Best way to get maven dependencies - maven-2

I only need to use this class org.apache.commons.io.FileUtils, and yet I'm downloading all commons Classes which I actually don't need, is there a way to say to maven download just FileUtils class? Not whole commons like from dependency below
<dependency>
<groupId>commons-io</groupId>
<artifactId>commons-io</artifactId>
<version>1.4</version>
</dependency>

is there a way to say to maven download just FileUtils class?
No. But depending on your exact use case, you could maybe use the Maven Shade Plugin to create an uber-jar and filter the content of the included dependencies:
Selecting Contents for Uber JAR
...
For fine-grained control of which
classes from the selected dependencies
are included, artifact filters can be
used:
<project>
...
<build>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-shade-plugin</artifactId>
<version>1.3.3</version>
<executions>
<execution>
<phase>package</phase>
<goals>
<goal>shade</goal>
</goals>
<configuration>
<filters>
<filter>
<artifact>junit:junit</artifact>
<includes>
<include>junit/framework/**</include>
<include>org/junit/**</include>
</includes>
<excludes>
<exclude>org/junit/experimental/**</exclude>
<exclude>org/junit/runners/**</exclude>
</excludes>
</filter>
<filter>
<artifact>*:*</artifact>
<excludes>
<exclude>META-INF/*.SF</exclude>
<exclude>META-INF/*.DSA</exclude>
<exclude>META-INF/*.RSA</exclude>
</excludes>
</filter>
</filters>
</configuration>
</execution>
</executions>
</plugin>
</plugins>
</build>
...
</project>
Here, Ant-like patterns are used to
specify that from the dependency
junit:junit only certain
classes/resources should be included
in the uber JAR. The second filter
demonstrates the use of wildcards for
the artifact identity which was
introduced in plugin version 1.3. It
excludes all signature related files
from every artifact, regardless of its
group or artifact id.
But note that FileUtils depends on other classes:
import org.apache.commons.io.filefilter.DirectoryFileFilter;
import org.apache.commons.io.filefilter.FalseFileFilter;
import org.apache.commons.io.filefilter.FileFilterUtils;
import org.apache.commons.io.filefilter.IOFileFilter;
import org.apache.commons.io.filefilter.SuffixFileFilter; // depends on org.apache.commons.io.IOCase
import org.apache.commons.io.filefilter.TrueFileFilter;
import org.apache.commons.io.output.NullOutputStream;
That you'll obviously need to include too.

Apache commons io has no dependencies to other apache commons projects. You get only commons io, no other commons libraries. That is one jar with about 100 classes, not very much.
You cannot get only one class into your project - this would propably also violating the license!
A look at FileUtils source also shows a lot of imports of other commons io classes. It will not work without the rest of the jar!

Use the dependency <exclusion> element
<dependency>
<groupId>sample.ProjectA</groupId>
<artifactId>Project-A</artifactId>
<version>1.0</version>
<scope>compile</scope>
<exclusions>
<exclusion> <!-- declare the exclusion here -->
<groupId>sample.ProjectB</groupId>
<artifactId>Project-B</artifactId>
</exclusion>
</exclusions>
</dependency>
to exclude those transitive dependencies that you don't need.
It is also a good practice to use mvn dependency:analyze-only and mvn dependecy:tree to understand how your depency graph is actually structured and what dependencies are you really using and not declaring and/or declaring and not using.
Regards.

I don't think so. The artifacts are packaged as jar files; you can't get them individually (that is, on a file-by-file basis). At least, not to my knowledge.
Also, think about it a little more - it is entirely possible that the FileUtils class has dependencies on other classes. But you can't really tell what they are without examining the source. That is information the user of the package does not need to know. You wouldn't want to figure out every other class that FileUtils uses (or what other classes the dependencies of FileUtils uses and so on and so forth). This is why the entire artifact is distributed as a discrete and self-contained entity. The artifact as a whole, if it is mavenized, will know what dependencies it needs and maven will go grab those for you as well.

Related

JPA persistence unit definitions are conflicting

I want to use one class and two JPA persistence units and as such to be able to store data in different tables (or even databases) and different definitions.
According to the JPA2.2 specification this should be possible but I experience weird behaviour. I'm using payara which uses eclipselink.
For a complete description and a reproducer see this github project.
I hope someone can help me.
The reason why this does not work is eclipselink weaving.
What this does is manipulate the bytecode of entity classes to extend them with functionality so provide all kinds of optimizations.
The problem however is that the resulting 'woven' classes depend on the definitions as defined in the persistence.xml and orm.xml.
This means in my case, because I have two different persistence.xml/orm.xml combinations, I would need two Foo.class and Bar.class files which reflect the different functionality. Of course this won't work.
The solution is to turn off weaving and this can be done using a property in the persistence.xml
<property name="eclipselink.weaving" value="off"/>
If you want to see the actual 'woven' classes you can use static weaving. This can be done using this property
<property name="eclipselink.weaving" value="static"/>
and this maven plugin
<plugin>
<groupId>de.empulse.eclipselink</groupId>
<artifactId>staticweave-maven-plugin</artifactId>
<version>1.0.0</version>
<executions>
<execution>
<phase>process-classes</phase>
<goals>
<goal>weave</goal>
</goals>
<configuration>
<persistenceXMLLocation>META-INF/persistence.xml</persistenceXMLLocation>
<logLevel>FINE</logLevel>
</configuration>
</execution>
</executions>
<dependencies>
<dependency>
<groupId>org.eclipse.persistence</groupId>
<artifactId>org.eclipse.persistence.jpa</artifactId>
<version>2.7.7.payara-p3</version>
</dependency>
</dependencies>
</plugin>
When you then build the project you can decompile the classes in the jar to see what is happening.
Thanks you Cris for pointing me to eclipselink weaving.

New at Maven: Using the shade plugin and 3rd party jars

This should be pretty simple, but I can't the around to it. I need to create an uberjar using 3rd party jars. I've already followed these instructions: Including a non-Mavenized dependency so it works with maven-shade-plugin and added them to the local repository. But now what? Every Maven tutorial/example is kinda shady (pun intended) and I just don't know how to edit the .pom file properly in order to make it work.
Besides, I'm confused about the shade "plugin" overall. I mean, I followed the basic Maven tutorials and it went all fine and dandy. But when I look into the shade plugin page, I can't find it to download, except for the source code. I mean, isn't it a plugin? Shouldn't I download the binaries and install it into Maven somehow?
Sorry about the extreme noobish question but, needless to say, I know squat about Maven.
To create your shaded (uber) jar, you just need to declare the shade plugin in your pom.xml.
With regards to installation of the shade plugin, simply declaring it in the plugins section of your pom.xml is all you need do. Maven plugins are not installed manually, but are automatically downloaded by Maven (if not already downloaded; just like dependencies), stored in your local repository, and used whenever a project needs them.
As to using it, much like other plugins, declare it in your pom.xml by adding a <plugin> element with your configuration needs. This plugin does nothing automatically (some do, some don't) - you have to specify which "goal" to execute (think "method of a class"), and in which "phase" (think "step" of the build process). Unless you have strange needs, specify the "shade" goal in the "package" phase (see below).
For more configuration possibilities, see the shade usage page, and their examples (especially selecting contents for uber jar). Here is a simple example which, when you run mvn package, replaces your original jar in the target/ directory with the uber jar. It only includes the runtime dependencies, not the ones used at test time (notice the <scope> element of the junit dependency, which is not included in the uber jar).
<project>
<groupId>com.sample</groupId>
<artifactId>test</artifactId>
<version>1.0</version>
<build>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-shade-plugin</artifactId>
<version>1.5</version>
<executions>
<execution>
<phase>package</phase>
<goals>
<goal>shade</goal>
</goals>
</execution>
</executions>
</plugin>
</plugins>
</build>
<dependencies>
<dependency>
<groupId>com.sample</groupId>
<artifactId>test-core</artifactId>
<version>1.0</version>
</dependency>
<dependency>
<groupId>junit</groupId>
<artifactId>junit</artifactId>
<version>3.8.1</version>
<scope>test</scope>
</dependency>
</dependencies>
</project>
If you have already downloaded the binaries and installed them into your local Maven repository all that remains is to declare them as dependencies in the POM. If the shade plugin is also declared in the POM a simple "mvn install" should generate a standalone JAR in the target directory.

Which maven2 lifecycle-phase to choose?

I have a Java EE-web-application and for using my project with oc4j application server it must be patched in my build-lifecycle to avoid several issues. Actually i do this via maven-antrun-plugin which works great. I have to remove, copy some special libraries into WEB-INF/lib and edit the web.xml, to avoid clashes with EL functions and classloading issues.
According to the maven lifecycle phases i chosed the phase prepare-package: this phase is executed before the war file is packaged, but unfortunately also before the (re-)sources are copied into the temporary working dir. I dislike working on the source folders because they're under version control and i don't want to have my coworkers to accidently commit them in cause the build-tool modified them.
So maven copies all the (re-)source stuff to target/__finalName__ where i want to fix the project for the use with oc4j. because this folder is temporary and will be packaged into the war file. Unfortunately the copying and packaging is isolated done in lifecycle package.
So how can i get between the copying of the sources and resources and the real packaging?
Example with prepare-package
This example doesn't work because the ${project.build.directory}/${build.finalName} doesn't exists and the ojdbc14.jar wasn't copied there in this phase.
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-antrun-plugin</artifactId>
<version>1.4</version>
<executions>
<execution>
<id>patch-oc4j</id>
<phase>prepare-package</phase>
<goals>
<goal>run</goal>
</goals>
<configuration>
<tasks>
<echo>Patching distribution for OC4J</echo>
<echo>Deleting the obsolete OJDBC library</echo>
<delete file="${project.build.directory}/${build.finalName}
/WEB-INF/lib/ojdbc14.jar" />
[... more patching ...]
</tasks>
</configuration>
</execution>
</executions>
</plugin>
Couldn't you use a profile for this? Maybe something like this:
<profiles>
<profile>
<id>oc4j</id>
<dependencies>
<dependency>
<groupId>com.oracle</groupId>
<artifactId>ojdbc14</artifactId>
<version>10.2.0.4.0</version>
<scope>provided</scope>
</dependency>
</dependencies>
</profile>
</profiles>
I have to remove, copy some special libraries into WEB-INF/lib and edit the web.xml, to avoid clashes with EL functions and classloading issues.
Sounds like you could, in part at least, do this with Build Profiles instead.. Your motivation for the problem above is a bit short, but if you elaborate we can judge this better..

maven dependencies groovy

I'm running a project that has a dependency on groovy 1.7-beta-1. The gmaven plugin uses groovy version 1.6 as a dependency. In my pom, I specify in the dependency management section the grooyv-all version as :
<dependencyManagement>
<dependencies>
<dependency>
<groupId>org.codehaus.groovy</groupId>
<artifactId>groovy-all</artifactId>
<version>1.7-beta-1</version>
</dependency>
</dependencies>
</dependencyManagement>
Yet when I run maven in debug mode I see that groovy 1.6 is being used for a dependency to the gmaven plugin. I thought my dependency management section would override this so they all use 1.7-beta-1, but I'm getting errors due to the different groovy versions. any help here would be appreciated.
thanks,
Jeff
Here's a refined version of Pascal's answer. I upgraded the main plugin version to 1.2, the dependency to Groovy 1.7, and wrapped it all in a pluginManagement tag so that it will nicely leverage the inheritance model.
Keep in mind that the 1.3-SNAPSHOT of the GMaven plugin has already begun using the 1.7-rc2 Groovy provider.
<!-- I wrapped everything in a plugin management section so that this can be neatly inherited across all your poms -->
<pluginManagement>
<plugins>
<plugin>
<groupId>org.codehaus.gmaven</groupId>
<artifactId>gmaven-plugin</artifactId>
<!-- Notice I upgraded it to 1.2 -->
<!-- Details here http://repo1.maven.org/maven2/org/codehaus/gmaven/gmaven-plugin/1.2/gmaven-plugin-1.2.pom -->
<version>1.2</version>
<dependencies>
<dependency>
<groupId>org.codehaus.gmaven.runtime</groupId>
<artifactId>gmaven-runtime-1.7</artifactId>
<version>1.2</version>
</dependency>
</dependencies>
</plugin>
</plugins>
</pluginManagement>
Overriding a dependency used by a plugin is a nice ability that was actually introduced by Maven 2.0.9.
To do so, at least with a plugin that you are using as a normal build plugin - as opposed to a report which is not the case with the the gmaven-plugin so I won't cover this case here - simply add a dependency block inside the plugin block, like this (this is a sample so versions may be inaccurate):
<plugin>
<groupId>org.codehaus.groovy.maven</groupId>
<artifactId>gmaven-plugin</artifactId>
<version>1.0</version>
<executions>
<execution>
<goals>
<goal>generateStubs</goal>
<goal>compile</goal>
<goal>generateTestStubs</goal>
<goal>testCompile</goal>
</goals>
</execution>
</executions>
<dependencies>
<dependency>
<groupId>org.codehaus.groovy</groupId>
<artifactId>groovy-all</artifactId>
<version>1.7-beta-1</version>
</dependency>
</dependencies>
</plugin>
As long as the new version of the dependency is "API compatible" with the version the plugin was linked against, you should be ok. If not, then you'll obviously have to upgrade to a newer version of the plugin compatible with the new API (i.e. likely using it as dependency), which is what you did.
To make gmaven accurately picks the right runtime is by configuring the "providerSelection" value, e.g.
<plugin>
<groupId>org.codehaus.gmaven</groupId>
<artifactId>gmaven-plugin</artifactId>
<configuration>
<providerSelection>1.7</providerSelection>
</configuration>
FYI, for the groovy:providers mojo, these are the configurations it expects (I extracted them by debugging to org.apache.maven.lifecycle.DefaultLifecycleExecutor.executeGoals(List, Stack, MavenSession, MavenProject) (look for XmlPlexusConfiguration):
<configuration>
<remoteRepositories implementation="java.util.List">${project.pluginArtifactRepositories}</remoteRepositories>
<project implementation="org.apache.maven.project.MavenProject">${project}</project>
<artifactRepository implementation="org.apache.maven.artifact.repository.ArtifactRepository">${localRepository}</artifactRepository>
<pluginArtifactMap implementation="java.util.Map">${plugin.artifactMap}</pluginArtifactMap>
<providerSelection implementation="java.lang.String">${gmaven.runtime}</providerSelection>
</configuration>
You need to add a similar 1.7 dependency to the dependencies of the plugin in a similarly structured <plugin> or <pluginManagement> section. Your dependency management section you are adding is proper, but does not affect the plugin dependencies. I'll try to review this reply and post an example later when I'm back at my desk.

Sharing Test code in Maven

How can you depend on test code from another module in Maven?
Example, I have 2 modules:
Base
Main
I would like a test case in Main to extend a base test class in Base. Is this possible?
Update: Found an acceptable answer, which involves creating a test jar.
I recommend using type instead of classifier (see also: classifier). It tells Maven a bit more explicitly what you are doing (and I've found that m2eclipse and q4e both like it better).
<dependency>
<groupId>com.myco.app</groupId>
<artifactId>foo</artifactId>
<version>1.0-SNAPSHOT</version>
<type>test-jar</type>
<scope>test</scope>
</dependency>
Thanks for the base module suggestion. However, I'd rather not create a new module for just this purpose.
Found an acceptable answer in the Surefire Maven documentation and a blog. See also "How to create a jar containing test classes".
This creates jar file of code from src/test/java using the jar plugin so that modules with tests can share code.
<project>
<build>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-jar-plugin</artifactId>
<version>2.4</version>
<executions>
<execution>
<goals>
<goal>test-jar</goal>
</goals>
</execution>
</executions>
</plugin>
</plugins>
</build>
</project>
In order to use the attached test JAR that was created above you simply specify a dependency on the main artifact with a specified classifier of tests:
<project>
...
<dependencies>
<dependency>
<groupId>com.myco.app</groupId>
<artifactId>foo</artifactId>
<version>1.0-SNAPSHOT</version>
<type>test-jar</type>
<scope>test</scope>
</dependency>
</dependencies>
...
</project>
We solved this by making a maven project with test code as the src/main/java and adding the following dependency to projects:
<dependency>
<groupId>foo</groupId>
<artifactId>test-base</artifactId>
<version>1</version>
<scope>test</scope>
</dependency>
Worked for me for 1 project, but I didn't for another after doing exactly the same steps.
So I debugged:
After mvn clean install I checked /target directory: .jar was there so thats good
Ran mvn dependency:tree on a project which should use those test classes. Noticed that generated jar file with test classes is marked as dependency, so thats good.
Conclusion could be only one - I restarted my Intellj. At first class import was still not visible, but after a minute it started to see it!
Note: I only restarted Intellj, no caches removal etc
Yep ... just include the Base module as a dependency in Main. If you're only inheriting test code, then you can use the scope tag to make sure Maven doesn't include the code in your artifact when deployed. Something like this should work:
<dependency>
<groupId>BaseGroup</groupId>
<artifactId>Base</artifactId>
<version>0.1.0-SNAPSHOT</version>
<scope>test</scope>
</dependency>