utils/html: Provide full control of allowed HTML elements

- Added new configuration option "strictly-allowed-html-elements" to specify only allowed HTML tags in the generated output. - Allowed "mark" and "u" elements for "highlight" and "underline" Markup extensions. - Updated "allowed-elements" in configuration files to include "tr". Fixes #751
isso-comments · May 26, 2024 · 863c4b7 · 863c4b7
1 parent 6f3874c
commit 863c4b7
Show file tree

Hide file tree

Showing 6 changed files with 69 additions and 20 deletions.
diff --git a/CHANGES.rst b/CHANGES.rst
@@ -50,6 +50,7 @@ Bugfixes & Improvements
 - Make 'text' field in 'comments' table NOT NULL and handling data migration (`#1019`_, pkvach)
 - Python 3.12 support (`#1015`_, ix5)
 - Disable Postbox submit button on click, enable after response (`#993`_, pkvach)
+- Provide full control of allowed HTML elements via the configuration file (`#1007`_, pkvach)
 
 .. _#951: https://github.com/posativ/isso/pull/951
 .. _#967: https://github.com/posativ/isso/pull/967
@@ -64,6 +65,7 @@ Bugfixes & Improvements
 .. _#1019: https://github.com/isso-comments/isso/pull/1019
 .. _#1015: https://github.com/isso-comments/isso/pull/1015
 .. _#993: https://github.com/isso-comments/isso/pull/993
+.. _#1007: https://github.com/isso-comments/isso/pull/1007
 
 0.13.1.dev0 (2023-02-05)
 ------------------------

diff --git a/contrib/isso-dev.cfg b/contrib/isso-dev.cfg
@@ -39,6 +39,7 @@ reply-to-self = true
 options = autolink, fenced-code, no-intra-emphasis, strikethrough, superscript
 flags =
 allowed-elements =
+strictly-allowed-html-elements =
 allowed-attributes =
 
 [hash]

diff --git a/docs/docs/reference/server-config.rst b/docs/docs/reference/server-config.rst
@@ -432,7 +432,7 @@ allowed-elements
 
     By default, only ``a``, ``blockquote``, ``br``, ``code``, ``del``, ``em``,
     ``h1``, ``h2``, ``h3``, ``h4``, ``h5``, ``h6``, ``hr``, ``ins``, ``li``,
-    ``ol``, ``p``, ``pre``, ``strong``, ``table``, ``tbody``, ``td``, ``th``,
+    ``ol``, ``p``, ``pre``, ``strong``, ``table``, ``tbody``, ``tr``, ``td``, ``th``,
     ``thead`` and ``ul`` are allowed.
 
     For a more detailed explanation, see :doc:`/docs/reference/markdown-config`.
@@ -444,11 +444,20 @@ allowed-elements
        mean that ``br, code, del, ...`` and all other default allowed tags are
        still allowed. You can only add *additional* elements here.
 
-       It is planned to change this behavior, see
-       `this issue <https://github.com/isso-comments/isso/issues/751>`_.
+       To specify a list of *only* allowed elements, use the
+       ``strictly-allowed-html-elements`` option instead.
 
     Default: (empty)
 
+strictly-allowed-html-elements
+
+    **Only** allow the specified HTML tags in the generated output, comma-separated.
+    If this option is set, the ``allowed-elements`` option is ignored.
+
+    Default: (empty)
+
+    .. versionadded:: 0.13.1
+
 allowed-attributes
     **Additional** HTML attributes (independent from elements) to allow in the
     generated output, comma-separated.

diff --git a/isso/isso.cfg b/isso/isso.cfg
@@ -213,9 +213,13 @@ flags =
 
 # Additional HTML tags to allow in the generated output, comma-separated. By
 # default, only a, blockquote, br, code, del, em, h1, h2, h3, h4, h5, h6, hr,
-# ins, li, ol, p, pre, strong, table, tbody, td, th, thead and ul are allowed.
+# ins, li, ol, p, pre, strong, table, tbody, tr, td, th, thead and ul are allowed.
 allowed-elements =
 
+# Only allow the specified HTML tags in the generated output, comma-separated.
+# If this option is set, the "allowed-elements" option is ignored.
+strictly-allowed-html-elements =
+
 # Additional HTML attributes (independent from elements) to allow in the
 # generated output, comma-separated. By default, only align and href are
 # allowed.

diff --git a/isso/tests/test_html.py b/isso/tests/test_html.py
@@ -60,7 +60,7 @@ def test_github_flavoured_markdown(self):
             </code></pre>""")
 
     def test_sanitizer(self):
-        sanitizer = html.Sanitizer(elements=[], attributes=[])
+        sanitizer = html.Sanitizer(elements=["p", "a", "code"], attributes=["href"])
         examples = [
             ('Look: <img src="..." />', 'Look: '),
             ('<a href="http://example.org/">Ha</a>',
@@ -94,23 +94,39 @@ def test_render(self):
             "markup": {
                 "options": "autolink",
                 "flags": "",
-                "allowed-elements": "",
-                "allowed-attributes": ""
+                "allowed-elements": "a, p",
+                "allowed-attributes": "href",
+                "strictly-allowed-html-elements": ""
             }
         })
         renderer = html.Markup(conf.section("markup")).render
         self.assertIn(renderer("http://example.org/ and sms:+1234567890"),
                       ['<p><a href="http://example.org/" rel="nofollow noopener">http://example.org/</a> and sms:+1234567890</p>',
                        '<p><a rel="nofollow noopener" href="http://example.org/">http://example.org/</a> and sms:+1234567890</p>'])
 
+    def test_render_with_strictly_allowed_elements(self):
+        conf = config.new({
+            "markup": {
+                "options": "autolink",
+                "flags": "",
+                "allowed-elements": "a, p",
+                "strictly-allowed-html-elements": "p",
+                "allowed-attributes": "href"
+            }
+        })
+        renderer = html.Markup(conf.section("markup")).render
+        self.assertEqual(renderer("http://example.org/ and sms:+1234567890"),
+                         '<p>http://example.org/ and sms:+1234567890</p>')
+
     def test_sanitized_render_extensions(self):
         """Options should be normalized from both dashed-case or snake_case (legacy)"""
         conf = config.new({
             "markup": {
                 "options": "no_intra_emphasis",  # Deliberately snake_case
                 "flags": "",
-                "allowed-elements": "",
-                "allowed-attributes": ""
+                "allowed-elements": "p",
+                "allowed-attributes": "",
+                "strictly-allowed-html-elements": ""
             }
         })
         renderer = html.Markup(conf.section("markup")).render

diff --git a/isso/utils/html.py b/isso/utils/html.py
@@ -17,16 +17,7 @@ def allow_attribute_class(tag, name, value):
         return name == "class" and bool(Sanitizer.code_language_pattern.match(value))
 
     def __init__(self, elements, attributes):
-        # attributes found in Sundown's HTML serializer [1]
-        # - except for <img> tag, because images are not generated anyways.
-        # - sub and sup added
-        #
-        # [1] https://github.com/vmg/sundown/blob/master/html/html.c
-        self.elements = ["a", "p", "hr", "br", "ol", "ul", "li",
-                         "pre", "code", "blockquote",
-                         "del", "ins", "strong", "em",
-                         "h1", "h2", "h3", "h4", "h5", "h6", "sub", "sup",
-                         "table", "thead", "tbody", "th", "td"] + elements
+        self.elements = elements
 
         # allowed attributes for tags
         self.attributes = {
@@ -108,12 +99,38 @@ def __init__(self, conf):
         parser = Markdown(extensions=self.extensions,
                           flags=self.flags)
         # Filter out empty strings:
-        allowed_elements = [x for x in conf.getlist("allowed-elements") if x]
+        strictly_allowed_html_elements = [x for x in conf.getlist("strictly-allowed-html-elements") if x]
         allowed_attributes = [x for x in conf.getlist("allowed-attributes") if x]
 
+        # if "strictly-allowed-html-elements" option is set, use it instead of "allowed-elements"
+        if strictly_allowed_html_elements:
+            allowed_elements = strictly_allowed_html_elements
+        else:
+            allowed_elements = [x for x in conf.getlist("allowed-elements") if x]
+
+            # attributes found in Sundown's HTML serializer [1]
+            # - except for <img> tag, because images are not generated anyways.
+            # - sub and sup added
+            #
+            # [1] https://github.com/vmg/sundown/blob/master/html/html.c
+            allowed_elements = ["a", "p", "hr", "br", "ol", "ul", "li",
+                                "pre", "code", "blockquote",
+                                "del", "ins", "strong", "em",
+                                "h1", "h2", "h3", "h4", "h5", "h6", "sub", "sup",
+                                "table", "thead", "tbody", "tr", "th", "td"] + allowed_elements
+
         # If images are allowed, source element should be allowed as well
         if 'img' in allowed_elements and 'src' not in allowed_attributes:
             allowed_attributes.append('src')
+
+        # If 'highlight' extension is enabled, allow 'mark' element
+        if 'highlight' in self.extensions and 'mark' not in allowed_elements:
+            allowed_elements.append('mark')
+
+        # If 'underline' extension is enabled, allow 'u' element
+        if 'underline' in self.extensions and 'u' not in allowed_elements:
+            allowed_elements.append('u')
+
         sanitizer = Sanitizer(allowed_elements, allowed_attributes)
 
         self._render = lambda text: sanitizer.sanitize(parser(text))