Repository navigation
feat(site): publish full entry content in the blog and changelog feeds - #9267
Conversation
Both RSS feeds carried only each entry's one-line description. Entries now render through the page's own MDX pipeline into plain, semantic HTML, so everything an entry can show on the site reaches the reader. The output follows Kevin Cox's feed best practices: every URL absolute, no classes or inline styles a reader would strip or misapply, no scripts, and standalone widgets replaced with a pointer back to the page. The feeds now link to themselves, get a 15-minute cache, CORS, and their proper media type, and both are advertised on every page. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LXU532pFoAphw9KaeUxuYj
…Markdown twins Extract the DOM normalizers from the llms-markdown integration into src/utils/rendered-content.ts so feed items and agent Markdown read the same content: asides become labeled blockquotes, tab groups and code frames flatten to labeled panels, data-llms-ignore drops page chrome, and data-llms-only text alternatives are revealed. Feeds now render code frames with the real component instead of a feed-only override, and islands keep their server-rendered markup like the Markdown twins rather than being swapped for a link back to the page. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LXU532pFoAphw9KaeUxuYj
Flattened code frames left <div><div><pre> wrappers behind. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LXU532pFoAphw9KaeUxuYj
@videojs/cdn
@videojs/cli
@videojs/core
@videojs/element
@videojs/html
@videojs/media
@videojs/react
@videojs/spf
@videojs/store
@videojs/utils
@videojs/cloudflare-video
@videojs/dash-video
@videojs/hlsjs-video
@videojs/mux-audio
@videojs/mux-video
@videojs/native-hls-video
@videojs/playerjs-video
@videojs/shaka-video
@videojs/spotify-audio
@videojs/tiktok-video
@videojs/twitch-video
@videojs/vimeo-video
@videojs/wistia-video
@videojs/youtube-video
@videojs/google-cast
@videojs/mux-data
commit: |
📦 Bundle Size Report🎨 @videojs/html — no changesPresets (7)
Media (19)
Players (5)
Skins (53)
UI Components (62)
⚛️ @videojs/react — no changesPresets (7)
Media (23)
Extensions (2)
Players (5)
Skins (34)
UI Components (39)
🧩 @videojs/core — no changesEntries (76)
🏷️ @videojs/element — no changesEntries (2)
📦 @videojs/store — no changesEntries (3)
🔧 @videojs/utils — no changesEntries (13)
📦 @videojs/cdn — no changes📦 @videojs/cloudflare-video — no changes📦 @videojs/dash-video — no changes📦 @videojs/google-cast — no changes📦 @videojs/hlsjs-video — no changes📦 @videojs/media — no changesEntries (3)
📦 @videojs/mux — no changes📦 @videojs/mux-audio — no changesEntries (2)
📦 @videojs/mux-data — no changes📦 @videojs/mux-video — no changesEntries (2)
📦 @videojs/native-hls-video — no changes📦 @videojs/playerjs-video — no changes📦 @videojs/shaka-video — no changes📦 @videojs/spf — no changesEntries (7)
📦 @videojs/spotify-audio — no changes📦 @videojs/tiktok-video — no changes📦 @videojs/twitch-video — no changes📦 @videojs/other — no changesEntries (3)
📦 @videojs/vimeo-video — no changes📦 @videojs/wistia-video — no changesEntries (2)
📦 @videojs/youtube-video — no changesℹ️ How to interpretEach entry is independently bundled, minified, and brotli-compressed. Initial size includes its static import graph; lazy dynamic chunks are reported separately. Entries are not additive because their dependency graphs overlap. Preset rows represent realistic combined bundles. Changes of 300 B or less across initial, lazy, and total size are collapsed, not discarded. Run |
✅ Deploy Preview for vjs-registry ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
✅ Deploy Preview for vjs10-site ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
|
The red checks on this head come from outside this PR, which only touches
Generated by Claude Code |
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LXU532pFoAphw9KaeUxuYj
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LXU532pFoAphw9KaeUxuYj
…log indexes Feed readers still discover both feeds through the alternate links in every page's head. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LXU532pFoAphw9KaeUxuYj
The layout-div pass unwrapped any attribute-less div without its own text, so a div holding only inline elements (a link, a code span) lost its line and ran into its neighbour. Unwrap only divs whose children are all blocks, as the comment already promised. No current blog or changelog item changes. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LXU532pFoAphw9KaeUxuYj
#9267) Render each blog post and changelog entry through the page's MDX pipeline and ship it as content:encoded, cleaned for feed readers (absolute URLs, no presentation hooks, components flattened to plain HTML). Extract the llms-markdown DOM normalizers into utils/rendered-content.ts so feed items and the Markdown twins share one authoring contract (data-llms-ignore, data-llms-only). Add self links, language, lastBuildDate, image, release categories, authors, and card images; cap feeds at the 20 newest entries; serve them as application/rss+xml with CORS and short reader caching; advertise both feeds on every page. Closes #9286 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LXU532pFoAphw9KaeUxuYj (cherry picked from commit d067696)
Before and after
To see the difference, paste a feed URL into any feed reader.
Before: an item is a title and a one-line description, so the reader has to open the site to read anything.
After: the same item carries the full release notes, plus a category. The content is unescaped and trimmed here for readability.
Site components come through as plain HTML. For example, an aside becomes a labeled blockquote:
What
/rss.xmland/changelog/rss.xmlcarried only each entry's one-line description. Each entry now renders through the page's own MDX pipeline and ships ascontent:encoded, so everything an entry can show on the site reaches the reader: asides, code, images, tables and footnotes.Checked against Kevin Cox's RSS feed best practices.
How content is rendered
utils/feeds/feedContent.tsrenders each entry in an Astro container with the MDX and React renderers registered. It passes the site's capitalized MDX components (CodeFrame,Footnotes,RelatedLinks, …), since a body can use them without importing them and rendering fails without them. It leaves out the lowercase overrides (p,a,h2): those only restyle plain elements, and a feed wants them plain.The markup then goes through two passes.
Shared with the Markdown twins:
utils/rendered-content.ts. The DOM normalizers from thellms-markdownintegration (videojs/v10#2883) now live here, and both pipelines callnormalizeRenderedContent(). Feed items and agent Markdown therefore read the same content:data-llms-ignoredrops page-only chrome, anddata-llms-onlytext alternatives are revealed.That gives components one authoring contract for every reader that runs without the site's scripts. Missing feed content should be fixed in these shared normalizers, not with a feed-only path. The extraction is a move with two small additions:
astro-static-slotis unwrapped alongsideastro-slot, andhiddenis removed fromdata-llms-only. I verified that all 430.mdtwins andllms*.txtfiles indistare byte-identical before and after.Feed-specific:
utils/feeds/feedHtml.ts.#fragmentlinks (including footnote references) still land.class,styleanddata-*are dropped. That also drops Shiki's inline token colors: both configured themes are dark, so on a reader's light background they were light-on-light. Code ships as plainpre/code.aria-hiddendecoration, Astro's render comments, and bare layout wrappers.Feed metadata
atom:link rel="self",language,lastBuildDate(newest item's date, so rebuilds don't change it), andimage. The channel<link>now points at/blogor/changeloginstead of the site root.StableorPrerelease, plusBreaking changes, as categories. The compare-on-GitHub link from the page header is appended to the content.dc:creatorper author, and the post'sogImageasmedia:contentwhen it has one.Serving and discovery
public/_headersserves both feeds asapplication/rss+xml(production serves them asapplication/xml) withaccess-control-allow-origin: *for browser-based readers. Readers get a 15-minutemax-ageand revalidate with Netlify's ETag. The CDN keeps the repo's purge-on-deploy TTL. The llms integration appends to this file rather than overwriting it.<link rel="alternate">, blog first, so readers and browser extensions find them. There is no visible subscribe link for now.Choices, with reasons
atom:updated.updatedDatecomes from git history and moves on every typo fix; readers resurface updated items as new.Notable choices worth a second opinion
llms-markdown.tsis refactored, since its normalizers moved out. Thedata-llms-*names now govern feeds too; renaming them seemed like churn for little gain.linkedommoves todependencies, since the cleanup runs at build time.Verification
utils/feeds/tests/,utils/tests/rendered-content.test.ts,releaseCategories, and the existingllms-markdownsuite.astro check: 0 errors.class/style/data-*attribute, island, comment, or relative URL left; OG images resolve to/_astro/.main), checked against the dev-only component test post: references stay superscript links to the entry, definitions render as a list under a "Footnotes" heading, and back-links and the popover script are dropped.🤖 Generated with Claude Code
https://claude.ai/code/session_01LXU532pFoAphw9KaeUxuYj
Closes #9286