1 files changed, 41 insertions, 0 deletions
diff --git a/src/diagnostics/top_level_character_data.rs b/src/diagnostics/top_level_character_data.rs
new file mode 100644
index 0000000..5deb595
--- /dev/null
+++ b/src/diagnostics/top_level_character_data.rs
@@ -0,0 +1,41 @@
+use miette::Diagnostic;
+use thiserror::Error;
+
+use crate::{
+	diagnostics::AstDiagnostic,
+	parse::Node,
+	tokenize::{Span, TokenType},
+};
+
+#[derive(Error, Debug, Diagnostic)]
+#[error("Top-level character data is present")]
+#[diagnostic(code(TopLevelCharacterData), severity(Error))]
+pub struct TopLevelCharacterData {
+	#[label(collection)]
+	nodes: Vec<Span>,
+}
+
+impl AstDiagnostic for TopLevelCharacterData {
+	fn handle_document(document: &crate::parse::Document<'_>) -> Option<Self> {
+		let cdata_nodes = document
+			.children
+			.iter()
+			.filter_map(|node| {
+				if let Node::CharacterData(token) = node
+					&& let TokenType::Data(data) = token.ty
+					&& data.chars().any(|c| !c.is_whitespace())
+				{
+					Some(token.span)
+				} else {
+					None
+				}
+			})
+			.collect::<Vec<_>>();
+
+		if cdata_nodes.is_empty() {
+			return None;
+		}
+
+		Some(Self { nodes: cdata_nodes })
+	}
+}