This is the source code for the Clojars jar repository. If you're looking for user documentation, try the wiki.
This is only of interest to you if you want to hack on the Clojars source code (send me a pull request with your patches) or if you want to try to run your own copy (for example inside a company).
There are several ways to run Clojars depending on what you intend to do with it. Regardless of how you run it, you first need to do some setup:
-
Install Leiningen
- Mac OS X Homebrwe:
brew install leiningen
- Mac OS X Homebrwe:
-
Install SQLite3
- Debian:
apt-get install sqlite3
- Mac OS X Homebrew:
brew install sqlite
- Debian:
-
Create an initial sqlite database:
mkdir data; sqlite3 data/db < clojars.sql
To run the application as standlone from the console:
-
Compile with:
lein uberjar
-
Run the webapp:
java -jar clojars-web-*-standalone.jar 8080 8701
-
Now try hitting localhost:8080 in your web browser.
To run the application in auto-reload mode, from the console:
- Run
lein ring server
and that's it, it should automatically open a browser in localhost:3000.
If you'd like to run it out of an editor/IDE environment you can
probably eval a call to the -main
function in
src/clojars/core.clj
.
- Install VirtualBox 4.
- Install Vagrant by running: gem install vagrant. You may need to install rubygems first.
- Run vagrant up and wait, it'll take a while.
Once the virtual machine is up and running you can access the web app at
localhost:8088. To ssh into it you can just run
vagrant ssh
. To push new jars, ssh is forwarded from the local port 2222 so
the command is:
scp -P 2222 foobar-0.0.1.jar pom.xml clojars@localhost:
. You can also ssh
directly to port 2222 to the vagrant user, ssh -p 2222 vagrant:localhost
,
with the password "vagrant".
When something goes wrong, inside the virtual machine, you can find the logs in
/var/log/clojar
.
If you'd like to hack on the UI or search it might be useful to have production-like metadata. I've put up a production database dump (with password hashes and email addresses stripped of course) which you can use like this:
wget http://meshy.org/~ato/clojars-test-data.sql.gz
rm data/db
zcat clojars-test-data.sql.gz | sqlite3 data/db
If you want all the actual jar files as well you can grab them via rsync.
The SSH integration is kind of a hack and needs improvement. Currently it uses Nailgun but the current setup has threading problems due to the way it does IO.
Basically clojars.scp
implements the SCP protocol
and acts as a Nailgun "nail" so it can be called from the
command-line. Clojars writes an SSH authorized_keys
file
with a line for each user's public key setting the login command to
ng --nailgun-port 8700 clojars.scp USERNAME
.
To set it up:
-
Install the Nailgun
ng
C program. Just unpack the Nailgun source and runmake
then copy theng
executable somewhere like/usr/local/bin
-
Create a "clojars" unix account with a disabled password.
-
Disable password authentication for that user by adding this to
/etc/ssh/sshd_config
:Match User clojars PasswordAuthentication no
-
Symlink in the auth_keys file the webapp generates:
cd /home/clojars mkdir .ssh cd .ssh ln -s ...../clojars-web/data/authorized_keys authorized_keys
There's a public mailing list clojars@googlegroups.com for general discussion. There's also a separate maintainers list for people looking after the repository and server.