@dukebot/astro-html-validator 1.1.1 → 1.1.3

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@dukebot/astro-html-validator",
-  "version": "1.1.1",
+  "version": "1.1.3",
   "description": "Validate Astro-generated HTML output for SEO metadata, JSON-LD, and internal links.",
   "type": "module",
   "main": "./src/index.mjs",
@@ -19,6 +19,8 @@
   ],
   "scripts": {
     "check": "node ./bin/cli.mjs --help",
+    "test": "node --test",
+    "test:links": "node --test tests/links-utils.test.mjs",
     "validate:dist": "node ./bin/cli.mjs"
   },
   "keywords": [
@@ -35,5 +37,8 @@
   },
   "publishConfig": {
     "access": "public"
+  },
+  "dependencies": {
+    "parse5": "^8.0.0"
   }
 }

package/src/utils/links.mjs

@@ -1,4 +1,5 @@
 import path from 'node:path';
+import { parse } from 'parse5';
 import { pathExists } from './common.mjs';
 
 /**
@@ -29,17 +30,46 @@ function toLocalPathFromAbsolute(rawUrl, absolutePrefixes) {
   return null;
 }
 
+/**
+ * Collects href/src attribute values from parsed HTML element nodes.
+ */
+function collectHtmlLinkAttributes(html = '') {
+  if (!html) return [];
+
+  const urls = [];
+  const document = parse(html, { sourceCodeLocationInfo: false });
+  const queue = [document];
+
+  while (queue.length > 0) {
+    const node = queue.shift();
+    if (!node) continue;
+
+    if (Array.isArray(node.attrs)) {
+      for (const attr of node.attrs) {
+        if (!attr?.name || !attr?.value) continue;
+        if (attr.name !== 'href' && attr.name !== 'src') continue;
+        urls.push(attr.value);
+      }
+    }
+
+    if (node.content) queue.push(node.content);
+    if (Array.isArray(node.childNodes) && node.childNodes.length > 0) {
+      queue.push(...node.childNodes);
+    }
+  }
+
+  return urls;
+}
+
 /**
  * Extracts local (root-relative) URLs from href/src attributes.
  */
 export function extractInternalUrls(html, { absoluteUrlPrefixes = [] } = {}) {
   const urls = new Set();
-  const regex = /(?:href|src)=["']([^"']+)["']/gi;
   const absolutePrefixes = normalizeAbsolutePrefixes(absoluteUrlPrefixes);
 
-  let match;
-  while ((match = regex.exec(html)) !== null) {
-    const raw = match[1]?.trim();
+  for (const value of collectHtmlLinkAttributes(html)) {
+    const raw = value?.trim();
     if (!raw) continue;
 
     if (
@@ -57,7 +87,7 @@ export function extractInternalUrls(html, { absoluteUrlPrefixes = [] } = {}) {
     if (!clean) continue;
 
     if (clean.startsWith('/')) {
-      if (clean) urls.add(clean);
+      urls.add(clean);
       continue;
     }