Merge commit 'c58967c3bd7b7eee0b284ee2b77a48ed9ce6863e' into glitch-soc/merge-upstream

Conflicts:
- `lib/sanitize_ext/sanitize_config.rb`:
  Conflict due to upstream adding fallback support for MathML Core while
  glitch-soc supported slightly more tags than upstream.
  Added the MathML Core fallback code to glitch-soc.
- `spec/lib/sanitize/config_spec.rb`:
  Ditto.

lib/sanitize_ext/sanitize_config.rb

@@ -74,6 +74,44 @@ class Sanitize
       current_node.replace(current_node.document.create_text_node(current_node.text)) unless LINK_PROTOCOLS.include?(scheme)
     end
 
+    # We assume that incomming <math> nodes are of the form
+    # <math><semantics>...<annotation>...</annotation></semantics></math>
+    # according to the [FEP]. We try to grab the most relevant plain-text
+    # annotation from the semantics node, and use it to display a representation
+    # of the mathematics.
+    #
+    # FEP: https://codeberg.org/fediverse/fep/src/branch/main/fep/dc88/fep-dc88.md
+    MATH_TRANSFORMER = lambda do |env|
+      math = env[:node]
+      return if env[:is_allowlisted]
+      return unless math.element? && env[:node_name] == 'math'
+
+      semantics = math.element_children[0]
+      return if semantics.nil? || semantics.name != 'semantics'
+
+      # next, we find the plain-text description
+      is_annotation_with_encoding = lambda do |encoding, node|
+        return false unless node.name == 'annotation'
+
+        node.attributes['encoding'].value == encoding
+      end
+
+      annotation = semantics.children.find(&is_annotation_with_encoding.curry['application/x-tex'])
+      if annotation
+        text = if math.attributes['display']&.value == 'block'
+                 "$$#{annotation.text}$$"
+               else
+                 "$#{annotation.text}$"
+               end
+        math.replace(math.document.create_text_node(text))
+        return
+      end
+      # Don't bother surrounding 'text/plain' annotations with dollar signs,
+      # since it isn't LaTeX
+      annotation = semantics.children.find(&is_annotation_with_encoding.curry['text/plain'])
+      math.replace(math.document.create_text_node(annotation.text)) unless annotation.nil?
+    end
+
     MASTODON_STRICT = freeze_config(
       elements: %w(p br span a abbr del s pre blockquote code b strong u sub sup i em h1 h2 h3 h4 h5 ul ol li ruby rt rp),
 
@@ -102,6 +140,7 @@ class Sanitize
         ALLOWED_CLASS_TRANSFORMER,
         IMG_TAG_TRANSFORMER,
         TRANSLATE_TRANSFORMER,
+        MATH_TRANSFORMER,
         UNSUPPORTED_HREF_TRANSFORMER,
       ]
     )