Don’t use "text-transform"ed text for content-based uses

This solution is far from perfect. What we could do is to have a getter/setter for the "text" attribute of TextBox using the "original_text" attribute, so that we’re sure that we change one when we change the other. But… That would be much more complicated and much slower, only to solve corner use cases. Of course, an issue will be open soon, when someone finds one of these cases ;). Fix #137.
Kozea · Sep 6, 2021 · 85e1ea1 · 85e1ea1
1 parent 083cab0
commit 85e1ea1
Show file tree

Hide file tree

Showing 3 changed files with 19 additions and 8 deletions.
diff --git a/tests/test_pdf.py b/tests/test_pdf.py
@@ -266,6 +266,19 @@ def test_bookmarks_12():
     assert re.findall(b'/Title \\((.*)\\)', pdf) == [b'a']
 
 
+@assert_no_logs
+def test_bookmarks_13():
+    pdf = FakeHTML(string='''
+      <div style="bookmark-level:1; bookmark-label:contents;
+                  text-transform:uppercase">a</div>
+    ''').write_pdf()
+    # a
+    counts = re.findall(b'/Count ([0-9-]*)', pdf)
+    outlines = counts.pop()
+    assert outlines == b'1'
+    assert re.findall(b'/Title \\((.*)\\)', pdf) == [b'a']
+
+
 @assert_no_logs
 def test_links_none():
     pdf = FakeHTML(string='<body>').write_pdf()

diff --git a/weasyprint/formatting_structure/boxes.py b/weasyprint/formatting_structure/boxes.py
@@ -475,6 +475,7 @@ class TextBox(InlineLevelBox):
     def __init__(self, element_tag, style, element, text):
         assert text
         super().__init__(element_tag, style, element)
+        self.original_text = text
         text_transform = style['text_transform']
         if text_transform != 'none':
             text = {

diff --git a/weasyprint/formatting_structure/build.py b/weasyprint/formatting_structure/build.py
@@ -374,6 +374,7 @@ def add_text(text):
         has_text.add(True)
         if text:
             if content_boxes and isinstance(content_boxes[-1], boxes.TextBox):
+                content_boxes[-1].original_text += text
                 content_boxes[-1].text += text
             else:
                 content_boxes.append(
@@ -649,12 +650,8 @@ def parse_again(mixin_pagebased_counters={}):
     box_list = compute_content_list(
         content_list, box, counter_values, css_token, parse_again,
         target_collector, counter_style, element=element)
-
-    if box_list is None:
-        box.bookmark_label = ''
-    else:
-        box.bookmark_label = ''.join(
-            box.text for box in box_list if isinstance(box, boxes.TextBox))
+    if box_list:
+        box.bookmark_label = ''.join(box_text(box) for box in box_list)
 
 
 def set_content_lists(element, box, style, counter_values, target_collector,
@@ -1550,10 +1547,10 @@ def set_viewport_overflow(root_box):
 
 def box_text(box):
     if isinstance(box, boxes.TextBox):
-        return box.text
+        return box.original_text
     elif isinstance(box, boxes.ParentBox):
         return ''.join(
-            child.text for child in box.descendants()
+            child.original_text for child in box.descendants()
             if not child.element_tag.endswith('::before') and
             not child.element_tag.endswith('::after') and
             not child.element_tag.endswith('::marker') and