Due to an API bug in GNOME Online Accounts, the asynchronous
goa_provider_get_all method doesn't accept a GCancellable argument.
This makes it difficult to cancel an ongoing call when the CcGoaPanel
gets destroyed.
Prior to commit c26f8ae018, this was hacked around by taking a
reference on the panel for the duration of the call. Instead of
cancelling a pending call on destruction, it would keep the panel alive
until the call was over. However, that was lost during commit
c26f8ae018.
One thing to bear in mind is that GtkWidgets, CcGoaPanel is one, can
be destroyed by a gtk_widget_destroy call, which is subtly different
than a simple sequence of g_object_unref calls. When gtk_widget_destroy
is used, it invokes the GObject::dispose virtual method of the widget.
It is expected this will cause anything holding a reference to this
widget to drop their references, leading to GObject::finalize being
called. However, there is no guarantee that this will happen in the
same iteration of the GMainLoop. Therefore, it is possible that when
the goa_provider_get_all call finishes, the CcGoaPanel might be in a
disposed, but not yet finalized state.
When a GObject is in a disposed-but-not-finalized state, only a very
limited number of operations can be performed on it. Its reference
count can be altered, the memory used by the instance struct can be
accessed, but none of the member GObjects can be assumed to be valid.
eg., it's definitely illegal to add new rows to the member GtkListBox.
Hence a boolean flag is used to mark the destroyed state of the panel.
This second part is a small improvement over the earlier hack.
https://gitlab.gnome.org/GNOME/gnome-control-center/issues/208
There was an issue where the "minor" axis snapping would not be done if
the "major" axis snapping had a zero distance. This could be seen when e.g.
moving a monitor on the right up/down slightly. In that case, no
snapping to align the bottom/top edges were done unless you also moved
the mouse sideways a bit.
Fixes#211
It is not useful to assert that task_data != NULL after
using it to build the D-Bus key.
Move the assertion to before it is first used and get
rid of this inconsistency.
As per doumentation: "It is a programming error to create an identical proxy
while asynchronously creating one. Not cancelling this operation will result in
an assertion failure when calling cc_object_storage_create_dbus_proxy_finish()."
In order to fullfill the second part we need to check for errors (including
cancellation ones) before we generate an assertion failure.
Fixes: https://gitlab.gnome.org/GNOME/gnome-control-center/issues/158
The code must not access the passed panel if the operation to create the
keyboard proxy has been cancelled. This fixes a possible crash when
switching away from the power panel.
The entry was not bound. Due to this we could run into a segfault trying
to query the hostname_entry. This would only happen if avahi is not
running.
Fixes#178
The build option `-ltinfo` is required, otherwise a few errors related
to libreadline.so will be thrown like:
libreadline.so: undefined reference to `tputs'
libreadline.so: undefined reference to `tgoto'
libreadline.so: undefined reference to `tgetflag'
Correct a typoed variable name so that calibration works again, instead
of crashing.
GLib-GObject-WARNING **: 15:52:50.561: invalid cast from 'GtkBox' to 'GtkEntry'
Automatic timezone does not work if location sharing is disabled in privacy.
Grays out automatic timezone setting if location sharing is disabled in privacy
settings. Enable automatic timezone seting if location sharing is enabled.
https://bugzilla.gnome.org/show_bug.cgi?id=788714
The row widgets are not full GtkWidgets. This means they are harder
to understand, use g_object_set_data which looses type safety and
cause the one .c file to be overly large.
Update this code to use a CcInputRow widget that uses GTK+ best
practice.
Users panel crashes with g_error() if some of libpwquality calls fails,
e.g. when pwquality.conf is broken. The default values should be used
instead of crash. Let's use g_warning() instead of g_error() in that
case.
https://gitlab.gnome.org/GNOME/gnome-control-center/issues/171
um_password_dialog_set_user() is called from on_permission_changed()
respective show_user(), which more or less means on every change in
the panel. It it pretty enough to call this before opening the password
dialog, which is already done over change_password() callback. Let's
remove the redundant call which may lead to unwanted failures among
others.
https://gitlab.gnome.org/GNOME/gnome-control-center/issues/171
The battery power level level bars are using the default offsets for low, high and full battery level, 0.25, 0.75 and 1.0.
Based on the suggestion from #68, this commit changes the high and full offset to 0.1 and 0.8, and the low battery level offset to 0.03 for primary batteries and UPSs, and 0.05 for other batteries and devices.
https://gitlab.gnome.org/GNOME/gnome-control-center/issues/68
Creating the restart file leaked GFile objects. It also stored the path in the
object, which is unnecessary as it can just be generated on demand.
This leak has was introduced when the feature was in
db551f1c5d
