Added support for Select fields as combo boxes

I've decoded two PDFs, one with the combo box and another blank and extracted the binary info necessary to render it and applied to the `anchors.add_inputs` function.

docs/api_reference.rst

@@ -765,7 +765,7 @@ The ``resize``, ``cursor``, ``caret-*`` and ``nav-*`` properties are **not**
 supported.
 
 The ``appearance`` property is supported. When set to ``auto``, it displays
-form fields as PDF form fields (supported for text inputs, check boxes and
-text areas only).
+form fields as PDF form fields (supported for text inputs, check boxes, text
+areas, and select only).
 
 The ``accent-color`` property is **not** supported.

tests/test_api.py

@@ -482,15 +482,23 @@ def test_partial_pdf_custom_metadata():
     assert b'value' in stdout
 
 
-@pytest.mark.parametrize('html, field', (
-    (b'<input>', b'/Tx'),
-    (b'<input type="checkbox">', b'/Btn'),
-    (b'<textarea></textarea>', b'/Tx'),
+@pytest.mark.parametrize('html, fields', (
+    (b'<input>', [b'/Tx']),
+    (b'<input type="checkbox">', [b'/Btn']),
+    (b'<textarea></textarea>', [b'/Tx']),
+    (b'<select><option value="a">A</option></select>', [b'/Ch', b'/Opt']),
+    # The selected values will be (b) and (c) in the PDF.
+    (b'<select multiple>'
+     b'<option value="a">A</option>'
+     b'<option value="b" selected>B</option>'
+     b'<option value="c" selected>C</option>'
+     b'</select>'
+     , [b'/Ch', b'/Opt', b'[(b) (c)]']),
 ))
-def test_pdf_inputs(html, field):
+def test_pdf_inputs(html, fields):
     stdout = _run('--pdf-forms --uncompressed-pdf - -', html)
     assert b'AcroForm' in stdout
-    assert field in stdout
+    assert all(field in stdout for field in fields)
     stdout = _run('--uncompressed-pdf - -', html)
     assert b'AcroForm' not in stdout
 

weasyprint/layout/block.py

@@ -833,6 +833,29 @@ def block_container_layout(context, box, bottom_space, skip_stack,
     if next_page['page'] is None:
         next_page['page'] = new_box.page_values()[1]
 
+    if (
+        box.element and
+        box.element.tag == 'select'
+    ):
+        if (
+            box.element.attrib.get('multiple') is not None and
+            not box.element.attrib.get('height') and
+            not box.element.attrib.get('max-height')
+        ):
+            # Overriding the height when the select has the multiple attribute
+            # and no height attribute
+            options_number = len([child for child in box.element])
+            new_box.height = min(box.height * 4, box.height * options_number)
+
+        if (
+            (padding_right:=box.style.get("padding_right")) and
+            (padding_left:=box.style.get("padding_left")) and
+            padding_right > padding_left
+        ):
+            # Overriding padding right because it's greater due to the select
+            # arrow
+            new_box.style["padding_right"] = padding_left
+
     return (
         new_box, resume_at, next_page, adjoining_margins, collapsing_through,
         max_lines)

weasyprint/pdf/anchors.py

@@ -160,6 +160,79 @@ def add_inputs(inputs, matrix, pdf, page, resources, stream, font_map,
                 'AS': '/Yes' if checked else '/Off',
                 'DA': pydyf.String(b' '.join(field_stream.stream)),
             })
+        elif element.tag == "select":
+            # Text, password, textarea, files, and unknown
+            font_description = get_font_description(style)
+            font = pango.pango_font_map_load_font(
+                font_map, context, font_description)
+            font = stream.add_font(font)
+            font.used_in_forms = True
+
+            field_stream.set_font_size(font.hash, font_size)
+            multiple = element.attrib.get("multiple") is not None
+            options = []
+            selected_values = []
+            for option in element:
+                options.append(
+                    pydyf.Array(
+                        [f'({option.attrib.get("value")})', f'({option.text})']
+                    )
+                )
+                if option.attrib.get("selected") is not None:
+                    selected_values.append(
+                        pydyf.String(option.attrib.get("value"))
+                    )
+
+            if multiple:
+                field = pydyf.Dictionary({
+                    'DA': pydyf.String(b' '.join(field_stream.stream)),
+                    'F': 2 ** (3 - 1),  # Print flag
+                    'FT': '/Ch',
+                    # To be a multiselect list we need to set the 21st bit to 1
+                    # outputing 2097152:
+                    # https://opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandards/PDF32000_2008.pdf # noqa: E501
+                    # This specification can be found on the link above on page
+                    # 445.
+                    'Ff': (1 << 21),
+                    'Opt': pydyf.Array(options),
+                    'P': page.reference,
+                    'Rect': pydyf.Array(rectangle),
+                    'Subtype': '/Widget',
+                    'T': pydyf.String(input_name),
+                    'Type': '/Annot',
+                    # The select value is a list of selected values that come
+                    # from the options with the selected property. If there are
+                    # no selected values, then the value is an empty string.
+                    'V': pydyf.Array(selected_values)
+                    if len(selected_values)
+                    else pydyf.String(''),
+                })
+            else:
+                field = pydyf.Dictionary({
+                    'DA': pydyf.String(b' '.join(field_stream.stream)),
+                    'F': 2 ** (3 - 1),  # Print flag
+                    'FT': '/Ch',
+                    # To be a combo box we need to set the 17th bit to 1
+                    # outputing 131072:
+                    # https://opensource.adobe.com/dc-acrobat-sdk-docs/pdfstandards/PDF32000_2008.pdf # noqa: E501
+                    # This specification can be found on the link above on page
+                    # 444.
+                    'Ff': (1 << 17),
+                    'Opt': pydyf.Array(options),
+                    'P': page.reference,
+                    'Rect': pydyf.Array(rectangle),
+                    'Subtype': '/Widget',
+                    'T': pydyf.String(input_name),
+                    'Type': '/Annot',
+                    # Get the last value in the list when there are multiple
+                    # selected values in a single select box (this is the same
+                    # approach used by browsers).
+                    # If there are no selected values, then the value is an
+                    # empty string.
+                    'V': selected_values[-1]
+                    if len(selected_values)
+                    else pydyf.String(''),
+                })
         else:
             # Text, password, textarea, files, and unknown
             font_description = get_font_description(style)