@adobe/helix-html-pipeline 4.1.5 → 5.0.1

package/CHANGELOG.md

@@ -1,3 +1,22 @@
+## [5.0.1](https://github.com/adobe/helix-html-pipeline/compare/v5.0.0...v5.0.1) (2023-09-09)
+
+
+### Bug Fixes
+
+* properly handle empty string ([#395](https://github.com/adobe/helix-html-pipeline/issues/395)) ([daac492](https://github.com/adobe/helix-html-pipeline/commit/daac49224642eeb69951ee68256fcb7f264f4f0c))
+
+# [5.0.0](https://github.com/adobe/helix-html-pipeline/compare/v4.1.5...v5.0.0) (2023-09-09)
+
+
+### Features
+
+* better metadata fallback ([2c7481c](https://github.com/adobe/helix-html-pipeline/commit/2c7481c19c52cce38559a50360c8c5eabc201b27)), closes [#382](https://github.com/adobe/helix-html-pipeline/issues/382)
+
+
+### BREAKING CHANGES
+
+* metadata fallback is different for empty values.
+
 ## [4.1.5](https://github.com/adobe/helix-html-pipeline/compare/v4.1.4...v4.1.5) (2023-09-02)
 
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@adobe/helix-html-pipeline",
-  "version": "4.1.5",
+  "version": "5.0.1",
   "description": "Helix HTML Pipeline",
   "main": "src/index.js",
   "types": "src/index.d.ts",
@@ -29,7 +29,10 @@
   },
   "homepage": "https://github.com/adobe/helix-html-pipeline#readme",
   "mocha": {
-    "require": "test/setup-env.js",
+    "require": [
+      "test/setup-env.js",
+      "mocha-suppress-logs"
+    ],
     "recursive": "true",
     "reporter": "mocha-multi-reporters",
     "reporter-options": "configFile=.mocha-multi.json",
@@ -50,7 +53,7 @@
     "hast-util-to-html": "8.0.4",
     "hast-util-to-string": "2.0.0",
     "hastscript": "7.2.0",
-    "jose": "4.14.5",
+    "jose": "4.14.6",
     "mdast-util-to-hast": "12.3.0",
     "mdast-util-to-string": "4.0.0",
     "mime": "3.0.0",
@@ -73,18 +76,19 @@
     "@semantic-release/git": "10.0.1",
     "@semantic-release/npm": "10.0.5",
     "c8": "8.0.1",
-    "eslint": "8.48.0",
+    "eslint": "8.49.0",
     "eslint-import-resolver-exports": "1.0.0-beta.5",
     "eslint-plugin-header": "3.1.1",
     "eslint-plugin-import": "2.28.1",
-    "esmock": "2.3.8",
+    "esmock": "2.4.1",
     "husky": "8.0.3",
     "js-yaml": "4.1.0",
     "jsdom": "22.1.0",
     "junit-report-builder": "3.0.1",
-    "lint-staged": "13.3.0",
+    "lint-staged": "14.0.1",
     "mocha": "10.2.0",
     "mocha-multi-reporters": "1.5.1",
+    "mocha-suppress-logs": "0.3.1",
     "semantic-release": "21.1.1"
   },
   "lint-staged": {

package/src/steps/extract-metadata.js

@@ -82,10 +82,7 @@ function readBlockConfig($block) {
             value = $img.properties.src;
           }
         }
-        if (value) {
-          // only keep non-empty value
-          config[name] = value;
-        }
+        config[name] = value;
       }
     }
   });
@@ -159,9 +156,20 @@ export default function extractMetaData(state, req) {
   const metaConfig = Object.assign(
     state.metadata.getModifiers(state.info.unmappedPath || state.info.path),
     state.mappedMetadata.getModifiers(state.info.unmappedPath),
-    getLocalMetadata(hast),
   );
 
+  // prune empty values and explicit "" strings from sheet based metadata
+  Object.entries(metaConfig).forEach(([name, value]) => {
+    if (!value) {
+      delete metaConfig[name];
+    } else if (value === '""') {
+      metaConfig[name] = undefined;
+    }
+  });
+
+  // apply document local overrides
+  Object.assign(metaConfig, getLocalMetadata(hast));
+
   // first process supported metadata properties
   [
     'title',
@@ -172,15 +180,18 @@ export default function extractMetaData(state, req) {
     'image-alt',
     'canonical',
     'feed',
+    'twitter:card',
   ].forEach((name) => {
-    if (metaConfig[name]) {
+    if (name in metaConfig) {
       meta[name] = metaConfig[name];
       delete metaConfig[name];
     }
   });
-  // default value for twitter:card (mandatory for rendering URLs as cards in tweets)
-  meta['twitter:card'] = metaConfig['twitter:card'] || 'summary_large_image';
-  delete metaConfig['twitter:card'];
+
+  if (!('twitter:card' in meta)) {
+    // default value for twitter:card (mandatory for rendering URLs as cards in tweets)
+    meta['twitter:card'] = 'summary_large_image';
+  }
 
   // add rest to meta.custom
   meta.custom = metaConfig;
@@ -193,7 +204,7 @@ export default function extractMetaData(state, req) {
   }
 
   // complete metadata with insights from content
-  if (!meta.title) {
+  if (!('title' in meta)) {
     // content.title is not correct if the h1 is in a page-block since the pipeline
     // only respects the heading nodes in the mdast
     const $title = select('div h1', hast);
@@ -202,13 +213,13 @@ export default function extractMetaData(state, req) {
     }
     meta.title = content.title;
   }
-  if (!meta.description) {
-    meta.description = extractDescription(hast);
+  if (!('description' in meta)) {
+    meta.description = extractDescription(hast) || undefined;
   }
 
   // use the req.url and not the state.info.path in case of folder mapping
   meta.url = makeCanonicalHtmlUrl(getAbsoluteUrl(state, req.url.pathname));
-  if (!meta.canonical) {
+  if (!('canonical' in meta)) {
     meta.canonical = meta.url;
   }
 
@@ -222,10 +233,12 @@ export default function extractMetaData(state, req) {
     }
   }
 
-  meta.image = getAbsoluteUrl(
-    state,
-    optimizeMetaImage(state.info.path, meta.image || content.image || '/default-meta-image.png'),
-  );
+  if (!('image' in meta)) {
+    meta.image = getAbsoluteUrl(
+      state,
+      optimizeMetaImage(state.info.path, meta.image || content.image || '/default-meta-image.png'),
+    );
+  }
 
-  meta.imageAlt = meta['image-alt'] || content.imageAlt;
+  meta.imageAlt = meta['image-alt'] ?? content.imageAlt;
 }

package/src/steps/render.js

@@ -26,7 +26,7 @@ function createElement(name, ...attrs) {
   const properties = {};
   for (let i = 0; i < attrs.length; i += 2) {
     const value = attrs[i + 1];
-    if (!value) {
+    if (value === undefined) {
       return null;
     }
     properties[attrs[i]] = value;
@@ -50,9 +50,10 @@ export default async function render(state, req, res) {
     res.document = hast;
     return;
   }
-  const $head = h('head', [
-    h('title', meta.title),
-  ]);
+  const $head = h('head');
+  if (meta.title !== undefined) {
+    $head.children.push(h('title', meta.title));
+  }
 
   // add meta
   // (this is so complicated to keep the order backward compatible to make the diff tests happy)
@@ -69,27 +70,37 @@ export default async function render(state, req, res) {
     'article:published_time': meta.published_time,
     'article:modified_time': meta.modified_time,
     'twitter:card': meta['twitter:card'],
-    'twitter:title': meta.title,
-    'twitter:description': meta.description,
-    'twitter:image': meta.image,
+    'twitter:title': '',
+    'twitter:description': '',
+    'twitter:image': '',
   };
-  // remove meta with no values
-  for (const name of Object.keys(metadata)) {
-    if (!metadata[name]) {
-      delete metadata[name];
-    }
-  }
+
   // append custom metadata
   Object.assign(metadata, meta.custom);
 
-  // remove og:url with explicit removal marker
-  if (metadata['og:url'] === '""') {
-    delete metadata['og:url'];
+  // fallback for twitter
+  const FALLBACKS = [
+    ['twitter:title', 'og:title'],
+    ['twitter:description', 'og:description'],
+    ['twitter:image', 'og:image'],
+  ];
+
+  for (const [from, to] of FALLBACKS) {
+    if (!(from in meta.custom)) {
+      metadata[from] = metadata[to];
+    }
+  }
+
+  // remove undefined metadata
+  for (const name of Object.keys(metadata)) {
+    if (metadata[name] === undefined) {
+      delete metadata[name];
+    }
   }
 
-  appendElement($head, createElement('link', 'rel', 'canonical', 'href', content.meta.canonical));
-  appendElement($head, createElement('meta', 'name', 'description', 'content', content.meta.description));
-  appendElement($head, createElement('meta', 'name', 'keywords', 'content', content.meta.keywords));
+  appendElement($head, createElement('link', 'rel', 'canonical', 'href', meta.canonical));
+  appendElement($head, createElement('meta', 'name', 'description', 'content', meta.description));
+  appendElement($head, createElement('meta', 'name', 'keywords', 'content', meta.keywords));
 
   for (const [name, value] of Object.entries(metadata)) {
     const attr = name.includes(':') && !name.startsWith('twitter:') ? 'property' : 'name';