Recent changes to gio omit the atime instead of setting it to an out of range value.

Fixes #556, but note this doesn't fix the fact that gio omitted atime on a system where atime should have been available.

In GDM sessions (greeter, initial-setup), it does not make sense to
automatically logout. This can happen if the system wide default is
changed to default to the "logout" action.

Note that we already use the RUNNING_UNDER_GDM environment variable in
the keyboard plugin currently. So doing this is likely sane, even if we
probably want a more elegant strategy to detect whether we are in a
"login" session.

We already suppress logout actions in GDM (10aa1714, power: Avoid
automatic logout in GDM/greeter). However, while this prevents the
action, we may still warn.

Change it so that the corresponding timeouts will never be registered.
Leave the guard in gnome_session_logout but add a warning as we should
never be hitting that code path.

It is expected for USBGuard to not be installed on many systems. We
should not warn about that DBus error, as it does not indicate an issue.

Closes: #567

The new FDO templates do not need privileged runners anymore. Also,
those do not exist and the asan should be run on the asan runner
instead.

A lot of wireless input devices, such as Logitech mice and touchpads, or
Bluetooth LE input devices, will disconnect and reconnect frequently
from the computer when unused.

This causes a problem when the battery on the device is low because
a new notification will be generated each time the device reconnects, as
if it was the first time it connected.

Saving the last warning-level for every external peripheral ensures that
we only emit one low battery notification for each warning-level, when
the threshold is crossed, rather than at every reconnect.

The last warning-level is not stored, so a new session, or a reboot will
cause devices to warn again once.

Helps: #108

We need the database only quite irregularly (i.e. when the location
change). Just grab the required information from GWeather on the fly and
destroy it immediately after. Also, use the API to only grab the cities
of one country instead of filtering them later.

See also: !168

We need the database only quite irregularly (i.e. when the location
change). Just load the database on the fly when we need it, and discard
it afterwards to free up the memory again.

See also: !168

The environment variables were not correctly cleared. Fix this and also
make sure that we don't have the GNOME_SETUP_DISPLAY environment
variable leaking in from the outside.

Access to a /dev/foo device should never use buffered mode.

While debugging a gsd-rfkill issue I noticed in the g_debug output
that the rfkill-glib.c code now seems to be receiving bogus events.

Doing a strace I noticed some read(dev_rfkill_fd, buf, 8) calls,
even though we call:
  g_io_channel_read_chars(..., sizeof(struct rfkill_event, ...)

Which requests 9 bytes. The problem is the kernel expects us to
read 1 event per read() system-call and it will throw away
excess data. The idea is here that the rfkill_event struct can
be extended by adding new fields at the end and then userspace code
compiled against older kernel headers will still work since it
will only read the fields it knows in a single call and the
extra fields are thrown away.

Since the rfkill-glib.c code was using buffered-io and asking
g_io_channel_read_chars for 9 bytes when compiled against recent
kernel headers, what would happen is that 2 events would be consumed
in 2 read(fd, buf, 8) syscalls and then the first byte of the
second event read would be appended to the previous event and
the remaining 7 bytes would be used as the first 7 bytes for the
next event (and eventually completed with the first 2 bytes of
the next event, etc.). Leading to completely bogus events.

Enabling unbuffered mode fixes this.

Note this is a relatively new problem, caused by the kernel
recently extending the rfkill_event struct with an extra byte-field:
"rfkill: add a reason to the HW rfkill state"
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=14486c82612a177cb910980c70ba900827ca0894

Before that kernel change the rfkill_event struct was 8 bytes large
which allowed us to get away with using buffered io here.