https://bugzilla.gnome.org/show_bug.cgi?id=710297

https://bugzilla.gnome.org/show_bug.cgi?id=710297

Now all DnD icons are dealt with as cairo_surface_t, if pixbufs are used
in the data model, a cairo surface is created out of it first.

https://bugzilla.gnome.org/show_bug.cgi?id=736046

With recent versions of Adwaita/GTK+.

Apply all the necessary changes by hand, rather than relying on a
GtkStyleContext being able to transfer its style to a PangoLayout.

https://bugzilla.gnome.org/show_bug.cgi?id=735966

Matthias Clasen authored Sep 02, 2014 and Bastien Nocera committed Sep 03, 2014

GTK+ changed from using the active state to a new checked
state for this. Adapt the toggle cell renderer.

https://bugzilla.gnome.org/show_bug.cgi?id=735646

Matthias Clasen authored Sep 02, 2014 and Bastien Nocera committed Sep 03, 2014

Adwaita provides styling only for .app-notification.frame, so
add the .frame style class as well. And don't draw a shadow.

https://bugzilla.gnome.org/show_bug.cgi?id=735680

Draw background and border for the check box so that it can be styled
with css.

https://bugzilla.gnome.org/show_bug.cgi?id=733823

So that prelight works.

https://bugzilla.gnome.org/show_bug.cgi?id=732656

libgd's Makefile.am refers to source files in subdirectories, which
causes warnings with automake 1.14. Enabling the 'subdir-objects'
option silences these.

libgd/Makefile.am:29: warning: source file 'libgd/gd-types-catalog.c' is in a subdirectory,
libgd/Makefile.am:29: but option 'subdir-objects' is disabled
automake: warning: possible forward-incompatibility.
automake: At least a source file is in a subdirectory, but the 'subdir-objects'
automake: automake option hasn't been enabled.  For now, the corresponding output
automake: object file(s) will be placed in the top-level directory.  However,
automake: this behaviour will change in future Automake versions: they will
automake: unconditionally cause object files to be placed in the same subdirectory
automake: of the corresponding sources.
automake: You are advised to start using 'subdir-objects' option throughout your
automake: project, to avoid future incompatibilities.

https://bugzilla.gnome.org/show_bug.cgi?id=732190

Fallout from 41820ae9

This is necessary to get GtkPopovers pointing to the tag, rather
than the entry altogether.

https://bugzilla.gnome.org/show_bug.cgi?id=723030

When we were calculating the text's sizes and coordinates,
we didn't apply the subtitle style to the 2nd line's Pango layout,
leading to the subtitle being placed as if it was normal sized.

So that they are smaller with the latest Adwaita master.

Otherwise all the items in the (tree or list) store would be
selected, instead of the (filtered) items shown on the screen.

We only supported list stores, but the rest of the widget supports
tree stores just as well.

Similar to a7a62888

If the signal handler for "selection-mode-request" didn't set
the view in selection mode, don't start poking at rows and changing
the selected state that we won't show.

If the view model was a filter model, the selection icon would only
appear or disappear on mouse hover, when the tree view was fetching the
value of the selected column again. This makes it look like selection
is horribly slow.

This patch makes sure that the icon view is instantly updated.

We only supported list stores, but the rest of the widget supports
tree stores just as well.

When clicking on a tag's close button, the close buttons of all tags
where drawn with the active state (well, except the button you
actually clicked on, which kept showing the hover state), because
of a wrong if-clause and setting two states at the same time.

I'm not sure if it is necessary to avoid setting GTK_STATE_FLAG_ACTIVE
and GTK_STATE_FLAG_PRELIGHT at the same time on the tag buttons, I
guess ACTIVE should take precedence here, but I'll leave that to the
CSS experts to figure out.

https://bugzilla.gnome.org/show_bug.cgi?id=720764

When a GdTaggedEntryTag was inserted into an unmapped but realized
GdTaggedEntry, the tag's GDK window wasn't created. This resulted in a
drawing failure on showing the entry again. Fix this by making sure to
realize the tag at once if the entry is already realized.

Also enhance test-tagged-entry to allow testing this.

https://bugzilla.gnome.org/show_bug.cgi?id=712755

No point in clearing the existing attributes in set_attributes_from_model
if we are not going to actually set any

No point in clearing the existing attributes in set_attributes_from_model
if we are not going to actually set any. Besides this fixes a crash in
Boxes (on app exit).

If the application keeps a reference to the tag, the tag is not
finalized and thus the tag was not unrealized.

When the tag is removed but not unrealized, the secondary icon doesn't
work. When we click on the secondary icon, the event->window is not the
good one.

This is needed to suport HiDpi.

This is needed to support HiDpi icons.

And default it to zero. This way, by default the spinner will not be
drawn when we create an empty model column in the store.

GD_TYPE_TOGGLE_PIXBUF_RENDERER was used in the types catalog when
LIBGD__VIEW_COMMON was defined but the Makefile.am included it
when LIBGD_MAIN_ICON_VIEW was defined.

This will fix distcheck in gnome-control-center if nothing else
is broken ;)

gd-toggle-pixbuf-renderer.h was included in the header when
LIBGD__VIEW_COMMON was defined but the Makefile.am included it
when LIBGD_MAIN_ICON_VIEW was defined.

This fixes distcheck in gnome-control-center.