Support images in <embed> or <object>

CHANGES

@@ -1,11 +1,15 @@
 Version 0.6, released on 2012-XX-XX
 ===================================
 
-* *Backward incompatible*: completely change the API. See the documentation:
-  http://weasyprint.org/using/#as-a-python-library
+* *Backward incompatible*: completely change the Python API.
+  See the documentation: http://weasyprint.org/using/#as-a-python-library
+* *Backward incompatible*: Proper margin collapsing.
+  This changes how blocks are rendered: adjoining margins "collapse"
+  (their maximum is used) instead of accumulating.
+* Support images in ``embed`` or ``object`` elements.
 * Switch to pystacia instead of PIL for raster images
-* Add compatibility with CPython 2.6 and 3.2 (was only 2.7 previously)
-* Proper margin collapsing
+* Add compatibility with CPython 2.6 and 3.2. (Previously only 2.7
+  was supported)
 * Many bug fixes
 
 

weasyprint/document.py

@@ -86,14 +86,14 @@ def pages(self):
         """
         return layout.layout_document(self, self.formatting_structure)
 
-    def get_image_from_uri(self, uri):
+    def get_image_from_uri(self, uri, type_=None):
         """
         Same as ``weasy.images.get_image_from_uri`` but cache results
         """
         missing = object()
         surface = self._image_cache.get(uri, missing)
         if surface is missing:
-            surface = images.get_image_from_uri(uri)
+            surface = images.get_image_from_uri(uri, type_)
             self._image_cache[uri] = surface
         return surface
 

weasyprint/html.py

@@ -102,10 +102,9 @@ def make_text_box(element, box, text):
                  box.style, [text_box])
 
 
-# TODO: also support images (including SVG) in <embed> and <object> elements?
 @handler('img')
 def handle_img(document, element, box):
-    """Handle ``<img>`` tags, return either an image or the alt-text.
+    """Handle ``<img>`` elements, return either an image or the alt-text.
 
     See: http://www.w3.org/TR/html5/embedded-content-1.html#the-img-element
 
@@ -135,18 +134,54 @@ def handle_img(document, element, box):
             return []
 
 
+@handler('embed')
+def handle_embed(document, element, box):
+    """Handle ``<embed>`` elements, return either an image or nothing.
+
+    See: http://www.w3.org/TR/html5/the-iframe-element.html#the-embed-element
+
+    """
+    src = get_url_attribute(element, 'src')
+    type_ = element.get('type', '').strip()
+    if src:
+        image = document.get_image_from_uri(src, type_)
+        if image is not None:
+            return [make_replaced_box(element, box, image)]
+    # No fallback.
+    return []
+
+
+@handler('object')
+def handle_object(document, element, box):
+    """Handle ``<object>`` elements, return either an image or the fallback
+    content.
+
+    See: http://www.w3.org/TR/html5/the-iframe-element.html#the-object-element
+
+    """
+    data = get_url_attribute(element, 'data')
+    type_ = element.get('type', '').strip()
+    if data:
+        image = document.get_image_from_uri(data, type_)
+        if image is not None:
+            return [make_replaced_box(element, box, image)]
+    # The element’s children are the fallback.
+    return [box]
+
+
 def integer_attribute(element, box, name, minimum=1):
     """Read an integer attribute from the HTML element and set it on the box.
 
     """
     value = element.get(name, '').strip()
-    try:
-        value = int(value)
-    except ValueError:
-        pass
-    else:
-        if value >= minimum:
-            setattr(box, name, value)
+    if value:
+        try:
+            value = int(value)
+        except ValueError:
+            pass
+        else:
+            if value >= minimum:
+                setattr(box, name, value)
 
 
 @handler('colgroup')
@@ -157,6 +192,9 @@ def handle_colgroup(_document, element, box):
             box.span = None  # sum of the children’s spans
         else:
             integer_attribute(element, box, 'span')
+            box.children = (
+                boxes.TableColumnBox.anonymous_from(box, [])
+                for _i in xrange(box.span))
     return [box]
 
 

weasyprint/images.py

@@ -116,15 +116,18 @@ def fallback_handler(file_like, uri):
         return png_handler(BytesIO(png_bytes), uri)
 
 
-def get_image_from_uri(uri):
+def get_image_from_uri(uri, type_=None):
     """Get a :class:`cairo.Surface`` from an image URI."""
     try:
         file_like, mime_type, _charset = urlopen(uri)
     except (IOError, ValueError) as exc:
         LOGGER.warn('Error while fetching an image from %s : %r', uri, exc)
         return None
 
-    handler = FORMAT_HANDLERS.get(mime_type, fallback_handler)
+    if not type_:
+        type_ = mime_type  # Use eg. the HTTP header
+    #else: the type was forced by eg. a 'type' attribute on <embed>
+    handler = FORMAT_HANDLERS.get(type_, fallback_handler)
     exception = None
     try:
         image = handler(file_like, uri)

weasyprint/layout/blocks.py

@@ -351,7 +351,6 @@ def block_level_height(document, box, max_position_y, skip_stack,
     # TODO: See corner cases in
     # http://www.w3.org/TR/CSS21/visudet.html#normal-block
     if new_box.height == 'auto':
-        print(box, position_y, new_box.content_box_y())
         new_box.height = position_y - new_box.content_box_y()
 
     if resume_at is not None:

weasyprint/tests/test_layout.py

@@ -842,11 +842,24 @@ def test_whitespace_processing():
 def test_with_images():
     """Test that width, height and ratio of images are respected."""
     # Try a few image formats
-    for url in ['pattern.png', 'pattern.gif', 'blue.jpg', 'pattern.svg',
-                "data:image/svg+xml,<svg width='4' height='4'></svg>",
-                "data:image/svg+xml,<svg width='4px' height='4px'></svg>",
-                ]:
-        page, = parse('<img src="%s">' % url)
+    for html in [
+        '<img src="%s">' % url for url in [
+            'pattern.png', 'pattern.gif', 'blue.jpg', 'pattern.svg',
+            "data:image/svg+xml,<svg width='4' height='4'></svg>",
+            "data:image/svg+xml,<svg width='4px' height='4px'></svg>",
+        ]
+    ] + [
+        '<embed src=pattern.png>',
+        '<embed src=pattern.svg>',
+        '<embed src=really-a-png.svg type=image/png>',
+        '<embed src=really-a-svg.png type=image/svg+xml>',
+
+        '<object data=pattern.png>',
+        '<object data=pattern.svg>',
+        '<object data=really-a-png.svg type=image/png>',
+        '<object data=really-a-svg.png type=image/svg+xml>',
+    ]:
+        page, = parse(html)
         html, = page.children
         body, = html.children
         line, = body.children
@@ -877,11 +890,12 @@ def test_with_images():
         'data:image/png,Not a PNG',
         'data:image/jpeg,Not a JPEG',
         'data:image/svg+xml,<svg>invalid xml',
+        'really-a-svg.png',
     ]:
         with capture_logs() as logs:
             page, = parse("<p><img src='%s' alt='invalid image'>" % url)
         assert len(logs) == 1
-        if url.startswith('data:'):
+        if url.startswith('data:') or 'really' in url:
             assert 'WARNING: Error while parsing an image' in logs[0]
         else:
             assert 'WARNING: Error while fetching an image' in logs[0]

weasyprint/tests/w3_test_suite/templates/run_test.html

@@ -4,9 +4,9 @@
     iframe {
         border: 2px inset;
         padding: 0;
-        width: 660px;
-        height: 660px;
-        box-sizing: border-box
+        width: 640px;
+        height: 640px;
+        box-sizing: content-box
     }
     .render {
         display: table-cell;
@@ -23,7 +23,8 @@ <h2>{{ test.title }}</h2>
 
 <div class=render>
     <h3>WeasyPrint</h3>
-    <iframe src="{{ url_for('render', test_id=test.test_id) }}"></iframe>
+    <iframe src="{{ url_for('render', test_id=test.test_id) }}"
+            marginwidth=0 marginheight=0></iframe>
 </div>
 
 <div class=render>