Feeds

RSS and Atom feeds

Feeds are built into core and require no module activation or additional configuration beyond a valid site_url (or OCMS_SITE_URL at startup).

Subscription	RSS 2.0	Atom 1.0
All posts	/rss.xml	/atom.xml
Category	/category/{slug}/rss.xml	/category/{slug}/atom.xml
Tag	/tag/{slug}/rss.xml	/tag/{slug}/atom.xml

Unprefixed URLs always use the active default language. Other active languages use their URL prefix, for example /ru/rss.xml or /ru/category/technology/rss.xml. Cookies, Accept-Language and ?lang= do not change a feed URL's language. A taxonomy slug must belong to that language; missing terms and unknown or inactive language prefixes return 404.

Included content

Each feed contains at most 20 published posts, newest publication first, with the page ID as the descending tie-breaker. If a published post has no publication timestamp, its creation timestamp supplies both the displayed date and ordering. Ordering compares actual instants across time zones and retains fractional seconds. Language, taxonomy and valid article-slug filters apply before the limit, so unroutable legacy posts do not consume subscription slots. Static pages, drafts, scheduled-but-unpublished posts and posts marked Exclude from lists are omitted. No index controls search engine indexing and does not remove a public post from feeds.

Entries contain the title, absolute article link, public author name, publication and modification dates, and a plain-text summary. The saved summary takes precedence; otherwise the body becomes an excerpt of at most 300 Unicode characters, including the ellipsis. Markup and script/style contents are removed. RSS escapes literal markup for display as text; Atom summaries use type="text". Full article HTML and media enclosures are not included.

Entry identifiers use the configured site origin and page ID (/page/{id}), independent of the article slug. Renaming or updating a post retains its identity; changing the configured origin changes the identifier. Author email addresses are never emitted. Posts without a publication timestamp use their creation timestamp. Empty feeds remain valid documents with site/language/taxonomy metadata supplying the feed update date.

HTTP behavior

GET returns UTF-8 XML with application/rss+xml or application/atom+xml. HEAD returns the same representation headers without a body, including when gzip or deflate compression is negotiated. Feeds are generated from current database data on every request, with Cache-Control: public, no-cache and a deterministic weak ETag that remains valid under compression. A matching If-None-Match returns 304 without a body; publishing, editing, unpublishing, deletion and taxonomy association changes are reflected on the next request.

Absolute URLs come from site_url, never request or proxy headers. An absent or invalid HTTP(S) origin returns 503. Database or serialization failures return 500 rather than an apparently empty feed. Error responses use no-store. Internal analytics already excludes .xml requests, so reader polling does not count as page views.

Theme discovery

The shipped themes, starter and fallback layout emit subscription discovery links in the HTML head. Every normal frontend page has the current language's site subscriptions; category and tag archives add their contextual subscriptions. No feed links are advertised when the site origin is unconfigured or invalid.

Custom HTML themes can render the additive BaseTemplateData.FeedLinks list:

{{range .FeedLinks}}
<link rel="alternate" type="{{.MIMEType}}" title="{{.Title}}" href="{{.URL}}">
{{end}}

Each FeedLink exposes Title, MIMEType and absolute URL. Existing themes continue working without this addition. Feed paths are reserved by the Migrator alongside the other core public routes.

The serializers follow the RSS 2.0 specification and Atom 1.0 standard. RSS includes Atom self/update and Dublin Core creator extensions for discovery, modification dates and author display names.