Merge pull request #270 from openzim/onxxx_events

Add support for onxxx HTML events
openzim · May 24, 2024 · a98e6cf · a98e6cf
2 parents ab2f17d + 51a50a6
commit a98e6cf
Show file tree

Hide file tree

Showing 5 changed files with 128 additions and 8 deletions.
diff --git a/README.md b/README.md
@@ -56,6 +56,8 @@ Scenario which are known to work well:
 - Redirections with `meta http-equiv` are not yet supported (see https://github.com/openzim/warc2zim/issues/237)
 - Web workers are not yet supported (see https://github.com/openzim/warc2zim/issues/272)
 - Service workers are not supported and will most probably never be
+- Inline JS code inside an onxxx HTML event (e.g. onclick, onhover, ...) is rewritten, so for instance redirection to another handled with these events is working
+  - However since URL rewriting is performed with dynamic JS rewriting, at this stage scraper has no clue on what is inside the ZIM and what is external ; all URLs are hence supposed to be internal, which might break some dynamic redirection to an online website
 
 It is also important to note that warc2zim is inherently limited to what is present inside the WARC. A bad WARC can only produce a bad ZIM. Garbage in, garbage out.
 

diff --git a/src/warc2zim/content_rewriting/html.py b/src/warc2zim/content_rewriting/html.py
@@ -67,6 +67,12 @@ def rewrite(self, content: str) -> RewritenHtml:
             self.url_rewriter, base_href=self.base_href, rewrite_all_url=False
         )
         self.css_rewriter = CssRewriter(self.url_rewriter, self.base_href)
+        self.js_rewriter = JsRewriter(
+            url_rewriter=self.url_rewriter,
+            base_href=self.base_href,
+            extra_rules=get_ds_rules(self.url_rewriter.article_url.value),
+            notify_js_module=self.notify_js_module,
+        )
 
         self.feed(content)
         self.close()
@@ -149,12 +155,7 @@ def handle_data(self, data: str):
                     data = RxRewriter(rules).rewrite(data, {})
         elif self.html_rewrite_context and self.html_rewrite_context.startswith("js-"):
             if data.strip():
-                data = JsRewriter(
-                    url_rewriter=self.url_rewriter,
-                    base_href=self.base_href,
-                    extra_rules=get_ds_rules(self.url_rewriter.article_url.value),
-                    notify_js_module=self.notify_js_module,
-                ).rewrite(
+                data = self.js_rewriter.rewrite(
                     data,
                     opts={"isModule": self.html_rewrite_context == "js-module"},
                 )
@@ -206,6 +207,8 @@ def process_attr(
             return (attr_name, ", ".join(new_value_list))
         if attr_name == "style":
             return (attr_name, self.css_rewriter.rewrite_inline(attr_value))
+        if attr_name.startswith("on") and not attr_name.startswith("on-"):
+            return (attr_name, self.js_rewriter.rewrite(attr_value))
         return (attr_name, attr_value)
 
     def format_attr(self, name: str, value: str | None) -> str:

diff --git a/test-website/content/index.html b/test-website/content/index.html
@@ -45,6 +45,7 @@
         <li><a href="./j%C3%A0v%C3%A0scr%C3%AFpt.html">Javascript with special chars</a></li>
         <li><a href="./http-return-codes.html">HTTP return codes</a></li>
         <li><a href="./base-href.html">Base href</a></li>
+        <li><a href="./onxxx.html">onxxx HTML events</a></li>
     </ul>
 </body>
 

diff --git a/test-website/content/onxxx.html b/test-website/content/onxxx.html
@@ -0,0 +1,38 @@
+<!DOCTYPE html>
+<html lang="en">
+
+<head>
+  <meta charset="utf-8">
+  <title>Test website</title>
+  <link rel="apple-touch-icon" sizes="180x180" href="./icons/apple-touch-icon.png">
+  <link rel="icon" type="image/png" sizes="32x32" href="./icons/favicon-32x32.png">
+  <link rel="icon" type="image/png" sizes="16x16" href="./icons/favicon-16x16.png">
+  <link rel="manifest" href="./icons/site.webmanifest">
+  <link rel="shortcut icon" href="./icons/favicon.ico">
+  <style>
+    .status {
+        color: red;
+    }
+    .status.green {
+        color: green;
+    }
+</style>
+</head>
+
+<body>
+
+  <h2>onxxx HTML events</h2>
+
+  <p>When clicking images below, it should redirect your browser to the another page</p>
+
+  <p>Relative (<a href="./index.html">this site home page</a>): <img src="images/image1.png" onclick="document.location.href='./index.html';"> </p>
+
+  <p>Absolute internal 1 (<a href="https://standard_netloc/index.html">this site home page</a>): <img src="images/image1.png" onclick="document.location.href='https://standard_netloc/index.html';"></p>
+
+  <p>Absolute internal 2 (<a href="https://not_standard_netloc_punny_encoded/index.html">another site home page</a>): <img src="images/image1.png" onclick="document.location.href='https://not_standard_netloc_punny_encoded/index.html';"></p>
+
+  <p>Absolute external (kiwix homepage): <img src="images/image1.png" onclick="document.location.href='https://www.kiwix.org';"></p>
+
+</body>
+
+</html>
diff --git a/tests/test_html_rewriting.py b/tests/test_html_rewriting.py
@@ -229,9 +229,11 @@ def test_extract_title(no_js_notify):
     </html>"""
 
     assert (
-        # Nota: lambda below is a trick, we should assign an ArticleUrlRewriter
         HtmlRewriter(
-            lambda _: "kiwix.org",  # pyright: ignore[reportGeneralTypeIssues, reportArgumentType]
+            ArticleUrlRewriter(
+                HttpUrl("http://example.com"),
+                {ZimPath("exemple.com/a/long/path")},
+            ),
             "",
             "",
             no_js_notify,
@@ -667,3 +669,77 @@ def test_simple_rewrite(input_content, expected_output, no_js_notify):
         .content
         == expected_output
     )
+
+
+@pytest.fixture(
+    params=[
+        ContentForTests(
+            """<img onclick="">""",
+        ),
+        ContentForTests(
+            """<img on-whatever="foo">""",
+        ),
+        ContentForTests(
+            """<img on="foo">""",
+        ),
+        ContentForTests(
+            """<img to="foo">""",
+        ),
+        ContentForTests(
+            """<img onclick="document.location.href='./index.html';">""",
+            (
+                """<img onclick="var _____WB$wombat$assign$function_____ = """
+                "function(name) {return (self._wb_wombat &amp;&amp; "
+                "self._wb_wombat.local_init &amp;&amp; "
+                "self._wb_wombat.local_init(name)) || self[name]; };\n"
+                "if (!self.__WB_pmw) { self.__WB_pmw = function(obj) "
+                "{ this.__WB_source = obj; return this; } }\n"
+                "{\n"
+                "let window = _____WB$wombat$assign$function_____(&quot;window&quot;);"
+                "\n"
+                "let globalThis = _____WB$wombat$assign$function_____"
+                "(&quot;globalThis&quot;);\n"
+                "let self = _____WB$wombat$assign$function_____(&quot;self&quot;);\n"
+                "let document = "
+                "_____WB$wombat$assign$function_____(&quot;document&quot;);\n"
+                "let location = "
+                "_____WB$wombat$assign$function_____(&quot;location&quot;);\n"
+                "let top = _____WB$wombat$assign$function_____(&quot;top&quot;);\n"
+                "let parent = "
+                "_____WB$wombat$assign$function_____(&quot;parent&quot;);\n"
+                "let frames = "
+                "_____WB$wombat$assign$function_____(&quot;frames&quot;);\n"
+                "let opener = "
+                "_____WB$wombat$assign$function_____(&quot;opener&quot;);\n"
+                "let arguments;\n\n"
+                "document.location.href=&#x27;./index.html&#x27;;\n"
+                """}">"""
+            ),  # NOTA: quotes and ampersand are escaped since we are inside HTML attr
+        ),
+    ]
+)
+def rewrite_onxxx_content(request):
+    yield request.param
+
+
+def test_rewrite_onxxx_event(rewrite_onxxx_content, no_js_notify):
+    assert (
+        HtmlRewriter(
+            ArticleUrlRewriter(
+                HttpUrl(f"http://{rewrite_onxxx_content.article_url}"),
+                {
+                    ZimPath("kiwix.org/foo.html"),
+                    ZimPath("kiwix.org/foo.js"),
+                    ZimPath("kiwix.org/foo.css"),
+                    ZimPath("kiwix.org/foo.png"),
+                    ZimPath("kiwix.org/favicon.png"),
+                },
+            ),
+            "",
+            "",
+            no_js_notify,
+        )
+        .rewrite(rewrite_onxxx_content.input_str)
+        .content
+        == rewrite_onxxx_content.expected_str
+    )