Add page boxes.

weasy/formatting_structure/boxes.py

@@ -40,6 +40,7 @@ Abstract classes, should not be instanciated:
 
 Concrete classes:
 
+ * PageBox
  * BlockBox
  * AnonymousBlockBox
  * InlineBox
@@ -49,8 +50,8 @@ Concrete classes:
  * TextBox
  * LineBox
 
-Apart from LineBox, all concrete box classes have one of the following "outside"
-behavior:
+Apart from PageBox and LineBox, all concrete box classes have one of the
+following "outside" behavior:
 
  * Block-level (inherits from BlockLevelBox)
  * Inline-level (inherits from InlineLevelBox)
@@ -139,6 +140,32 @@ class BlockContainerBox(Box):
     """
 
 
+class PageBox(BlockContainerBox):
+    """
+    Initially the whole document will be in a single page box. During layout
+    a new page box is created after every page break.
+    """
+    def __init__(self, document, page_number):
+        self.document = document
+        # starting at 1 for the first page.
+        self.page_number = page_number
+        # Page boxes are not linked to any element.
+        super(PageBox, self).__init__(element=None)
+
+    def _init_style(self):
+        # First page is a right page. TODO: this "should depend on the major
+        # writing direction of the document".
+        first_is_right = True
+        is_right = (self.page_number % 2) == (1 if first_is_right else 0)
+        page_type = 'right' if is_right else 'left'
+        if self.page_number == 1:
+            page_type = 'first_' + page_type
+        style = self.document.page_pseudo_elements[page_type].style
+        # Copying might not be needed, but let’s be careful with mutable
+        # objects.
+        self.style = style.copy()
+
+
 class BlockBox(BlockContainerBox, BlockLevelBox):
     """
     A block-level box that is also a block container.
@@ -180,7 +207,8 @@ class LineBox(AnonymousBox):
     inline-level boxes.
 
     In early stages of building the box tree a single line box contains many
-    consecutive inline boxes and will be split later when wrapping lines.
+    consecutive inline boxes. Later, during layout phase, each line boxes will
+    be split into multiple line boxes, one for each actual line.
     """
 
 

weasy/formatting_structure/build.py

@@ -32,11 +32,14 @@ def build_formatting_structure(document):
     """
     Build a formatting structure (box tree) from a DOM document.
     """
-    box_tree = dom_to_box(document)
-    inline_in_block(box_tree)
-    block_in_inline(box_tree)
-    process_whitespace(box_tree)
-    return box_tree
+    box = dom_to_box(document)
+    inline_in_block(box)
+    block_in_inline(box)
+    process_whitespace(box)
+
+    page = PageBox(document, 1)
+    page.add_child(box)
+    return page
 
 
 def dom_to_box(element):

weasy/layout.py

@@ -28,6 +28,6 @@ def compute_dimensions(box):
     """
 
 
-@compute_dimensions.register(boxes.BlockLevelBox)
+@compute_dimensions.register(boxes.PageBox)
 def block_level_box_dimensions(box):
     return box

weasy/tests/test_boxes.py

@@ -55,6 +55,10 @@ def unwrap_html_body(box):
     and remove them to simplify further tests. These are always at the root
     of HTML documents.
     """
+    if isinstance(box, boxes.PageBox):
+        assert len(box.children) == 1
+        box = box.children[0]
+
     assert isinstance(box, boxes.BlockBox)
     assert box.element.tag == 'html'
     assert len(box.children) == 1

weasy/utils.py

@@ -25,7 +25,7 @@ class MultiFunction(object):
     A callable with different implementations depending on the type of the
     first argument.
     
-    This objects takes __name__, __module__ and __doc__ from base_function
+    This object takes __name__, __module__ and __doc__ from base_function
     if it is given, but does not use it’s body.
     """