@adobe/helix-html-pipeline 6.30.0 → 6.31.0

package/CHANGELOG.md

@@ -1,3 +1,17 @@
+# [6.31.0](https://github.com/adobe/helix-html-pipeline/compare/v6.30.1...v6.31.0) (2026-08-17)
+
+
+### Features
+
+* update og:image from <a> also along with img ([9768a57](https://github.com/adobe/helix-html-pipeline/commit/9768a57e5d96aef18af081e4f9aec67ba9a4d586))
+
+## [6.30.1](https://github.com/adobe/helix-html-pipeline/compare/v6.30.0...v6.30.1) (2026-08-17)
+
+
+### Bug Fixes
+
+* **sitemap-index:** static version in code, dynamic version in content ([#1125](https://github.com/adobe/helix-html-pipeline/issues/1125)) ([5c76a06](https://github.com/adobe/helix-html-pipeline/commit/5c76a060b3ff5a1c198c565f256b3ee7a5bbbc0b))
+
 # [6.30.0](https://github.com/adobe/helix-html-pipeline/compare/v6.29.12...v6.30.0) (2026-08-14)
 
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@adobe/helix-html-pipeline",
-  "version": "6.30.0",
+  "version": "6.31.0",
   "description": "Helix HTML Pipeline",
   "main": "src/index.js",
   "types": "src/index.d.ts",

package/src/sitemap-index-pipe.js

@@ -77,10 +77,12 @@ export async function sitemapIndexPipe(state, req) {
   try {
     await initConfig(state, req, res);
 
-    // fetch sitemap.xml
+    // fetch sitemap-index.xml
     state.timer?.update('content-fetch');
+    state.content.sourceBus = 'code';
     await fetchContent(state, req, res);
     if (res.status === 404) {
+      state.content.sourceBus = 'content';
       const ret = await generateSitemapIndex(state);
       if (ret.status === 200) {
         res.status = 200;

package/src/steps/extract-metadata.js

@@ -19,6 +19,48 @@ import {
 import { toMetaName } from '../utils/modifiers.js';
 import { childNodes } from '../utils/hast-utils.js';
 
+// common web image extensions
+const IMAGE_EXTENSIONS = new Set(['avif', 'webp', 'png', 'jpg', 'jpeg', 'gif']);
+
+/**
+ * Checks whether an anchor points directly to an image, based on its href ending in a
+ * recognizable image extension.
+ * @param {Element} $a The anchor element
+ * @returns {boolean} true if the anchor should be treated as an image link
+ */
+function isImageAnchor($a) {
+  try {
+    // parse via URL (with a dummy base for relative hrefs) so the extension check is based
+    // on the pathname alone, ignoring any query string or fragment
+    const { pathname } = new URL($a.properties.href, 'https://localhost/');
+    const ext = pathname.slice(pathname.lastIndexOf('.') + 1).toLowerCase();
+    return IMAGE_EXTENSIONS.has(ext);
+  } catch {
+    // href isn't a URL that can be parsed (e.g. mailto:, malformed)
+    return false;
+  }
+}
+
+/**
+ * Checks whether two URL-ish strings refer to the same resource, e.g. so that a link
+ * text that only differs from its href by encoding (say, a literal space vs. "%20") isn't
+ * mistaken for a real caption. Falls back to plain string equality if either side can't be
+ * parsed as a URL - a genuine caption isn't expected to parse as one anyway.
+ * @param {string} a first URL-ish string
+ * @param {string} b second URL-ish string
+ * @returns {boolean} true if a and b are the same URL, ignoring encoding differences
+ */
+function isSameUrl(a, b) {
+  if (a === b) {
+    return true;
+  }
+  try {
+    return new URL(a, 'https://localhost/').href === new URL(b, 'https://localhost/').href;
+  } catch {
+    return false;
+  }
+}
+
 /**
  * Cleans up comma-separated string lists and returns an array.
  * @param {string} list A comma-separated list
@@ -262,11 +304,32 @@ export default function extractMetaData(state, req) {
 
   // content.image is not correct if the first image is in a page-block. since the pipeline
   // only respects the image nodes in the mdast
-  const $hero = select('div img', hast);
+  //
+  // Approach B (https://www.aem.live/docs/media#approach-b-asset-management-delivery)
+  // images are authored as <a> links, not <img>, so they never show up as image nodes -
+  // match the anchor too, via isImageAnchor, which recognizes any href pointing directly
+  // at an image file by its extension, regardless of the asset source. Gathered in document
+  // order (rather than trying img first, falling back to the anchor only if no img exists at
+  // all) so whichever appears first in the page wins - otherwise an unrelated <img> elsewhere
+  // on the page (e.g. a card thumbnail below the actual hero anchor) would incorrectly take
+  // priority over the real hero.
+  const $hero = selectAll('div img, div a[href]', hast)
+    .find(($el) => $el.tagName === 'img' || isImageAnchor($el));
   if ($hero) {
-    content.image = $hero.properties.src;
-    if ($hero.properties.alt) {
-      content.imageAlt = $hero.properties.alt;
+    if ($hero.tagName === 'a') {
+      content.image = $hero.properties.href;
+      // prefer the anchor's title attribute as alt text; fall back to the link text, unless
+      // it's empty or authors pasted the raw URL as the caption (see the "no caption" test) -
+      // neither is meaningful to read out, so leave it unset rather than exposing the raw URL
+      // to assistive tech and social crawlers.
+      const text = toString($hero).trim();
+      content.imageAlt = $hero.properties.title
+        || (text && !isSameUrl(text, $hero.properties.href) ? text : undefined);
+    } else {
+      content.image = $hero.properties.src;
+      if ($hero.properties.alt) {
+        content.imageAlt = $hero.properties.alt;
+      }
     }
   }