Feeds, sitemap and search

Folia writes these outputs itself, from what it built, without a template. Each is optional or automatic as noted.

Redirects

A page's redirect_from makes a stub page at each old URL that redirects to the new one, and all of them are listed in /redirects.json. The stub's target uses the configuration's [site] url.

A stub refreshes to its target at once, and names it as its canonical URL. On a dark site, give the stubs the site's colours with [redirects], so that following an old URL does not flash a white page:

[redirects]
background = "#1d1f21"
color-scheme = "dark"

Atom feeds

Written only when the configuration has [feeds]:

Feed	URL
the blog	/blog/feed.xml
each tag	/blog/tag/TAG.xml (a tag with a space too: /blog/tag/stem cells.xml, linked percent-encoded)
the microblog, and each kind	/micro/feed.xml, /feed/micro/KIND.xml
the garden	/dg/feed.xml
combined (optional)	the page you name in [feeds.combined], such as /feed.xml

Each lists the newest limit items (10 by default). A feed's <updated> is its newest item's date, never the build time, so rebuilding does not make every feed look new. There are no category feeds. The generator table decides what the feeds say generated them.

Sitemap

/sitemap.xml lists every page by URL except /404.html (redirect stubs are not pages), then the .htm, .html, .xhtml and .pdf static files. A page's <lastmod> is its last modification; a page with no source, such as a category archive, has none. A static file's is its last git commit.

robots.txt

Folia writes a default /robots.txt that lets every crawler fetch everything and points to the sitemap, using the configuration's [site] url:

User-agent: *
Allow: /
Sitemap: https://example.org/sitemap.xml

To write your own, put a robots.txt at the root of the site. A file of the site takes the URL, so it is copied as it is and Folia writes no default. A preview build gets the same default; keep it off the public web by where you put it.

Search

Written only when the configuration has [search]. Folia writes documents, the array of records that the lunr search needs ({id, url, title, body}), followed by your own script, the fragment file, as it is. So the search UI and the lunr setup are yours; Folia supplies the data.

• Records: the site's pages by URL (not .xml, assets, category or tag URLs), then the RSVPs, likes and bookmarks, the garden notes and the posts, oldest first. A micro entry's body starts with its date.
• A body is the content's text: markup removed, a space after each full stop, and a line break is a space.
• A listing page is indexed by the text of the template block content-block names, so the navigation and footer are not in the index. Without content-block, the whole rendered page is.

The script ends up at [search] source (assets/js/search.js in the example).