The new and improved 2.0 version of relcfp.
- Thank you to @leno4ka94 for the Jekyll theme.
Made with contrib.rocks.
Use Ruby 3.3.6 (see .ruby-version) and Node.js with npm. The site deploys on Netlify.
bundle config set --local path vendor/bundle
bundle install
npm ci
npm run build
npm run devnpm run build generates _site/, minifies HTML, inline and standalone CSS/JS,
and then generates the Pagefind search index. After a new
checkout, build once before starting the development server so search assets exist.
Edit _includes/main.css; the shared head inlines it to avoid a render-blocking
request. assets/main.css publishes the same source for existing external links.
scripts/minify.cjs changes only generated files; source remains readable.
It preserves whitespace in code blocks and form fields. Pagefind's optimized,
hashed assets are left to Pagefind. JSON-LD is serialized compactly with safe
script escaping. There is no Sass compilation step for site styles.
Jekyll 3 still installs its own Sass converter as a transitive dependency.
_posts/,pages/,_data/: content and navigation data.pages/index.htmlis the homepage;pages/feed/holds the feed templates._layouts/,_includes/: shared Liquid templates.assets/: plain CSS, browser JavaScript, images, and downloadable PDFs.scripts/get_rss.py: feed maintenance, run from the repository root.tests/: feed regressions, browser-script checks, and built-site checks.docs/: audit results, maintenance run log, historical technology reports, and archived theme files._site/,vendor/,node_modules/,.venv/: ignored local build/dependency output.
python3 -m venv .venv
.venv/bin/python -m pip install -r scripts/requirements.txt
.venv/bin/python -m unittest discover -s tests -v
bundle exec ruby tests/test_firehose.rb
bundle exec ruby tests/test_metadata.rb
npm test
npm run build
python3 tests/check_site.pyTo refresh the feed, run .venv/bin/python scripts/get_rss.py from this directory.
This writes pages/feed/feed.xml, _data/feed.json, and docs/RUNLOG.md. The homepage renders the
saved JSON during the build. The scheduled GitHub workflow refreshes it separately.
See the optimization audit for findings and measurements.
The responsive stylesheet keeps the existing colors and background, uses fluid text
and bounded reading widths, and changes card grids by screen size. Archive images
load lazily; all archive posts remain visible if JavaScript is disabled. Search
scripts are deferred, and keep_files preserves the generated search bundle during
Jekyll's local rebuilds. Run a production build again to refresh search content.
See responsive checks for screen sizes and browser results.
The shared head uses _plugins/metadata.rb and _includes/metadata.html to publish
one compact, script-safe Schema.org JSON-LD graph per page. It connects the
organization, website, page, visible breadcrumbs, and (on posts) blog and
BlogPosting. Open Graph, Twitter cards, Dublin Core, and Zotero-compatible
Highwire citation tags use the same values. Citation tags appear only on posts;
z.itemType identifies them as blog posts rather than journal articles.
Set author to a key in _data/authors.yml, a name, a mapping with name and
url (or website/uri), or a list of those values. Missing authors use the
RELCFP organization. Author bylines use the standard p-author h-card
microformat; posts and listing entries use h-entry, with e-content around
the full post text. There is no standard h-author root microformat.
Optional description, image (path or mapping with path and alt), lang,
tags, and last_modified_at enrich the metadata. Dates come from content,
never build time. Existing external canonical_url values remain search
canonicals and are recorded as isBasedOn in the post schema; citations and
microformat identifiers refer to the local curated post. No journal, event,
deadline, or scholarly-paper details are inferred from the text of a CFP.
The RELCFP profile has an organization h-card with p-org and a canonical
u-uid; listing feeds use their visible headings as p-name. Reference guides:
Microformats, h-card,
h-entry, h-feed,
and IndieWebify.Me. These cover identity and publishing
markup; Webmention sending/receiving requires a separate service or backend.
The production build publishes every curated post in three formats:
- RSS 2.0
- JSON Feed 1.1
- twtxt, with the oldest posts first
The imported feed remains at /feed.xml, and the existing curated Atom feed at
/posts/feed.atom. The Atom feeds are published without an XSL stylesheet.
Feed templates have no HTML layout
and are excluded from the sitemap. Netlify sets the feed content types.
Firehose Atom combines both Atom feeds on every
Jekyll build, newest updates first. Entries linking to the same URL are included
once, preferring the curated post's full content. Its template lives in
pages/feed/firehose.xml; _plugins/firehose.rb merges the rendered source feeds.
humans.txt credits all human GitHub contributors
checked on October 28, 2026, the theme author thanked in this README, and the
existing Creativitas acknowledgment. _data/contributors.json records the source
and excludes the two automation accounts from the human list. Refresh that list
from GitHub when contributors change; builds do not need a GitHub connection.
Root files are limited to conventional project/build entry points and the favicon.
The Python requirements are in scripts/requirements.txt; maintenance logs and
historical theme material are under docs/. Pagefind's default glob and indexed
content boundaries replace the redundant root configuration file.