@electron/lint-roller 1.12.0 → 1.13.0

package/markdownlint-rules/emd002.mjs

@@ -0,0 +1,47 @@
+import { addError, filterTokens } from 'markdownlint/helpers';
+
+import { fromMarkdown } from 'mdast-util-from-markdown';
+import { visit } from 'unist-util-visit';
+
+export const names = ['EMD002', 'no-angle-brackets'];
+export const description = 'No unescaped opening angle brackets in text (does not play nice with MDX)';
+export const tags = ['brackets'];
+
+const UNESCAPED_REGEX = /(?<!\\)<(?!\s)/g;
+
+function EMD002(params, onError) {
+  filterTokens(params, 'inline', (token) => {
+    for (const childToken of token.children) {
+      // childToken.line has the raw content, but may also contain
+      // more content than just childToken.content. Since we need
+      // the raw content to detect escaped angle brackets, parse
+      // the raw content of any match a second time to avoid any
+      // false positives.
+      if (
+        childToken.type === 'text' &&
+        childToken.markup !== '&lt;' &&
+        childToken.markup !== '&#60;' &&
+        childToken.content.includes('<') &&
+        childToken.line.match(UNESCAPED_REGEX) !== null
+      ) {
+        // The AST produced by mdast is easier to work with here
+        // since it gives position data within the line
+        const tree = fromMarkdown(childToken.line);
+
+        visit(
+          tree,
+          (node) => node.type === 'text',
+          (node) => {
+            const rawContent = childToken.line.slice(node.position.start.offset, node.position.end.offset);
+
+            if (rawContent.match(UNESCAPED_REGEX) !== null) {
+              addError(onError, token.lineNumber, 'Unescaped opening angle bracket');
+            }
+          },
+        );
+      }
+    }
+  });
+};
+
+export { EMD002 as function };