We didn't made up our minds until now from some of them, and for
the others no progress has been made or bugs reported, so probably
they are fine.

We were using GTkTreeView in a simple list. Also, as we know,
GtkCellRenderers are not the best way to theme and manipulate
widgets.

So instead use a GtkListBox to modernize the GtkPlacesSidebar,
and in the way clean up some parts of the code (like headings)
which were not used anymore.
Also we don't use a model anymore, since the data is simple
enough to manage it in a subclass of the row itself.

It was somewhat annoying given the "trace" of it. Decrease
the time to avoid too much "trace"

Now with the API addition on gtkplacessidebar, we can show some
hints for the drop targets.

https://bugzilla.gnome.org/show_bug.cgi?id=747793

https://bugzilla.gnome.org/show_bug.cgi?id=747793

Following design mockups

https://bugzilla.gnome.org/show_bug.cgi?id=747793

We didn't use those for some time now.

It is convenient to allow applications to show all the drop
targets at once. This improves the user experience with drag
an drop.

The new API allows the application to set the gtkplacessidebar
in a mode where invalid drop targets are insensitive and it
adds a "new bookmark" row. This mode is intended to be set
when the application is aware of a dnd operation and needs to
be stopped kwhen the application is aware that dnd operation
was cancelled or ended in a different part than gtkplacesisdebar.

The context parameter is unused in this patch, but will be
used in next patches when the sidebar will use a GtkListBox.
The reason of being unused now is just convenience.

https://bugzilla.gnome.org/show_bug.cgi?id=747793

Following the new mockups, put the new bookmark row always
in the first position, so it's easier to drop an item.

https://bugzilla.gnome.org/show_bug.cgi?id=747793

GtkInspector is opening a separate display connection, which makes
it more likely that gtk_get_current_event() returns an event from
the "wrong" display.

This was broken in 0796d7b6.

https://bugzilla.gnome.org/show_bug.cgi?id=751018

We were getting ourselves in trouble by casting touch events
to GdkEventButton and poking directly at their internals. Instead,
use GdkEvent API to get what we need.

This fixes a crash when using the gear menu in epiphany with
touch. The same crash also occurred in testmenubutton.

The iteration now progresses past a delete, so make sure we iterate
safely. Also, don't chain up if we removed a child.

https://bugzilla.gnome.org/show_bug.cgi?id=750568

https://bugs.freedesktop.org/show_bug.cgi?id=90917

https://bugzilla.gnome.org/show_bug.cgi?id=750568

https://bugs.freedesktop.org/show_bug.cgi?id=90917

For these widgets we set pass-through on the child window so that
input over these widgets (that are not on a child input window) goes
to the window below the overlay.

https://bugzilla.gnome.org/show_bug.cgi?id=750568

https://bugs.freedesktop.org/show_bug.cgi?id=90917

This allows you to control the z-ordering of overlay children

https://bugzilla.gnome.org/show_bug.cgi?id=750568

https://bugs.freedesktop.org/show_bug.cgi?id=90917

An pass_through window is something you can draw in but does not
affect event handling. Normally if a window has with no event mask set
for a particular event then input events in it go to its parent window
(X11 semantics), whereas if pass_through is enabled the window below
the window will get the event. The later mode is useful when the
window is partially transparent. Note that an pass-through windows can
have child windows that are not pass-through so they can still get events
on some parts.

Semantically, this behaves the same as an regular window with
gdk_window_set_child_input_shapes() called on it (and re-called any
time a child is changed), but its far more efficient and easy to use.

This allows us to fix the testoverlay input stacking test.

https://bugzilla.gnome.org/show_bug.cgi?id=750568

https://bugs.freedesktop.org/show_bug.cgi?id=90917

In this case we have a bunch of interactive main children
of the overlay, and then a centered overlay that contains both
non-interactive (labels) and interactive (entry) widgets.

This shows off a problem where the non-interactive parts (the labels)
steals input from the overlay main children (breaks button click and
hover effects).

https://bugzilla.gnome.org/show_bug.cgi?id=750568

https://bugs.freedesktop.org/show_bug.cgi?id=90917

For functions that take state flags as an argument we need to special
case the situation where the passed in flags don't match the current
state.

Previously we would create a copy of the style info, change its state
and do the lookup from there.

Now that GtkCssNode has replaced style infos, this doesn't work as well
anymore as copying a GtkCssNode is not possible.
However, unike style infos, GtkCssNodes are instant-apply, so we don't
need to copy anymore, we can just change the state of the node.

This causes some invalidations to be queued, but we can take that
performance hit as this is fallback code.

https://bugzilla.redhat.com/show_bug.cgi?id=1228852

This tests the new assistant api to remove padding from
pages.

This can be useful when embedding complex widgets like
a file chooser as a page.

https://bugzilla.gnome.org/show_bug.cgi?id=750631

This is one of the cases where one wants a page without
any padding.

Instead of having padding outside the notebook containing
all pages, put each page in an extra box and add the padding
there. This is in preparation for allowing pages without
padding.

This means the pages themselves will be responsible for setting
suitable padding.

https://bugzilla.gnome.org/show_bug.cgi?id=750631

We are using this to test pango functionality that
uses fontconfig apis, so this doesn't build on OS X
or Windows anymore.

https://bugzilla.gnome.org/show_bug.cgi?id=750892

The popups on the color editor are somewhat hard to get rid of.
Arrange things so that clicking the same control again dismisses
the current popup.

There's nothing to customize here, so no need to show a menu.

clang complains that these may be used uninitialized.

clang complained about transient_for being used uninitialized.