According to the latest mockups for the connection editor dialog [1],
the IPv4 and IPv6 pages are supposed to use a table-like editor to
manage the addresses, in a similar fashion of what was done to the
routes editor. This way of editing is not only easier to comprehend,
but also improves the size of the dialog, requiring much less vertical
space to present the routes.
The current implementation, however, uses a vertical layout and a toolbar,
which is inefficient in its usage of space.
Fix that by implementing the table-like editor widget, both in IPv4
and IPv6 pages.
[1] https://raw.githubusercontent.com/gnome-design-team/gnome-mockups/master/system-settings/network/aday2/network-wires.pnghttps://bugzilla.gnome.org/show_bug.cgi?id=779841
According to the latest mockups for the connection editor dialog [1],
the IPv4 and IPv6 pages are supposed to use a table-like editor to
manage the routes. This editor is not only easier to comprehend, but
also improves the size of the dialog, requiring much less vertical
space to present the routes.
The current implementation, however, uses a vertical layout and a toolbar,
which is inefficient in its usage of space.
Fix that by implementing the table-like editor widget, both in IPv4
and IPv6 pages.
[1] https://raw.githubusercontent.com/gnome-design-team/gnome-mockups/master/system-settings/network/aday2/network-wires.pnghttps://bugzilla.gnome.org/show_bug.cgi?id=779841
When calling for the wireless security widgets, the code
simply assumes that the corresponding GType is initialized.
This may not always be true, which leads to a nasty crash
every time e.g. we open the network connection editor dialog.
This commit fixes that by introducing a new standard macro
wrapping wireless_security_get_type(), and ensuring the type
is initializing when calling wireless_security_init(), thus
protecting every code path from this crash.
This commit also makes CePageSecurity use the new macro for
better legibility.
https://bugzilla.gnome.org/show_bug.cgi?id=785581
Firewalld is a Fedora-only daemon, and it doesn't integrate all that
well with Fedora Workstation, the version of Fedora that uses GNOME, so
remove the support.
In Fedora Workstation, firewalld only ever supports one zone, as the
other ones are badly defined, and not translatable or translated.
The pointer has already been dereferenced. Might have made more sense
to check if it's the empty string and treat that as error, but
nm_utils_ipaddr_valid also returns error on empty string so lets just
rely on the utility function doing the right thing here.
https://bugzilla.gnome.org/show_bug.cgi?id=708500
Null-checking the text_gateway makes no sense as it has already been
dereferenced earlier. Instead check if it's the empty string and
treat that as not-an-error. Also prevent the empty string from
being set in default_gateway.
https://bugzilla.gnome.org/show_bug.cgi?id=708500
From https://bugzilla.gnome.org/show_bug.cgi?id=765969 as explained by
Dan Winship:
"
libnm-util/libnm-glib had a buggy data model, which nm-connection-editor
(and then gnome-control-center) copied, in which each manually-configured
IP address has an associated gateway address. In reality, NM always just
took the first non-empty gateway value from the address array, and
completely ignored any other gateway values.
libnm represents this more accurately, by having a single gateway
value which is separate from the address array. Ideally, the editors should
show it this way as well (eg, like nmtui does). Failing that, it would
be nice to at least make it so that only the first row in the address
table can have a non-empty gateway value.
"
We went for the second option, only showing a gateway entry for the
first address in the list.
This isn't related to route-specific gateway addresses.
https://bugzilla.gnome.org/show_bug.cgi?id=765969
The implementation of ->validate in the parent class will return TRUE if
there's no implementation in the child class, so no need to implement
those.
https://bugzilla.gnome.org/show_bug.cgi?id=769230
When changing the method from Manual to Automatic, we need to clear the
gateway setting, otherwise the settings verification will fail:
ipv6.gateway: gateway cannot be set if there are no addresses configured
Another fallout of the libnm 1.2 port
https://bugzilla.gnome.org/show_bug.cgi?id=769230
We still offered to add Bond, Team, Bridge and VLAN connections even
though we don't support them any more. Remove the first page as we only
offer to add VPN connections.
https://bugzilla.gnome.org/show_bug.cgi?id=767614
We also remove support for WiMAX (now unsupported by NetworkManager),
and InfiniBand (Enterprise feature), and the use of
the deprecated NM_SETTING_WIRELESS_SEC property.
With help from network-manager-applet patches by Jiří Klimeš and
Dan Winship.
https://bugzilla.gnome.org/show_bug.cgi?id=765910
Currently, if firewalld is not running, then the firewall zone combo is
always insensitive. This would make sense if firewalld was great and we
were to consider it a distro bug to not have firewalld running and
enabled, then users would be able to see the option is disabled and
think "maybe my distro has done something wrong."
But in actuality, firewalld is not really great, only Fedora ships it
and we're not really recommending it to other distros. So hide the combo
if it's not running, else it's just broken on $NOTFEDORA.
https://bugzilla.gnome.org/show_bug.cgi?id=763477
Since LIBDIR of gnome-control-center and the VPN plugins aren't
necessarily the same, use the paths as specified by the VPN plugins and
only reconstruct the path if it is not absolute or we fail to load the
plugin.
Patch from Debian package, by Michael Biebl <biebl@debian.org>
https://bugzilla.gnome.org/show_bug.cgi?id=764509
It doesn't make sense to try to validate pages because widgets are
changing due to the whole dialog being destroyed and it causes a bunch
of warnings because some resources are cleared on each page dispose
method.
Avoid all that by disconnecting the page "changed" signal handler
before we start destroying widgets.
We don't need secrets for new connections and, in fact, trying to
retrieve secrets in that case will fail because we have a plain
NMConnection instead of a NMRemoteConnection.
The above mentioned error would result in the page never being
initialized.
