bug: possible regression in CSS parsing of advanced expressions

[eht16]

I noticed a difference when parsing and removing CSS styles between Bleach 4.x and 5.x.
This is probably related to #633 resp. #648.

Before, styles like list-style: url(https://www.example.com) seem to have been stripped even if list-style was in the styles list.
Recent Bleach versions (5.x) do keep this list-style style and this seems correct to me.

** python and bleach versions (please complete the following information):**

• Python Version: 3.9
• Bleach Version: 4.1 and 5.0

To Reproduce

Steps to reproduce the behavior:

import bleach

allowed_css_styles = ['width', 'list-style']
allowed_attributes = {'*': ['class', 'style', 'id']}
allowed_tags = ['div']

test_html_list_style = '<div style="list-style: url(https://www.example.com); width: 100%;"></div>'
test_html_width = '<div style="width: expression(test);"></div>'

try:
    # Bleach 5.x
    from bleach.css_sanitizer import CSSSanitizer

    css_sanitizer = CSSSanitizer(allowed_css_properties=allowed_css_styles)

    bleached_text = bleach.clean(test_html_list_style, tags=allowed_tags, attributes=allowed_attributes, css_sanitizer=css_sanitizer)
    print(bleached_text)  # gives: <div style="list-style: url(https://www.example.com); width: 100%;"></div>

    bleached_text = bleach.clean(test_html_width, tags=allowed_tags, attributes=allowed_attributes, css_sanitizer=css_sanitizer)
    print(bleached_text)  # gives: <div style="width: expression(test);"></div>

except ImportError:
    # Bleach 4.x
    bleached_text = bleach.clean(test_html_list_style, tags=allowed_tags, attributes=allowed_attributes, styles=allowed_css_styles)
    print(bleached_text)  # gives: <div style="width: 100%;"></div>

    bleached_text = bleach.clean(test_html_width, tags=allowed_tags, attributes=allowed_attributes, styles=allowed_css_styles)
    print(bleached_text)  # gives: <div style=""></div>

Expected behavior

I think the current behavior is correct but I just want to get sure and that I'm not misunderstanding or misusing the new CSS sanitizer.

So, is the previous behavior more or less a bug in the old CSS sanitizer and has been fixed implicitly by the new one?

[g-k]

So, is the previous behavior more or less a bug in the old CSS sanitizer and has been fixed implicitly by the new one?

Yep! The bug was #529.