The new wpa_supplicant D-Bus interface only passes back the 'level'
of the scanned BSS, which with nl80211 drivers is almost always dBm,
which NM handled fine. But WEXT-based drivers (ipw2x00, other older
ones, and some vendor drivers) use a mix of values for the 'level'
parameter, including the old WEXT 8-bit signed-value-in-unsigned-int
scheme. Handle that.
Alternatively, we could have the supplicant expose the 'flags' value
from its internal BSS list over the bus.
Ran into a case right underneath a tower where the Intel WiMAX SDK
reported signal strength as 32767 which is pretty clearly an overflow
of a 16-bit int somewhere in the WiMAX stack. Since our internal
NMWimaxNsp GObject property for signal-quality has a max value of
100, dbus-glib got pretty angry when GObject failed to return
something because the property value was out of range, leading
to malformed message and the bus daemon kicking us off, and finally
an assertion somewhere. Fun.
Apparently dbus-glib will pass non-NULL parameters to proxy
callbacks in the error case, so we've got to make sure we don't
touch any return parameters if there's been an error.
It's easier to be able to pass both a whole hashed connection, or just
a hashed setting to the function, and have it figure out what needs
to be updated based on the given setting_name.
Add some testcases to make sure that all works correctly too.
Timestamps are no longer written to the connection file itself, but
are kept in a lookaside file in /var to allow for read-only or
stateless /etc and to ease system administration and deployment.
NM updates timestamp for active connections every 5 min. We don't
want to touch files in /etc due to this. This commit solves that
by not updating timestamp in the connection's property. Rather it
updates the timestamp internally. All timestamps are also kept track
of in /var/lib/NetworkManager/timestamps file.
When settings are requested via D-Bus GetSettings(), the proper
timestamp is put in the connection setting before returning.
We only need newer glib for G_DEFINE_BOXED_TYPE, which is only
required for introspection. If you don't want introspection, you
don't need that, so allow building on glib 2.24 without.
We may be compatible with even earlier glib versions too, if so
let me know and I'll downgrade the requirement further.
gtk-doc removes the 'html' directory as part of the build process
so we need to have spec.html generation depend on the gtk-doc html
target. Also, fix up the paths to make sure the generated spec.html
gets copied into the final 'html' directory.
