This commit does a number of things:
- remove some dead wchar configury from configure.ac and gdkconfig.h
- repurpose gdkconfig.h as header that contains GDK_WINDOWING_foo
  macros for each included backend, include it in gdk.h and install
  it in $includedir instead of below $libdir
- drop the backend from the library names
- build libgdk-3.0.la as a convenience lib and include it in libgtk-3.0.la
It does not yet enable building multiple backends at the same time.

This lets us change this without affecting apps.

There is no need for apps to access this class, as its only implemented
in the gdk backends, and we want to be free to change it later.

Now a cell can either have a "fixed" size or it can have
an "aligned" starting point or both. "fixed" size cells take
no space when they are invisible.

Added _gtk_cell_area_set_cell_data_func_with_proxy() to be called by
gtk_cell_layout_set_cell_data_func() when the layouting object itself
is not the underlying cell area.

Even if the default CSS contains values for these, handle non-existing
values as queries might happen on an incomplete style.

https://bugzilla.gnome.org/show_bug.cgi?id=637606

https://bugzilla.gnome.org/show_bug.cgi?id=637606

Since the default style also has a backing context, it wasn't
being replaced after initialization.

Since commit 9552152d

After the window removal a while ago, the calendar main window
was not properly moved in size_allocate. Also, we ought to hide/show
the windows in map/unmap, not keep them visible at all times.

Bug 634657

Docs have also been improved, to make explicit the device
type/source accepted.

So that custom settings can be added from the custom-widget-apply
callback.

https://bugzilla.gnome.org/show_bug.cgi?id=637069

Previously, for performance reasons we would sometimes
skip invoking the unmap signal (and associated vfunc)
in favor of simply unrealizing. However, widgets then
had no way to clean stuff up when they were hidden
(but still inside a parent which was shown).

This patch also removes _gtk_tooltip_hide() which
was done in both unmap and unrealize in gtkwidget.c,
now can only be in unmap.

There are probably lots of things cleaned up in
unrealize that would now be better to move to unmap.

https://bugzilla.gnome.org/show_bug.cgi?id=629923

Requires fixes to GtkContainer and GtkWindow to unmap their
children, rather than just withdrawing or hiding the container
window.

Requires fix to GtkHandleBox to chain up to GtkContainer unmap.

Historically we avoided these unmaps for efficiency reasons,
but these days it's a bigger problem that there's no way
for child widgets to know that one of their ancestors has
become unmapped.

These checks are a bit expensive so require --enable-debug=yes.
gtk_widget_verify_invariants() checks invariants mentioned
in docs/widget_system.txt in particular, and can verify
others in the future.

Some of the invariants in docs/widget_system.txt don't
in fact hold right now, so those are #if 0'd in this
patch pending someone fixing either the docs or the code.

When it was converted to no-window, we forgot to make it hide/show
its input window as necessary.

Also, remove some dead code, pointed out in bug 637608.

Fixes a bug in mutter where it would resort to a dummy style
to get iconview rubberband color.

Reported by Mathieu Bridon in
https://bugzilla.gnome.org/show_bug.cgi?id=637520

Pointed out in bug 637471.

The dialog will always show up with the last selected application.

These strings are likely user-provided values, so we should
respect locale settings.

Reported in bug 637189.

gdk_test_simulate_*() uses XSendEvent, which doesn't currently work
with XI2/GenericEvents, so make tests use core events for the time
being. Luckily there's a lot more to test than low-level event
handling in these tests.

This fixes a failure when running make test, where style queries
happened before there was even an screen.

This fixes commit d0b81b21