Add Raw Node

kivikakk · Dec 17, 2024 · 04eedf6 · 04eedf6
1 parent 8d1e90c
commit 04eedf6
Show file tree

Hide file tree

Showing 7 changed files with 105 additions and 2 deletions.
diff --git a/src/cm.rs b/src/cm.rs
@@ -366,6 +366,7 @@ impl<'a, 'o, 'c> CommonMarkFormatter<'a, 'o, 'c> {
             NodeValue::HtmlInline(ref literal) => {
                 self.format_html_inline(literal.as_bytes(), entering)
             }
+            NodeValue::Raw(ref literal) => self.format_raw(literal.as_bytes(), entering),
             NodeValue::Strong => {
                 if parent_node.is_none()
                     || !matches!(parent_node.unwrap().data.borrow().value, NodeValue::Strong)
@@ -686,6 +687,12 @@ impl<'a, 'o, 'c> CommonMarkFormatter<'a, 'o, 'c> {
         }
     }
 
+    fn format_raw(&mut self, literal: &[u8], entering: bool) {
+        if entering {
+            self.write_all(literal).unwrap();
+        }
+    }
+
     fn format_strong(&mut self) {
         write!(self, "**").unwrap();
     }

diff --git a/src/html.rs b/src/html.rs
@@ -380,6 +380,9 @@ where
                             | NodeValue::HtmlInline(ref literal) => {
                                 self.escape(literal.as_bytes())?;
                             }
+                            NodeValue::Raw(ref literal) => {
+                                self.output.write_all(literal.as_bytes())?;
+                            }
                             NodeValue::LineBreak | NodeValue::SoftBreak => {
                                 self.output.write_all(b" ")?;
                             }
@@ -757,6 +760,12 @@ where
                     }
                 }
             }
+            NodeValue::Raw(ref literal) => {
+                // No sourcepos.
+                if entering {
+                    self.output.write_all(literal.as_bytes())?;
+                }
+            }
             NodeValue::Strong => {
                 // Unreliable sourcepos.
                 let parent_node = node.parent();

diff --git a/src/nodes.rs b/src/nodes.rs
@@ -127,6 +127,9 @@ pub enum NodeValue {
     /// **Inline**.  [Raw HTML](https://github.github.com/gfm/#raw-html) contained inline.
     HtmlInline(String),
 
+    /// **Inline**.  A Raw text.
+    Raw(String),
+
     /// **Inline**.  [Emphasized](https://github.github.com/gfm/#emphasis-and-strong-emphasis)
     /// text.
     Emph,
@@ -511,6 +514,7 @@ impl NodeValue {
             NodeValue::Strong => "strong",
             NodeValue::Code(..) => "code",
             NodeValue::HtmlInline(..) => "html_inline",
+            NodeValue::Raw(..) => "raw",
             NodeValue::Strikethrough => "strikethrough",
             NodeValue::FrontMatter(_) => "frontmatter",
             NodeValue::TaskItem { .. } => "taskitem",

diff --git a/src/tests.rs b/src/tests.rs
@@ -21,6 +21,7 @@ mod multiline_block_quotes;
 mod options;
 mod pathological;
 mod plugins;
+mod raw;
 mod regressions;
 mod rewriter;
 mod shortcodes;

diff --git a/src/tests/api.rs b/src/tests/api.rs
@@ -239,6 +239,9 @@ fn exercise_full_api() {
         nodes::NodeValue::HtmlInline(html) => {
             let _: &String = html;
         }
+        nodes::NodeValue::Raw(text) => {
+            let _: &String = text;
+        }
         nodes::NodeValue::Emph => {}
         nodes::NodeValue::Strong => {}
         nodes::NodeValue::Strikethrough => {}

diff --git a/src/tests/raw.rs b/src/tests/raw.rs
@@ -0,0 +1,77 @@
+use std::cell::RefCell;
+
+use crate::{
+    arena_tree::Node,
+    nodes::{Ast, NodeValue},
+};
+
+use super::*;
+
+#[test]
+fn raw_node() {
+    let user_input = "User input: <iframe></iframe>";
+    let system_input_inline = "System Inline input: <iframe></iframe>";
+    let system_input_block = "System Block input: <iframe></iframe>";
+    let input = user_input.to_owned() + system_input_inline + "\n\n" + system_input_block + "\n";
+
+    let mut options = Options::default();
+    options.render.escape = true;
+    options.render.unsafe_ = false;
+    options.extension.tagfilter = true;
+
+    let arena = Arena::new();
+    let root = parse_document(&arena, user_input, &options);
+    let raw_ast_inline = Ast::new(
+        NodeValue::Raw(system_input_inline.to_string()),
+        (0, 0).into(),
+    );
+    let raw_node_inline = arena.alloc(Node::new(RefCell::new(raw_ast_inline)));
+    root.first_child()
+        .unwrap()
+        .last_child()
+        .unwrap()
+        .insert_after(raw_node_inline);
+    let raw_ast_block = Ast::new(
+        NodeValue::Raw(system_input_block.to_string()),
+        (0, 0).into(),
+    );
+    let raw_node_block = arena.alloc(Node::new(RefCell::new(raw_ast_block)));
+    root.first_child().unwrap().insert_after(raw_node_block);
+
+    let mut output = vec![];
+    html::format_document(root, &options, &mut output).unwrap();
+    compare_strs(
+        &String::from_utf8(output).unwrap(),
+        concat!(
+            "<p>User input: &lt;iframe&gt;&lt;/iframe&gt;",
+            "System Inline input: <iframe></iframe></p>\n",
+            "System Block input: <iframe></iframe>"
+        ),
+        "html",
+        &input,
+    );
+
+    let mut md = vec![];
+    cm::format_document_with_plugins(root, &options, &mut md, &Plugins::default()).unwrap();
+    compare_strs(&String::from_utf8(md).unwrap(), &input, "cm", &input);
+
+    let mut xml = vec![];
+    crate::xml::format_document(root, &options, &mut xml).unwrap();
+    compare_strs(
+        &String::from_utf8(xml).unwrap(),
+        concat!(
+            "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n",
+	    "<!DOCTYPE document SYSTEM \"CommonMark.dtd\">\n",
+	    "<document xmlns=\"http://commonmark.org/xml/1.0\">\n",
+	    "  <paragraph>\n",
+	    "    <text xml:space=\"preserve\">User input: </text>\n",
+	    "    <html_inline xml:space=\"preserve\">&lt;iframe&gt;</html_inline>\n",
+	    "    <html_inline xml:space=\"preserve\">&lt;/iframe&gt;</html_inline>\n",
+	    "    <raw xml:space=\"preserve\">System Inline input: &lt;iframe&gt;&lt;/iframe&gt;</raw>\n",
+	    "  </paragraph>\n",
+	    "  <raw xml:space=\"preserve\">System Block input: &lt;iframe&gt;&lt;/iframe&gt;</raw>\n</document>\n"
+        ),
+        "xml",
+        &input,
+    );
+}
diff --git a/src/xml.rs b/src/xml.rs
@@ -89,7 +89,8 @@ impl<'o, 'c> XmlFormatter<'o, 'c> {
                         match node.data.borrow().value {
                             NodeValue::Text(ref literal)
                             | NodeValue::Code(NodeCode { ref literal, .. })
-                            | NodeValue::HtmlInline(ref literal) => {
+                            | NodeValue::HtmlInline(ref literal)
+                            | NodeValue::Raw(ref literal) => {
                                 self.escape(literal.as_bytes())?;
                             }
                             NodeValue::LineBreak | NodeValue::SoftBreak => {
@@ -148,7 +149,8 @@ impl<'o, 'c> XmlFormatter<'o, 'c> {
                 NodeValue::Text(ref literal)
                 | NodeValue::Code(NodeCode { ref literal, .. })
                 | NodeValue::HtmlBlock(NodeHtmlBlock { ref literal, .. })
-                | NodeValue::HtmlInline(ref literal) => {
+                | NodeValue::HtmlInline(ref literal)
+                | NodeValue::Raw(ref literal) => {
                     self.output.write_all(b" xml:space=\"preserve\">")?;
                     self.escape(literal.as_bytes())?;
                     write!(self.output, "</{}", ast.value.xml_node_name())?;