: https://phabricator.apertis.org/D388 Reviewed-by: araujo
-
Summary: "?urn nie:dataSource tracker:extractor-data-source" is a property added by tracker-extract when it adds all metadata. Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D399
-
Summary: Tracker will soon index files from /home/share as well, just ignore them for now. Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D398
-
Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D397
-
Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D396
-
Summary: They are now assert_indexed and assert_not_indexer. This provides more useful backtrace if an error occures. Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D395
-
Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D394
-
Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D393
-
Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D392
-
Summary: This fix failing tests when destination directory does not already exists Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D384
-
Depends on D380, D382. Differential Revision: https://phabricator.apertis.org/D383 Reviewed-by: pwith
-
-
Summary: This is mostly a 1-1 translation. Tests are not smarter than what they were before, to the exception that they don't include sleep() of arbitrary duration everywhere. Reviewers: smcv, pwith Differential Revision: https://phabricator.apertis.org/D377
-
-
This is mostly a 1-1 translation. Tests are not smarter than what they were before, to the exception that they don't include sleep() of arbitrary duration everywhere.
-
-
Hopefully this will prevent the environment (LAVA? systemd?) from squashing leading spaces and making the tree hard to read. Differential Revision: https://phabricator.apertis.org/D376 Reviewed-by: pwith
-
thermald works with real hardware, and refuses to start otherwise. Bug: https://bugs.apertis.org/show_bug.cgi?id=485 Differential Revision: https://phabricator.apertis.org/D375 Reviewed-by: pwith
-
It works on real images, but not in LAVA. Bug: https://bugs.apertis.org/show_bug.cgi?id=483 Differential Revision: https://phabricator.apertis.org/D374 Reviewed-by: pwith
-
Differential Revision: https://phabricator.apertis.org/D373 Reviewed-by: pwith
-
This is good for usability, and might also prevent it from detecting inactivity and killing our tests, which upstream developer Neil Williams warned me might happen (although I haven't seen it on our installation). Differential Revision: https://phabricator.apertis.org/D372 Reviewed-by: pwith
-
-
Otherwise, leftover dbus-test-tool processes from some other test could be still on the bus, causing misleading results. Differential Revision: https://phabricator.apertis.org/D365 Reviewed-by: pwith
-
This syntax change was made when dbus-test-tool was merged upstream. Differential Revision: https://phabricator.apertis.org/D364 Reviewed-by: pwith
-
The way in which we reconfigure the dbus-daemons here has not worked since the bus setup moved to /usr/share/dbus-1/s*.conf in 1.9.18. It is no longer necessary anyway: root has access to Stats by default on the system bus, and users have access to Stats by default on their own session buses. Differential Revision: https://phabricator.apertis.org/D363 Reviewed-by: pwith
-