fix: broken regex for <NuxtImage> components

Fixes #298
nuxt-modules · Aug 22, 2024 · 469e7bd · 469e7bd
1 parent 691409a
commit 469e7bd
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 2 deletions.
diff --git a/src/util/extractSitemapMetaFromHtml.ts b/src/util/extractSitemapMetaFromHtml.ts
@@ -1,5 +1,6 @@
 import { withSiteUrl } from 'nuxt-site-config-kit/urls'
 import { parseURL } from 'ufo'
+import { tryUseNuxt } from '@nuxt/kit'
 import type { ResolvedSitemapUrl, SitemapUrl, VideoEntry } from '../runtime/types'
 
 export function extractSitemapMetaFromHtml(html: string, options?: { images?: boolean, videos?: boolean, lastmod?: boolean, alternatives?: boolean }) {
@@ -12,7 +13,7 @@ export function extractSitemapMetaFromHtml(html: string, options?: { images?: bo
     if (mainMatch?.[1] && mainMatch[1].includes('<img')) {
       // Extract image src attributes using regex on the HTML, but ignore elements with invalid values such as data:, blob:, or file:
       // eslint-disable-next-line regexp/no-useless-lazy
-      const imgRegex = /<img\s+src=["']((?!data:|blob:|file:)[^"']+?)["'][^>]*>/gi
+      const imgRegex = /<img\s+(?:[^>]*?\s)?src=["']((?!data:|blob:|file:)[^"']+?)["'][^>]*>/gi
 
       let match
       while ((match = imgRegex.exec(mainMatch[1])) !== null) {
@@ -22,7 +23,7 @@ export function extractSitemapMetaFromHtml(html: string, options?: { images?: bo
         let url = match[1]
         // if the match is relative
         if (url.startsWith('/'))
-          url = withSiteUrl(url)
+          url = tryUseNuxt() ? withSiteUrl(url) : url
         images.add(url)
       }
     }

diff --git a/test/unit/extractSitemapMetaFromHtml.test.ts b/test/unit/extractSitemapMetaFromHtml.test.ts
@@ -82,6 +82,18 @@ describe('extractSitemapMetaFromHtml', () => {
         ],
       }
     `)
+
+    const html3 = `<div id="__nuxt"><div><main><div class="document-driven-page"><!--[--><div><h1 id="index"><!--[-->index<!--]--></h1><ul><!--[--><li><!--[--><a href="/bar" class=""><!--[-->/bar<!--]--></a><!--]--></li><li><!--[--><a href="/foo" class=""><!--[-->/foo<!--]--></a><!--]--></li><!--]--></ul><img onerror="this.setAttribute(&#39;data-error&#39;, 1)" alt="Test image" data-nuxt-img srcset="/_ipx/_/logo.svg 1x, /_ipx/_/logo.svg 2x" src="/_ipx/_/logo.svg" class="test"><p><!--[--><a href="/sitemap.xml" class=""><!--[-->/sitemap.xml<!--]--></a><!--]--></p></div><!--]--></div></main></div></div><div id="teleports"></div>`
+    const testcase3 = extractSitemapMetaFromHtml(html3)
+    expect(testcase3).toMatchInlineSnapshot(`
+      {
+        "images": [
+          {
+            "loc": "/_ipx/_/logo.svg",
+          },
+        ],
+      }
+    `)
   })
 
   it('extracts videos from HTML', async () => {