fix bug in here where prev pointer was set to the wrong thing

fix bug in here where prev pointer was set to the wrong thing

2001-01-30  Havoc Pennington  <hp@redhat.com>

	* gtk/gtkliststore.c (gtk_list_store_insert_before): fix bug in
	here where prev pointer was set to the wrong thing

	* gtk/gtktreemodel.c (gtk_tree_path_is_ancestor): new function
	(gtk_tree_path_is_descendant): new function

	* gtk/gtkliststore.c (gtk_list_store_iter_n_children): return
	cached length
	(gtk_list_store_get_iter): don't modify iter if we can't get the
	path.

	* gtk/gtkliststore.h (struct _GtkListStore): cache the length

	* gtk/gtktreednd.h: add virtual function row_drop_possible() to
	GtkTreeDragDest

	* gtk/gtktreestore.c (copy_node_data): fix varargs type error that
	was causing segfault

	* gtk/gtktreedatalist.c (_gtk_tree_data_list_node_copy): set next
	pointer to NULL

	* gtk/gtktreestore.c (gtk_tree_store_append): fix memleak

	* gtk/gtkliststore.c (gtk_list_store_iter_next): don't modify iter
	on returning FALSE
	(gtk_list_store_iter_children): ditto
	(gtk_list_store_iter_nth_child): ditto
	(gtk_list_store_iter_nth_child): ditto
	(gtk_list_store_iter_parent): ditto

	* gtk/gtktreestore.c (gtk_tree_store_get_path): g_return_if_fail
	on iter->user_data != NULL instead of silently accepting it.
	(gtk_tree_store_iter_next): ditto. Also, don't modify iter unless
	we are returning TRUE.
	(gtk_tree_store_iter_children): ditto
	(gtk_tree_store_iter_nth_child): ditto
	(gtk_tree_store_iter_parent): ditto
	(gtk_tree_store_insert): remove handling of parent->user_data ==
	NULL, replace with parent == NULL

	* gtk/gtktreemodel.c (inserted_callback): put some fixes in here,
	and a comment explaining things

	* gtk/gtktreestore.c: add GtkTreeDragSource/GtkTreeDragDest
	interface support to GtkTreeStore.

	* gtk/gtktreemodel.c (gtk_tree_path_prev): didn't properly return
	FALSE if no prev, fix

	* gtk/gtktreeview.c (set_source_row): use a row reference
	(set_dest_row): use a row reference