Be more tolerant to invalid css.

When tinycss2 fails to parse, we use a regex rewriter to (trying to)
rewrite url in the css.

In case of invalid css tinycss2 doesn't not failed on parsing, but
generate a ParserError token
https://doc.courtbouillon.org/tinycss2/stable/api_reference.html?highlight=parser#tinycss2.ast.ParseError

This is on serialization that exception occurs.
We even try/except all transformation to be sure.

Fix #155

src/warc2zim/content_rewriting/css.py

@@ -1,3 +1,4 @@
+import re
 from collections.abc import Iterable
 
 from tinycss2 import (
@@ -10,27 +11,60 @@
 from tinycss2.serializer import serialize_url
 
 from warc2zim.content_rewriting import UrlRewriterProto
+from warc2zim.content_rewriting.rx_replacer import RxRewriter
+
+
+class FallbackRegexCssRewriter(RxRewriter):
+    def __init__(self, url_rewriter: UrlRewriterProto):
+        rules = [
+            (
+                re.compile(r"""url\((?P<quote>['"])?(?P<url>.+?)(?P=quote)(?<!\\)\)"""),
+                lambda m_object, _opts: "".join(
+                    [
+                        "url(",
+                        m_object["quote"],
+                        url_rewriter(m_object["url"]),
+                        m_object["quote"],
+                        ")",
+                    ]
+                ),
+            )
+        ]
+        super().__init__(rules)
 
 
 class CssRewriter:
     def __init__(self, url_rewriter: UrlRewriterProto):
         self.url_rewriter = url_rewriter
+        self.fallback_rewriter = FallbackRegexCssRewriter(url_rewriter)
 
     def rewrite(self, content: str | bytes) -> str:
-        if isinstance(content, bytes):
-            rules = parse_stylesheet_bytes(content)[0]
-        else:
-            rules = parse_stylesheet(content)
-        self.process_list(rules)
+        try:
+            if isinstance(content, bytes):
+                rules = parse_stylesheet_bytes(content)[0]
+            else:
+                rules = parse_stylesheet(content)
+            self.process_list(rules)
 
-        output = serialize(rules)
+            output = serialize(rules)
+        except Exception:
+            # If tinycss fail to parse css, it will generate a "Error" token.
+            # Exception is raised at serialization time.
+            # We try/catch the whole process to be sure anyway.
+            return self.fallback_rewriter.rewrite_content(content, {})
         return output
 
     def rewrite_inline(self, content: str) -> str:
-        rules = parse_declaration_list(content)
-        self.process_list(rules)
-        output = serialize(rules)
-        return output
+        try:
+            rules = parse_declaration_list(content)
+            self.process_list(rules)
+            output = serialize(rules)
+            return output
+        except Exception:
+            # If tinycss fail to parse css, it will generate a "Error" token.
+            # Exception is raised at serialization time.
+            # We try/catch the whole process to be sure anyway.
+            return self.fallback_rewriter.rewrite_content(content, {})
 
     def process_list(self, components: Iterable[ast.Node]):
         if components:  # May be null

tests/test_css_rewriting.py

@@ -5,14 +5,30 @@
 from warc2zim.content_rewriting.css import CssRewriter
 from warc2zim.url_rewriting import ArticleUrlRewriter
 
+from .utils import ContentForTests
+
 
 @pytest.fixture(
     params=[
-        b"p { color: red; }",
-        b"p {\n color: red;\n}",
-        b"p { background: blue; }",
-        b"p { background: rgb(15, 0, 52); }",
-        b"/* See bug issue at http://exemple.com/issue/link */ p { color: blue; }",
+        ContentForTests(b"p { color: red; }"),
+        ContentForTests(b"p {\n color: red;\n}"),
+        ContentForTests(b"p { background: blue; }"),
+        ContentForTests(b"p { background: rgb(15, 0, 52); }"),
+        ContentForTests(
+            b"/* See bug issue at http://exemple.com/issue/link */ p { color: blue; }"
+        ),
+        ContentForTests(
+            b"p { width= } div { background: url(http://exemple.com/img.png)}",
+            b"p { width= } div { background: url(exemple.com/img.png)}",
+        ),
+        ContentForTests(
+            b"p { width= } div { background: url('http://exemple.com/img.png')}",
+            b'p { width= } div { background: url("exemple.com/img.png")}',
+        ),
+        ContentForTests(
+            b'p { width= } div { background: url("http://exemple.com/img.png")}',
+            b'p { width= } div { background: url("exemple.com/img.png")}',
+        ),
     ]
 )
 def no_rewrite_content(request):
@@ -21,8 +37,66 @@ def no_rewrite_content(request):
 
 def test_no_rewrite(no_rewrite_content):
     assert (
-        CssRewriter(ArticleUrlRewriter("kiwix.org", set())).rewrite(no_rewrite_content)
-        == no_rewrite_content.decode()
+        CssRewriter(ArticleUrlRewriter(no_rewrite_content.article_url, set())).rewrite(
+            no_rewrite_content.input_bytes
+        )
+        == no_rewrite_content.expected_bytes.decode()
+    )
+
+
+@pytest.fixture(
+    params=[
+        ContentForTests('"border:'),
+        ContentForTests("border: solid 1px #c0c0c0; width= 100%"),
+        # Despite being invalid, tinycss parse it as "width" property without value.
+        ContentForTests("width:", "width:;"),
+        ContentForTests("border-bottom-width: 1px;border-bottom-color: #c0c0c0;w"),
+        ContentForTests(
+            'background: url("http://exemple.com/foo.png"); width=',
+            'background: url("exemple.com/foo.png"); width=',
+        ),
+    ]
+)
+def invalid_content_inline(request):
+    yield request.param
+
+
+def test_invalid_css_inline(invalid_content_inline):
+    assert (
+        CssRewriter(
+            ArticleUrlRewriter(invalid_content_inline.article_url, set())
+        ).rewrite_inline(invalid_content_inline.input_str)
+        == invalid_content_inline.expected_str
+    )
+
+
+@pytest.fixture(
+    params=[
+        # Tinycss parse `"border:}` as a string with an unexpected eof in string.
+        # At serialization, tiny try to recover and close the opened rule
+        ContentForTests(b'p {"border:}', b'p {"border:}}'),
+        ContentForTests(b'"p {border:}'),
+        ContentForTests(b"p { border: solid 1px #c0c0c0; width= 100% }"),
+        ContentForTests(b"p { width: }"),
+        ContentForTests(
+            b"p { border-bottom-width: 1px;border-bottom-color: #c0c0c0;w }"
+        ),
+        ContentForTests(
+            b'p { background: url("http://exemple.com/foo.png"); width= }',
+            b'p { background: url("exemple.com/foo.png"); width= }',
+        ),
+    ]
+)
+def invalid_content(request):
+    yield request.param
+
+
+def test_invalid_cssl(invalid_content):
+    assert (
+        CssRewriter(ArticleUrlRewriter(invalid_content.article_url, set())).rewrite(
+            invalid_content.input_bytes
+        )
+        == invalid_content.expected_bytes.decode()
     )