The clojuredocs.org webapp
A few ways to contribute:
- Find & report bugs: https://github.com/zk/clojuredocs/issues
- Suggestions on how to make development on the site more friendly (docs, codebase organization).
Let's use GH issues for discussion for now
If you're looking for a project:
- Content for namespaces (see
src/md/namespaces
) needs to be added / edited. Example of clojure.core.async - Listing of clojure training / classes / events on home page
- Stand-alone example page, maybe have the var info (signature, doc string, etc) at the top.
- Source-linking on libs not included in the standard library e.g. core.async.
Production is deployed on an AWS t2.micro instance. There's an nginx process running on the box, balancing to two JVMs managed by Upstart to support zero-downtime deploys.
To regenerate the upstart scripts:
cd $REPO
sudo foreman export -a clojuredocs -e ./.env -u ubuntu -c "web=2" upstart /etc/init/
To start the app processes:
sudo service clojuredocs-web-1 start
sudo service clojuredocs-web-2 start
To redeploy:
# in $REPO
sudo service clojuredocs-web-1 stop
git pull origin master
# This will compile assets & run tests
bin/build
sudo service clojuredocs-web-1 start
# Wait for proc 1 to start serving requests
sudo service clojuredocs-web-2 restart
- > JDK 7
- JCE Unlimited Strength Jurisdiction Policy Files
- lein
- foreman (see
Procfile
,bin/dev
) - less
- entr (available in homebrew)
- MongoDB
Run bin/dev
, which will start all the things (repl, web process,
scss compiler, etc). See Procfile
for more info.
Connect to the repl and / or visit http://localhost:4000
You'll notice that var information is already populated. In an effort to not make the same mistakes again, all core-related var info is loaded from the runtime version of Clojure on start up.
OTOH, examples, see-alsos, and notes (and any other user-generated content) are stored in the database.
The app uses a MongoDB database named clojuredocs
to store data. Run bin/db-reset
to seed the database with a recent production export (you must be running mongod
for this to work).
Occasionally you'll need to compile and run things as they would be in production (checking advanced cljs compilation, for example): bin/prod-local
.
The ClojureDocs project is set-up to emit source-maps for compiled javascript. To enable in Chrome, check the 'Enable JS source maps' option in the Developer Tools settings pane.
Clojure vars are pulled directly from the runtime, and are not stored in the database. When new versions of Clojure are released:
- Change the Clojure dep in
project.clj
- Update the version string, source base url, and gh tag url in
clojuredocs.search/clojure-lib
- Update the version for the mobile nav in
clojuredocs.pages.common
- Update the github URL in
clojuredocs.pages.vars/source-url
.
Interesting files:
src/clj/clojuredocs/main.clj
-- Main entry-point into the app, starts the jetty serversrc/clj/clojuredocs/entry.clj
-- Root routes and middlewaresrc/clj/clojuredocs/pages.clj
-- User-facing HTML pagessrc/clj/clojuredocs/api.clj
-- API endpoints for ajax calls from the frontend.src/cljs/clojuredocs/main.cljs
-- Main js entry-point into the app
- Functions that return hiccup structures should be prefixed with a
$
, like$layout
. - Mutable state should be prefixed with a
!
, ex:!my-atom
.
Most vars are picked up from Clojure at runtime, using core namespace
and var introspection utilities. Special forms and namespaces to
include on the site are specified explicitly in the
clojuredocs.search.static
namespace.
Vars are picked up automatically based on the namespaces specified in
clojuredocs.search.static/clojure-namespaces
vector. Any namespace
in this vector will be queried for public vars to be made searchable
and displayable on the site.
Special forms are specified in the
clojuredocs.search.static/special-forms
list, and require a server
restart to be picked up in a dev environment.
Sometimes we'd like to add core libraries that are not part of the standard Clojure distribution (like core.async) to the site. Here's how to do that:
- Add dependency to
project.clj
- Add ns sym to
clojure-namespaces
inclojuredocs.search.static
- Add a short description + links to community articles / videos /
other resources to
src/md/namespaces/
- Dev starts the environment using
lein repl :headless
, prod useslein run -m clojuredocs.main
. See:repl-options
inproject.clj
for initialization options.
- See also relations (ns/name -> [ns/name]), results cached for 1 minute.
Zachary Kim, Lee Hinman, and more.
Copyright © 2010-present Zachary Kim
Distributed under the Eclipse Public License either version 1.0 or (at your option) any later version.