Fix RSS discovery for pages without updates (#4626)

.changeset/fix-rss-discovery.md

@@ -0,0 +1,5 @@
+---
+"gitbook": patch
+---
+
+Only advertise RSS feeds on pages with Updates blocks.

packages/gitbook/src/components/PageBody/PageBody.tsx

@@ -25,6 +25,7 @@ import {
 import { getLLMsTxtURL, getPageMarkdownURL } from '@/lib/llms-directive';
 import type { AncestorRevisionPage } from '@/lib/pages';
 import { tcls } from '@/lib/tailwind';
+import { getPageRSSURL } from '@/routes/rss';
 
 const LINK_PREVIEW_MAX_COUNT = 500;
 
@@ -86,6 +87,14 @@ export async function PageBody(props: {
 
     return (
         <CurrentPageProvider page={{ spaceId: context.space.id, pageId: page.id }}>
+            {contentHasUpdates ? (
+                <link
+                    rel="alternate"
+                    type="application/rss+xml"
+                    href={getPageRSSURL(context, page)}
+                    title="RSS Feed"
+                />
+            ) : null}
             <main
                 className={tcls(
                     'relative min-w-0 flex-1',

packages/gitbook/src/components/SitePage/SitePage.tsx

@@ -39,7 +39,6 @@ import {
     getDocumentFilterableTags,
     updatesFilterStyleHref,
 } from '@/lib/updates';
-import { getPageRSSURL } from '@/routes/rss';
 
 export type SitePageProps = {
     context: GitBookSiteContext;
@@ -244,9 +243,6 @@ export async function generateSitePageMetadata(props: SitePageProps): Promise<Me
             languages: alternates?.languages,
             types: {
                 'text/markdown': `${linker.toAbsoluteURL(linker.toPathInSpace(page.path))}.md`,
-                // We always reference the RSS feed even if the page doesn't have updates blocks,
-                // It might result in 404, but we can't know here if the page has updates blocks.
-                'application/rss+xml': [{ url: getPageRSSURL(context, page), title: 'RSS Feed' }],
                 // Currently it will output with an empty "type" like <link rel="alternate" href="..." type />
                 // Team at Vercel is aware of this and will ensure it will be omitted when the value is empty in future versions of Next.js
                 // https://gitbook.slack.com/archives/C04K6MV5W1K/p1763034072958419?thread_ts=1762937203.511629&cid=C04K6MV5W1K

packages/gitbook/tests/rss.test.ts

@@ -29,3 +29,22 @@ it('should not expose a RSS feed for a page without updates (root page)', async
     expect(response.status).toBe(404);
     expect(await response.text()).toBe('No updates found in page');
 });
+
+for (const path of ['', '/text-page', '/blocks/updates']) {
+    it(`only advertises an RSS feed when the page has updates (${path || '/'})`, async () => {
+        const response = await fetch(
+            getContentTestURL(`https://gitbook.gitbook.io/test-gitbook-open${path}`),
+            { headers: { 'User-Agent': 'Googlebot' } }
+        );
+        expect(response.status).toBe(200);
+        const html = await response.text();
+        const links = Array.from(html.matchAll(/<link\b[^>]*>/g), ([link]) => link).filter((link) =>
+            link.includes('type="application/rss+xml"')
+        );
+        expect(links).toHaveLength(path === '/blocks/updates' ? 1 : 0);
+        if (path === '/blocks/updates') {
+            expect(links[0]).toContain('/blocks/updates/rss.xml');
+            expect(html.slice(0, html.indexOf('</head>'))).toContain('type="application/rss+xml"');
+        }
+    });
+}