Escaped characters in CSS are not properly kept

[benoit74]

This test is failing:

def test_rewrite_css2(no_js_notify):
    output = (
        HtmlRewriter(
            ArticleUrlRewriter(HttpUrl("http://kiwix.org/"), set()),
            "",
            "",
            no_js_notify,
        )
        .rewrite(
            r'<style>.test { content="\e123" }</style>',
        )
        .content
    )
    assert output == r'<style>.test { content="\e123" }</style>'

Looks like "\e123" is interpreted somehow by the parser.

This is encountered in the wild on https://geo.libretexts.org/Courses/Coastline_College/An_Introduction_To_Geology_-_Coastline_College/01%3A_Understanding_Science/1.01%3A_What_is_Science where we have such inline CSS.