| Feed Contents | |||
|---|---|---|---|
| Installation and Setup | by wolf | 21-4-2009 13:18 | |
| User Management | by wolf | 23-4-2009 0:03 | |
| Upgrading from earlier versions | by editor | 21-4-2009 0:22 | |
Installation and Setup
Building from Source
Prerequisites
- AtomicWiki requires a recent build of eXist, at least version 1.3. Since 1.3 has not been release yet, you have to check out the sources from SVN trunk.
- Next, check out AtomicWiki, e.g. using command line tools:
svn checkout http://atomicwiki.googlecode.com/svn/trunk/AtomicWiki AtomicWiki
You should also make sure that the environment variable JAVA_HOME points to the root directory of your Java JDK (the JDK, not the JRE!).
Building
You need to compile eXist first. Switch to the directory into which you checked out eXist. The search functions in AtomicWiki require eXist's Lucene index, so you need to enable it. Edit extensions/indexes/build.properties and change the corresponding property to 'true':
# Lucene integration include.index.lucene = true
Next, compile eXist by calling build.sh or build.bat once. This should create exist.jar and a few other jars.
Back in AtomicWiki, edit build.properties and change exist.home to point to your eXist installation directory. Leave all other properties unchanged. AtomicWiki will create its own, embedded eXist instance and should thus not interfere with other database installations.
Call build.sh or build.bat in the AtomicWiki directory once.
This will:
- copy necessary jars from the eXist installation into the local directories
- build a few extra Java classes needed by the wiki
Start the Server
Call startup.bat or startup.sh to launch the included web server.
If everything worked correct, you should be able to browse to

The first time you try to access this page, you will be presented with a setup page. Clicking on "Start Setup" will load the default AtomicWiki site into the db.
Generating a .war Archive
Calling ant with target "war" creates a file atomic.war in the root directory. This can be deployed into an external servlet engine like e.g. tomcat.
- 0 Comments
- Add Comment
User Management
Adding new users
AtomicWiki is based on eXist's Atom implementation and thus relies on its user management and access control. However, the wiki itself does currently not offer any user management tools. This will change in future versions.
For now, you have to use the Java client which comes with eXist to edit user accounts and permissions. In the default setup, all wiki users should be members of the group "bloggers".
Using the eXist client

You can access AtomicWiki like any other eXist database using the Java client that ships with eXist. Go into the eXist installation you used for building AtomicWiki and launch bin/client.sh or binlient.bat. In the login dialog, change the URL to
xmldb:exist://localhost:8000/xmlrpc
and enter your database admin password in the Password field (default: "atom").
Admin User
AtomicWiki sometimes requires blogadmin user rights for certain administrative tasks. It is impossible to reconstruct the clear text password from the database, so AtomicWiki needs to find it elsewhere.
The blogadmin user and password are written into /db/atom/configuration.xml. If you change the blogadmin user's password, you should also update configuration.xml accordingly.
Upgrading from earlier versions
Backup/Restore
When upgrading from earlier versions of AtomicWiki and/or eXist, make sure you always create a backup of your data first. This is done via the standard backup/restore tools that ship with eXist. When launching the eXist admin client, replace the server URL with
xmldb:exist://localhost:8000/xmlrpc
or the corresponding port and context of your AtomicWiki installation (you always need to add the /xmlrpc).
If you need to restore from a backup into an existing AtomicWiki install, make sure to remove the old data first by deleting the contents of the webapp/WEB-INF/data directory.
Automatic Restore
We prepared an Ant task to make restoring/upgrading easier. The task will restore a specified backup archive, then run an automatic migration process to upgrade old data. This is important as the database layout may have changed between AtomicWiki versions. There have been a few modifications recently. It is thus possible that your restored data does no longer match the expectations of the protocol implementation or AtomicWiki's internal organization. The restore task handles this. To run it, call
build.sh -Dbackup.zip restore
where backup.zip points to the backup archive you want to restore.
If you ever want to manually launch the migration process, just launch
build.sh migrate
The task will not cause any damage if your database is already up to date.
