I'm using Apache Ivy + IvyDE for getting my project's dependencies, which are:
<dependency org="com.google.guava" name="guava" rev="r08" />
<!-- logging -->
<dependency org="org.slf4j" name="jcl-over-slf4j" rev="1.6.1" />
<dependency org="ch.qos.logback" name="logback-classic" rev="0.9.27" />
<!-- database -->
<dependency org="org.hibernate" name="hibernate-entitymanager" rev="3.6.2.Final" />
<dependency org="org.hibernate" name="hibernate-validator" rev="4.1.0.Final" />
<dependency org="org.hibernate" name="hibernate-c3p0" rev="3.6.2.Final" />
<dependency org="mysql" name="mysql-connector-java" rev="5.1.14" />
Sources are the Maven and JBoss (Hibernate) repositories.
As you can see I'm using logback+SLF4J for logging, but for some reason Ivy will download log4j and slf4j-log4j as well, which causes a few small problem in my application.
Is there a way to see why this happens, to see which of the dependencies above depend on log4j? Can I get a dependency graph/tree generated from Ivy/IvyDE?
And is there then a way to prevent this from happening?
We have an ant target that looks like this:
<target name="report" depends="init">
<mkdir dir="report" />
<!--
The type attribute is optional, we're using it to exlude other dependcy types we're not interested in.
Note that each resolve uses that list (via a property) in our build.
-->
<ivy:resolve type="jar,ejb,tld,bundle"/>
<ivy:report todir="report" />
</target>
Then it's just a call ant report
and Ivy will generate the report as HTML in the given directory.
Take a look at the Ivy documentation for ivy:report.
Edit:
To prevent the inclusion of those artifacts/dependencies, you could try transitive="false"
on the <dependency ..>
element, or use <exclude>
. For example, we use Hibernate 3 but don't want to have JTA 1.1
, so our ivy.xml
containts this: <exclude module="jta"/>
to exclude all transitive JTA dependencies.