fix: ignore invalid values on image discovery (#290)

nuxt-modules · Jun 27, 2024 · edd0c88 · edd0c88
1 parent 477f7c8
commit edd0c88
Show file tree

Hide file tree

Showing 3 changed files with 22 additions and 2 deletions.
diff --git a/.playground/nuxt.config.ts b/.playground/nuxt.config.ts
@@ -60,6 +60,7 @@ export default defineNuxtConfig({
     prerender: {
       routes: [
         // '/sitemap_index.xml',
+        '/prerender',
         '/should-be-in-sitemap',
         '/foo.bar/',
         '/test.doc',

diff --git a/.playground/pages/prerender.vue b/.playground/pages/prerender.vue
diff --git a/src/util/extractSitemapMetaFromHtml.ts b/src/util/extractSitemapMetaFromHtml.ts
@@ -10,8 +10,9 @@ export function extractSitemapMetaFromHtml(html: string, options?: { images?: bo
     const mainRegex = /<main[^>]*>([\s\S]*?)<\/main>/
     const mainMatch = mainRegex.exec(html)
     if (mainMatch?.[1] && mainMatch[1].includes('<img')) {
-      // extract image src using regex on the html
-      const imgRegex = /<img[^>]+src="([^">]+)"/g
+      // Extract image src attributes using regex on the HTML, but ignore elements with invalid values such as data:, blob:, or file:
+      const imgRegex = /<img\s+src=["']((?!data:|blob:|file:)[^"']+?)["'][^>]*>/gi
+
       let match
       // eslint-disable-next-line no-cond-assign
       while ((match = imgRegex.exec(mainMatch[1])) !== null) {