datetime: Fix checking when timesyncd will run on a system
Fixes #2158 (closed).
When 'systemctl show' is used see the ConditionResult property, the value is correct only when the unit has been started. When the unit is not running but can run, ConditionResult has a value of 'no' leading to an incorrect result.
This in turn leads to _is_time_managed() returning incorrect value once the service has been stopped. FreedomBox would have noted that daemon can be enabled/disabled during startup while during attempts to enable it the action script will think that service can't be enabled/disabled.
Fix this by using a better approach to detect when the service can run. Newer versions of systemd (likely >=250) have the ability to run 'systemd-analzye condition --unit=systemd-timesyncd.service' which have been ideal to detect this. However, --unit option is not available in older versions. Use systemd-virt-detect (part of systemd package) to detect for containers instead.
Tests:
-
Boot the machine and run datetime functional tests
-
User interface should not show enable/disable button for the app in container but show in VM.
-
Running first setup (after removing /var/lib/plinth/plinth.sqlite3) should work on container and VM.
-
Run above tests on a container and on a VM
Signed-off-by: Sunil Mohan Adapa sunil@medhas.org