weasyprint 64.1__py3-none-any.whl → 65.0__py3-none-any.whl

weasyprint/layout/float.py

@@ -68,7 +68,7 @@ def float_layout(context, box, containing_block, absolute_boxes, fixed_boxes,
             context, box, bottom_space=bottom_space,
             skip_stack=skip_stack, containing_block=containing_block,
             page_is_empty=True, absolute_boxes=absolute_boxes,
-            fixed_boxes=fixed_boxes)
+            fixed_boxes=fixed_boxes, discard=False)
     elif isinstance(box, boxes.GridContainerBox):
         box, resume_at, _, _, _ = grid_layout(
             context, box, bottom_space=bottom_space,
@@ -216,9 +216,11 @@ def avoid_collisions(context, box, containing_block, outer=True):
                 # bound.
                 position_x = max_right_bound - box_width
             else:
-                # The position of the right border of the replaced box is at
-                # the right bound.
-                assert isinstance(box, boxes.BlockReplacedBox)
+                # The position of the right border of the replaced box or
+                # formatting context is at the right bound.
+                assert (
+                    isinstance(box, boxes.BlockReplacedBox) or
+                    box.establishes_formatting_context())
                 position_x = max_right_bound - box_width
 
     available_width = max_right_bound - max_left_bound

weasyprint/layout/inline.py

@@ -218,7 +218,7 @@ def skip_first_whitespace(box, skip_stack):
     return None
 
 
-def remove_last_whitespace(context, box):
+def remove_last_whitespace(context, line):
     """Remove in place space characters at the end of a line.
 
     This also reduces the width and position of the inline parents of the
@@ -226,6 +226,7 @@ def remove_last_whitespace(context, box):
 
     """
     ancestors = []
+    box = line
     while isinstance(box, (boxes.LineBox, boxes.InlineBox)):
         ancestors.append(box)
         if not box.children:
@@ -244,25 +245,16 @@ def remove_last_whitespace(context, box):
         assert resume is None
         space_width = box.width - new_box.width
         box.width = new_box.width
-
-        # RTL line, the trailing space is at the left of the box. We have to
-        # translate the box to align the stripped text with the right edge of
-        # the box.
-        if new_box.pango_layout.first_line_direction % 2:
-            box.position_x -= space_width
-            for ancestor in ancestors:
-                ancestor.position_x -= space_width
     else:
         space_width = box.width
         box.width = 0
         box.text = ''
 
-        # RTL line, the textbox with a trailing space is now empty at the left
-        # of the line. We have to translate the line to align it with the right
-        # edge of the box.
-        line = ancestors[0]
-        if line.style['direction'] == 'rtl':
-            line.translate(dx=-space_width, ignore_floats=True)
+    # RTL line, the trailing space is at the left of the box. We have to translate the
+    # box to align the stripped text with the right edge of the box.
+    if box.pango_layout.first_line_direction % 2:
+        for child in line.children:
+            child.translate(dx=-space_width, ignore_floats=True)
 
     for ancestor in ancestors:
         ancestor.width -= space_width
@@ -370,9 +362,12 @@ def atomic_box(context, box, position_x, skip_stack, containing_block,
         if box.is_table_wrapper:
             containing_size = (containing_block.width, containing_block.height)
             table_wrapper_width(context, box, containing_size)
+            width, min_width, max_width = box.width, box.min_width, box.max_width
         box = inline_block_box_layout(
             context, box, position_x, skip_stack, containing_block,
             absolute_boxes, fixed_boxes)
+        if box.is_table_wrapper:
+            box.width, box.min_width, box.max_width = width, min_width, max_width
     else:  # pragma: no cover
         raise TypeError(f'Layout for {type(box).__name__} not handled yet')
     return box
@@ -513,7 +508,7 @@ def split_inline_level(context, box, position_x, max_x, bottom_space,
                 setattr(box, f'margin_{side}', 0)
         new_box, resume_at, _, _, _ = flex_layout(
             context, box, -inf, skip_stack, containing_block, False,
-            absolute_boxes, fixed_boxes)
+            absolute_boxes, fixed_boxes, False)
         preserved_line_break = False
         first_letter = '\u2e80'
         last_letter = '\u2e80'
@@ -1049,8 +1044,7 @@ def inline_box_verticality(box, top_bottom_subtrees, baseline_y):
 
         # the child’s `top` is `child.baseline` above (lower y) its baseline.
         top = child_baseline_y - child.baseline
-        box_types = (
-            boxes.InlineBlockBox, boxes.InlineFlexBox, boxes.InlineGridBox)
+        box_types = (boxes.InlineBlockBox, boxes.InlineFlexBox, boxes.InlineGridBox)
         if isinstance(child, box_types):
             # This also includes table wrappers for inline tables.
             child.translate(dy=top - child.position_y)
@@ -1119,18 +1113,17 @@ def text_align(context, line, available_width, last):
 def justify_line(context, line, extra_width):
     # TODO: We should use a better algorithm here, see
     # https://www.w3.org/TR/css-text-3/#justify-algos
-    nb_spaces = count_spaces(line)
-    if nb_spaces == 0:
-        return
-    add_word_spacing(context, line, extra_width / nb_spaces, 0)
+    if (nb_spaces := count_expandable_spaces(line)):
+        add_word_spacing(context, line, extra_width / nb_spaces, 0)
 
 
-def count_spaces(box):
+def count_expandable_spaces(box):
+    """Count expandable spaces (space and nbsp) for justification."""
     if isinstance(box, boxes.TextBox):
         # TODO: remove trailing spaces correctly
-        return box.text.count(' ')
+        return box.text.count(' ') + box.text.count('\u00a0')
     elif isinstance(box, (boxes.LineBox, boxes.InlineBox)):
-        return sum(count_spaces(child) for child in box.children)
+        return sum(count_expandable_spaces(child) for child in box.children)
     else:
         return 0
 
@@ -1139,7 +1132,7 @@ def add_word_spacing(context, box, justification_spacing, x_advance):
     if isinstance(box, boxes.TextBox):
         box.justification_spacing = justification_spacing
         box.position_x += x_advance
-        nb_spaces = count_spaces(box)
+        nb_spaces = count_expandable_spaces(box)
         if nb_spaces > 0:
             layout = create_layout(
                 box.text, box.style, context, max_width=None,
@@ -1152,7 +1145,10 @@ def add_word_spacing(context, box, justification_spacing, x_advance):
     elif isinstance(box, (boxes.LineBox, boxes.InlineBox)):
         box.position_x += x_advance
         previous_x_advance = x_advance
-        for child in box.children:
+        children = box.children
+        if box.style['direction'] == 'rtl':
+            children = children[::-1]
+        for child in children:
             if child.is_in_normal_flow():
                 x_advance = add_word_spacing(
                     context, child, justification_spacing, x_advance)

weasyprint/layout/page.py

@@ -605,10 +605,8 @@ def make_page(context, root_box, page_type, resume_at, page_number,
         previous_resume_at = resume_at
         root_box = root_box.copy_with_children([])
 
-    # TODO: handle cases where the root element is something else.
-    # See https://www.w3.org/TR/CSS21/visuren.html#dis-pos-flo
-    assert isinstance(root_box, (
-        boxes.BlockBox, boxes.FlexContainerBox, boxes.GridContainerBox))
+    # https://www.w3.org/TR/css-display-4/#root
+    assert isinstance(root_box, boxes.BlockLevelBox)
     context.create_block_formatting_context()
     context.current_page = page_number
     context.current_page_footnotes = []

weasyprint/layout/percent.py

@@ -21,8 +21,7 @@ def percentage(value, refer_to):
         return refer_to * value.value / 100
 
 
-def resolve_one_percentage(box, property_name, refer_to,
-                           main_flex_direction=None):
+def resolve_one_percentage(box, property_name, refer_to):
     """Set a used length value from a computed length value.
 
     ``refer_to`` is the length for 100%. If ``refer_to`` is not a number, it
@@ -35,9 +34,7 @@ def resolve_one_percentage(box, property_name, refer_to,
     percent = percentage(value, refer_to)
     setattr(box, property_name, percent)
     if property_name in ('min_width', 'min_height') and percent == 'auto':
-        if (main_flex_direction is None or
-                property_name != (f'min_{main_flex_direction}')):
-            setattr(box, property_name, 0)
+        setattr(box, property_name, 0)
 
 
 def resolve_position_percentages(box, containing_block):
@@ -48,7 +45,7 @@ def resolve_position_percentages(box, containing_block):
     resolve_one_percentage(box, 'bottom', cb_height)
 
 
-def resolve_percentages(box, containing_block, main_flex_direction=None):
+def resolve_percentages(box, containing_block):
     """Set used values as attributes of the box object."""
     if isinstance(containing_block, boxes.Box):
         # cb is short for containing block
@@ -69,8 +66,8 @@ def resolve_percentages(box, containing_block, main_flex_direction=None):
     resolve_one_percentage(box, 'padding_top', maybe_height)
     resolve_one_percentage(box, 'padding_bottom', maybe_height)
     resolve_one_percentage(box, 'width', cb_width)
-    resolve_one_percentage(box, 'min_width', cb_width, main_flex_direction)
-    resolve_one_percentage(box, 'max_width', cb_width, main_flex_direction)
+    resolve_one_percentage(box, 'min_width', cb_width)
+    resolve_one_percentage(box, 'max_width', cb_width)
 
     # XXX later: top, bottom, left and right on positioned elements
 
@@ -83,14 +80,12 @@ def resolve_percentages(box, containing_block, main_flex_direction=None):
         else:
             assert height.unit == 'px'
             box.height = height.value
-        resolve_one_percentage(box, 'min_height', 0, main_flex_direction)
-        resolve_one_percentage(box, 'max_height', inf, main_flex_direction)
+        resolve_one_percentage(box, 'min_height', 0)
+        resolve_one_percentage(box, 'max_height', inf)
     else:
         resolve_one_percentage(box, 'height', cb_height)
-        resolve_one_percentage(
-            box, 'min_height', cb_height, main_flex_direction)
-        resolve_one_percentage(
-            box, 'max_height', cb_height, main_flex_direction)
+        resolve_one_percentage(box, 'min_height', cb_height)
+        resolve_one_percentage(box, 'max_height', cb_height)
 
     collapse = box.style['border_collapse'] == 'collapse'
     # Used value == computed value
@@ -102,38 +97,8 @@ def resolve_percentages(box, containing_block, main_flex_direction=None):
             setattr(box, prop, box.style[prop])
 
     # Shrink *content* widths and heights according to box-sizing
-    # Thanks heavens and the spec: Our validator rejects negative values
-    # for padding and border-width
-    if box.style['box_sizing'] == 'border-box':
-        horizontal_delta = (
-            box.padding_left + box.padding_right +
-            box.border_left_width + box.border_right_width)
-        vertical_delta = (
-            box.padding_top + box.padding_bottom +
-            box.border_top_width + box.border_bottom_width)
-    elif box.style['box_sizing'] == 'padding-box':
-        horizontal_delta = box.padding_left + box.padding_right
-        vertical_delta = box.padding_top + box.padding_bottom
-    else:
-        assert box.style['box_sizing'] == 'content-box'
-        horizontal_delta = 0
-        vertical_delta = 0
-
-    # Keep at least min_* >= 0 to prevent funny output in case box.width or
-    # box.height become negative.
-    # Restricting max_* seems reasonable, too.
-    if horizontal_delta > 0:
-        if box.width != 'auto':
-            box.width = max(0, box.width - horizontal_delta)
-        box.max_width = max(0, box.max_width - horizontal_delta)
-        if box.min_width != 'auto':
-            box.min_width = max(0, box.min_width - horizontal_delta)
-    if vertical_delta > 0:
-        if box.height != 'auto':
-            box.height = max(0, box.height - vertical_delta)
-        box.max_height = max(0, box.max_height - vertical_delta)
-        if box.min_height != 'auto':
-            box.min_height = max(0, box.min_height - vertical_delta)
+    adjust_box_sizing(box, 'width')
+    adjust_box_sizing(box, 'height')
 
 
 def resolve_radii_percentages(box):
@@ -154,3 +119,33 @@ def resolve_radii_percentages(box):
             rx = percentage(rx, box.border_width())
             ry = percentage(ry, box.border_height())
             setattr(box, property_name, (rx, ry))
+
+
+def adjust_box_sizing(box, axis):
+    if box.style['box_sizing'] == 'border-box':
+        if axis == 'width':
+            delta = (
+                box.padding_left + box.padding_right +
+                box.border_left_width + box.border_right_width)
+        else:
+            delta = (
+                box.padding_top + box.padding_bottom +
+                box.border_top_width + box.border_bottom_width)
+    elif box.style['box_sizing'] == 'padding-box':
+        if axis == 'width':
+            delta = box.padding_left + box.padding_right
+        else:
+            delta = box.padding_top + box.padding_bottom
+    else:
+        assert box.style['box_sizing'] == 'content-box'
+        delta = 0
+
+    # Keep at least min_* >= 0 to prevent funny output in case box.width or
+    # box.height become negative.
+    # Restricting max_* seems reasonable, too.
+    if delta > 0:
+        if getattr(box, axis) != 'auto':
+            setattr(box, axis, max(0, getattr(box, axis) - delta))
+        setattr(box, f'max_{axis}', max(0, getattr(box, f'max_{axis}') - delta))
+        if getattr(box, f'min_{axis}') != 'auto':
+            setattr(box, f'min_{axis}', max(0, getattr(box, f'min_{axis}') - delta))

weasyprint/layout/preferred.py

@@ -43,14 +43,12 @@ def min_content_width(context, box, outer=True):
         return table_and_columns_preferred_widths(context, box, outer)[0]
     elif isinstance(box, boxes.TableCellBox):
         return table_cell_min_content_width(context, box, outer)
-    elif isinstance(box, (
-            boxes.BlockContainerBox, boxes.TableColumnBox, boxes.FlexBox)):
+    elif isinstance(box, (boxes.BlockContainerBox, boxes.TableColumnBox)):
         return block_min_content_width(context, box, outer)
     elif isinstance(box, boxes.TableColumnGroupBox):
         return column_group_content_width(context, box)
     elif isinstance(box, (boxes.InlineBox, boxes.LineBox)):
-        return inline_min_content_width(
-            context, box, outer, is_line_start=True)
+        return inline_min_content_width(context, box, outer, is_line_start=True)
     elif isinstance(box, boxes.ReplacedBox):
         return replaced_min_content_width(box, outer)
     elif isinstance(box, boxes.FlexContainerBox):
@@ -59,8 +57,7 @@ def min_content_width(context, box, outer=True):
         # TODO: Get real grid size.
         return block_min_content_width(context, box, outer)
     else:
-        raise TypeError(
-            f'min-content width for {type(box).__name__} not handled yet')
+        raise TypeError(f'min-content width for {type(box).__name__} not handled yet')
 
 
 def max_content_width(context, box, outer=True):
@@ -73,14 +70,12 @@ def max_content_width(context, box, outer=True):
         return table_and_columns_preferred_widths(context, box, outer)[1]
     elif isinstance(box, boxes.TableCellBox):
         return table_cell_min_max_content_width(context, box, outer)[1]
-    elif isinstance(box, (
-            boxes.BlockContainerBox, boxes.TableColumnBox, boxes.FlexBox)):
+    elif isinstance(box, (boxes.BlockContainerBox, boxes.TableColumnBox)):
         return block_max_content_width(context, box, outer)
     elif isinstance(box, boxes.TableColumnGroupBox):
         return column_group_content_width(context, box)
     elif isinstance(box, (boxes.InlineBox, boxes.LineBox)):
-        return inline_max_content_width(
-            context, box, outer, is_line_start=True)
+        return inline_max_content_width(context, box, outer, is_line_start=True)
     elif isinstance(box, boxes.ReplacedBox):
         return replaced_max_content_width(box, outer)
     elif isinstance(box, boxes.FlexContainerBox):
@@ -89,8 +84,7 @@ def max_content_width(context, box, outer=True):
         # TODO: Get real grid size.
         return block_max_content_width(context, box, outer)
     else:
-        raise TypeError(
-            f'max-content width for {type(box).__name__} not handled yet')
+        raise TypeError(f'max-content width for {type(box).__name__} not handled yet')
 
 
 def _block_content_width(context, box, function, outer):
@@ -344,7 +338,7 @@ def inline_line_widths(context, box, outer, is_line_start, minimum, skip_stack=N
             # "By default, there is a break opportunity
             #  both before and after any inline object."
             if minimum:
-                lines = [0, max_content_width(context, child), 0]
+                lines = [0, min_content_width(context, child), 0]
             else:
                 lines = [max_content_width(context, child)]
         # The first text line goes on the current line
@@ -673,9 +667,12 @@ def replaced_min_content_width(box, outer=True):
             intrinsic_width, intrinsic_height, intrinsic_ratio = (
                 image.get_intrinsic_size(
                     box.style['image_resolution'], box.style['font_size']))
-            width, _ = default_image_sizing(
-                intrinsic_width, intrinsic_height, intrinsic_ratio, 'auto',
-                height, default_width=300, default_height=150)
+            if intrinsic_ratio and not intrinsic_width and not intrinsic_height:
+                width = 0
+            else:
+                width, _ = default_image_sizing(
+                    intrinsic_width, intrinsic_height, intrinsic_ratio, 'auto',
+                    height, default_width=300, default_height=150)
     elif box.style['width'].unit == '%':
         # See https://drafts.csswg.org/css-sizing/#intrinsic-contribution
         width = 0

weasyprint/pdf/fonts.py

@@ -1,6 +1,7 @@
 """Fonts integration in PDF."""
 
 import io
+import re
 from hashlib import md5
 from logging import WARNING
 from math import ceil
@@ -34,7 +35,7 @@ class Font:
                 part.split('=')[0]: float(part.split('=')[1])
                 for part in ffi.string(variations).decode().split(',')}
         if weight := self.variations.get('weight'):
-            self.weight = int(round(weight))
+            self.weight = round(weight)
             pango.pango_font_description_set_weight(description, weight)
         else:
             self.weight = pango.pango_font_description_get_weight(description)
@@ -58,24 +59,18 @@ class Font:
             in md5(description_string, usedforsecurity=False).digest()[:6])
 
         # Set font name.
-        fields = description_string.split(b' ')
-        if fields and b'=' in fields[-1]:
-            fields.pop()  # Remove variations
-        if fields:
-            fields.pop()  # Remove font size
-        else:
-            fields = [b'Unknown']
-        self.name = b'/' + self.hash.encode() + b'+' + b'-'.join(fields)
+        name = re.split(b' [#@]', description_string)[0]
+        self.name = b'/' + self.hash.encode() + b'+' + name.replace(b' ', b'-')
 
         # Set ascent and descent.
         if self.font_size:
             pango_metrics = pango.pango_font_get_metrics(pango_font, ffi.NULL)
-            self.ascent = int(round(
+            self.ascent = round(
                 pango.pango_font_metrics_get_ascent(pango_metrics) * FROM_UNITS /
-                self.font_size * 1000))
-            self.descent = -int(round(
+                self.font_size * 1000)
+            self.descent = -round(
                 pango.pango_font_metrics_get_descent(pango_metrics) * FROM_UNITS /
-                self.font_size * 1000))
+                self.font_size * 1000)
         else:
             self.ascent = self.descent = 0
 
@@ -112,7 +107,7 @@ class Font:
         self.flags = 2 ** (3 - 1)  # Symbolic, custom character set
         if self.style:
             self.flags += 2 ** (7 - 1)  # Italic
-        if b'Serif' in fields:
+        if b'Serif' in name.split(b' '):
             self.flags += 2 ** (2 - 1)  # Serif
 
     def clean(self, cmap, hinting):
@@ -342,15 +337,19 @@ def build_fonts_dictionary(pdf, fonts, compress, subset, options):
             b'/CMapType 2 def',
             b'1 begincodespacerange',
             b'<0000> <ffff>',
-            b'endcodespacerange',
-            f'{len(cmap)} beginbfchar'.encode()], compress=compress)
-        for glyph, text in cmap.items():
-            unicode_codepoints = ''.join(
-                f'{letter.encode("utf-16-be").hex()}' for letter in text)
-            to_unicode.stream.append(
-                f'<{glyph:04x}> <{unicode_codepoints}>'.encode())
+            b'endcodespacerange'], compress=compress)
+        cmap_length = len(cmap)
+        cmap_items = tuple(cmap.items())
+        for i in range(ceil(cmap_length / 100)):
+            batch_length = min(100, cmap_length - i * 100)
+            to_unicode.stream.append(f'{batch_length} beginbfchar'.encode())
+            for glyph, text in cmap_items[i*100:(i+1)*100]:
+                unicode_codepoints = ''.join(
+                    f'{letter.encode("utf-16-be").hex()}' for letter in text)
+                to_unicode.stream.append(
+                    f'<{glyph:04x}> <{unicode_codepoints}>'.encode())
+            to_unicode.stream.append(b'endbfchar')
         to_unicode.stream.extend([
-            b'endbfchar',
             b'endcmap',
             b'CMapName currentdict /CMap defineresource pop',
             b'end',
@@ -561,7 +560,7 @@ def _build_vector_font_dictionary(font_dictionary, pdf, font, widths, compress,
     })
     if str(pdf_version) <= '1.4':  # Cast for bytes and None
         cids = sorted(font.widths)
-        padded_width = int(ceil((cids[-1] + 1) / 8))
+        padded_width = ceil((cids[-1] + 1) / 8)
         bits = ['0'] * padded_width * 8
         for cid in cids:
             bits[cid] = '1'

weasyprint/stacking.py

@@ -82,6 +82,7 @@ def _dispatch(box, page, child_contexts, blocks, floats, blocks_and_cells):
         child_contexts.append(StackingContext.from_box(box, page))
         return
 
+    stacking_classes = (boxes.InlineBlockBox, boxes.InlineFlexBox, boxes.InlineGridBox)
     if style['position'] != 'static':
         assert style['z_index'] == 'auto'
         # "Fake" context: sub-contexts will go in this `child_contexts` list.
@@ -91,8 +92,7 @@ def _dispatch(box, page, child_contexts, blocks, floats, blocks_and_cells):
         child_contexts.insert(index, stacking_context)
     elif box.is_floated():
         floats.append(StackingContext.from_box(box, page, child_contexts))
-    elif isinstance(box, (
-            boxes.InlineBlockBox, boxes.InlineFlexBox, boxes.InlineGridBox)):
+    elif isinstance(box, stacking_classes):
         # Have this fake stacking context be part of the "normal" box tree,
         # because we need its position in the middle of a tree of inline boxes.
         return StackingContext.from_box(box, page, child_contexts)

weasyprint/svg/__init__.py

@@ -467,11 +467,13 @@ class SVG:
         # Draw node children
         if node.display and node.tag not in DEF_TYPES:
             for child in node:
-                if text_anchor_shift:
+                new_chunk = text_anchor_shift and (
+                    child.tag == 'text' or 'x' in child.attrib or 'y' in child.attrib)
+                if new_chunk:
                     new_stream = self.stream
                     self.stream = original_streams[-1]
                 self.draw_node(child, font_size, fill_stroke)
-                if text_anchor_shift:
+                if new_chunk:
                     self.stream = new_stream
                 visible_text_child = (
                     TAGS.get(node.tag) == text and
@@ -502,7 +504,8 @@ class SVG:
                     x - font_size, y - font_size,
                     x + width + font_size, y + height + font_size)
                 x_align = width / 2 if text_anchor == 'middle' else width
-                self.stream.transform(e=-x_align)
+                if node.tag == 'text' or 'x' in node.attrib or 'y' in node.attrib:
+                    self.stream.transform(e=-x_align)
             self.stream.draw_x_object(group_id)
             self.stream.pop_state()
 

weasyprint/text/constants.py

@@ -49,6 +49,11 @@ PANGO_VARIANT = {
     'titling-caps': pango.PANGO_VARIANT_TITLE_CAPS,
 }
 
+PANGO_DIRECTION = {
+    'ltr': pango.PANGO_DIRECTION_LTR,
+    'rtl': pango.PANGO_DIRECTION_RTL,
+}
+
 # Language system tags
 # From https://docs.microsoft.com/typography/opentype/spec/languagetags
 LST_TO_ISO = {

weasyprint/text/ffi.py

@@ -166,6 +166,16 @@ ffi.cdef('''
         PANGO_ELLIPSIZE_END
     } PangoEllipsizeMode;
 
+    typedef enum {
+        PANGO_DIRECTION_LTR,
+        PANGO_DIRECTION_RTL,
+        PANGO_DIRECTION_TTB_LTR,
+        PANGO_DIRECTION_TTB_RTL,
+        PANGO_DIRECTION_WEAK_LTR,
+        PANGO_DIRECTION_WEAK_RTL,
+        PANGO_DIRECTION_NEUTRAL
+    } PangoDirection;
+
     typedef struct GSList {
        gpointer data;
        struct GSList *next;
@@ -278,6 +288,7 @@ ffi.cdef('''
     void pango_layout_set_wrap (PangoLayout *layout, PangoWrapMode wrap);
     void pango_layout_set_single_paragraph_mode (PangoLayout *layout, gboolean setting);
     void pango_layout_set_ellipsize (PangoLayout *layout, PangoEllipsizeMode ellipsize);
+    void pango_layout_set_auto_dir (PangoLayout *layout, gboolean auto_dir);
     int pango_layout_get_baseline (PangoLayout *layout);
     void pango_layout_line_get_extents (
         PangoLayoutLine *line, PangoRectangle *ink_rect, PangoRectangle *logical_rect);
@@ -360,6 +371,7 @@ ffi.cdef('''
     PangoLanguage * pango_language_from_string (const char *language);
     PangoLanguage * pango_language_get_default (void);
     void pango_context_set_language (PangoContext *context, PangoLanguage *language);
+    void pango_context_set_base_dir (PangoContext *context, PangoDirection direction);
 
     void pango_get_log_attrs (
         const char *text, int length, int level, PangoLanguage *language,

weasyprint/text/fonts.py

@@ -208,14 +208,23 @@ class FontConfiguration:
             match = SubElement(root, 'match', target='font')
             test = SubElement(match, 'test', name='file', compare='eq')
             SubElement(test, 'string').text = str(font_path)
-            edit = SubElement(match, 'edit', name='fontfeatures', mode=mode)
             descriptors = {
                 rules[0][0].replace('-', '_'): rules[0][1] for rules in
                 rule_descriptors.get('font_variant', [])}
             settings = rule_descriptors.get('font_feature_settings', 'normal')
             features = font_features(font_feature_settings=settings, **descriptors)
-            for key, value in features.items():
-                SubElement(edit, 'string').text = f'{key} {value}'
+            if features:
+                edit = SubElement(match, 'edit', name='fontfeatures', mode=mode)
+                for key, value in features.items():
+                    SubElement(edit, 'string').text = f'{key} {value}'
+            if unicode_ranges := rule_descriptors.get('unicode_range'):
+                edit = SubElement(match, 'edit', name='charset', mode=mode)
+                plus = SubElement(edit, 'plus')
+                for unicode_range in unicode_ranges:
+                    charset = SubElement(plus, 'charset')
+                    range_ = SubElement(charset, 'range')
+                    for value in (unicode_range.start, unicode_range.end):
+                        SubElement(range_, 'int').text = f'0x{value:x}'
             header = (
                 b'<?xml version="1.0"?>',
                 b'<!DOCTYPE fontconfig SYSTEM "urn:fontconfig:fonts.dtd">')

weasyprint/text/line_break.py

@@ -5,7 +5,7 @@ from math import inf
 
 import pyphen
 
-from .constants import LST_TO_ISO, PANGO_WRAP_MODE
+from .constants import LST_TO_ISO, PANGO_DIRECTION, PANGO_WRAP_MODE
 from .ffi import FROM_UNITS, TO_UNITS, ffi, gobject, pango, pangoft2, unicode_to_char_p
 from .fonts import font_features, get_font_description
 
@@ -76,6 +76,8 @@ class Layout:
             pango.pango_font_map_create_context(font_map),
             gobject.g_object_unref)
         pango.pango_context_set_round_glyph_positions(pango_context, False)
+        pango.pango_context_set_base_dir(
+            pango_context, PANGO_DIRECTION[style['direction']])
 
         if style['font_language_override'] != 'normal':
             lang_p, lang = unicode_to_char_p(LST_TO_ISO.get(
@@ -96,6 +98,7 @@ class Layout:
         self.layout = ffi.gc(
             pango.pango_layout_new(pango_context),
             gobject.g_object_unref)
+        pango.pango_layout_set_auto_dir(self.layout, False)
         pango.pango_layout_set_font_description(self.layout, font_description)
 
         text_decoration = style['text_decoration_line']
@@ -187,13 +190,11 @@ class Layout:
                     pango.pango_layout_set_text(self.layout, text, -1)
 
                 space_spacing = int(word_spacing * TO_UNITS + letter_spacing)
-                position = bytestring.find(b' ')
                 # Pango gives only half of word-spacing on boundaries
                 boundary_positions = (0, len(bytestring) - 1)
-                while position != -1:
-                    factor = 1 + (position in boundary_positions)
-                    add_attr(position, position + 1, factor * space_spacing)
-                    position = bytestring.find(b' ', position + 1)
+                for match in re.finditer(' |\u00a0'.encode(), bytestring):
+                    factor = 1 + (match.start() in boundary_positions)
+                    add_attr(match.start(), match.end(), factor * space_spacing)
 
             if word_breaking:
                 attr = pango.pango_attr_insert_hyphens_new(False)
@@ -213,7 +214,7 @@ class Layout:
             layout.set_text(' ' * self.style['tab_size'])
             line, _ = layout.get_first_line()
             width, _ = line_size(line, self.style)
-            width = int(round(width))
+            width = round(width)
         else:
             width = int(self.style['tab_size'].value)
         # 0 is not handled correctly by Pango

{weasyprint-64.1.dist-info → weasyprint-65.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: weasyprint
-Version: 64.1
+Version: 65.0
 Summary: The Awesome Document Factory
 Keywords: html,css,pdf,converter
 Author-email: Simon Sapin <simon.sapin@exyr.org>
@@ -30,7 +30,7 @@ Requires-Dist: pydyf >=0.11.0
 Requires-Dist: cffi >=0.6
 Requires-Dist: tinyhtml5 >=2.0.0b1
 Requires-Dist: tinycss2 >=1.4.0
-Requires-Dist: cssselect2 >=0.1
+Requires-Dist: cssselect2 >=0.8.0
 Requires-Dist: Pyphen >=0.9.1
 Requires-Dist: Pillow >=9.1.0
 Requires-Dist: fonttools[woff] >=4.0.0