gtk_cell_area_foreach_alloc() itterates over an allocated list of cells
and calls a callback with the cell's allocation until the callback returns
TRUE.
  - Removed vfunc get_cell_allocation() and implemented
    gtk_cell_area_get_cell_allocation() using foreach_alloc()
  - Added gtk_cell_area_get_cell_at_position() using foreach_alloc()
  - Removed GtkCellAreaBox ->event() implementation and implemented
    mouse "click" cell activation in GtkCellArea class directly using
    gtk_cell_area_get_cell_at_position().

Also gave a boolean return value to the callback to allow breaking out of the loop.

The function has been re-implemented around GtkCellArea.  This commits
also brings back the invocation of this function in
gtk_tree_view_button_press().  I shouldn't have removed this.

By renaming it to x_offset.

This test displays a GtkTreeViewColumn rendering a vertically
oriented GtkCellArea along side a horizontally oriented column
in the same treeview.

Now consult gtk_cell_area_context_get_preferred_height().

It can be that height-for-widths requested here were too large
when multiple cells are in play because of the alignments stored
in the context... removing the temporary focus-line-width hack.

Allow feeding treeviewcolumn a custom cell-area (or not a custom one,
but allow sharing the cell-area with say, the combo-box area).

This patch also:

 - Fixes signal connections to the area (now they do eventually get
   disconnected at dispose time, they are handled regardless if a
   treeview is set but execute safely, at least there is only one
   connection/disconnection).

 - Fixes refcounting on the cell_area (GtkCellArea is GInitiallyUnowned).

 - Adds a constructor() in order to build the cell-area if one has
   not been provided by the caller before hand at g_object_new()
   construct time.

Fixed GtkTreeViewColumn allocation to only subtract the depth padding/indentation from the expander column.

Changed GtkTreeView to bookkeep the deepest expanded depth instead of recalculating it at column allocation time.

This is a premature patch, it traverses the tree's expanded
rows and fetches the deepest depth every time we allocate a
column. The deepest depth should rather be cached and pushed
when a row expands, then recalculated when the deepest expanded
row collapses.

Removed gtk_tree_view_get_real_requested_width_from_column() from
gtktreeview.c in favor of this function in the treeviewcolumn domain
(since this function goes and checks treeviewcolumn internals, settings
and derives the real requested width, seems logical this should be done
by the treeviewcolumn instead).

GtkTreeView duplicated the contents of this function word-for-word
inside gtk_tree_view_update_size().

Now the GtkTreeViewColumn takes care of move/resizing its window and
allocating its button (except for the special case of current drag_column
where the column doesnt actually get reallocated).

This actually much simplifies interaction with GtkTreeSelection
(at least reduces code size where dealing width the treeview anchor path).

Added them directly to gtktreeview.c.

Some details:
  - button_request was not needed, consult the minimum request of the button
  - gtk_tree_view_column_get_button() needed to be public as people can set
    tooltips on the button (and libgail accesses the button).

This fixes commit fb3429e5

Not a good idea to cast an instance to a class...

Grr, gdk_window_class_class_get_type should have been
gdk_window_window_class_get_type.

Even though the style properties have the GdkColor type, the other
gtk_widget_override_* API takes RGBA colors, so it is consistent
now.

Conflicts:
	gtk/gtkmarshalers.list
	tests/Makefile.am