Technology Updates by Abhishek Chitransh...: Maven

Apache Maven

Developer(s)
Apache Software Foundation

Stable release
2.2.1 / August 11, 2009; 11 months ago (2009-08-11)

Development status
Active

Written in
Java

Operating system
Cross-platform

Type
Build Tool

License
Apache License 2.0

Website
http://maven.apache.org

Maven is a software tool for project management and build automation. While Maven is primarily used for Java programming, it can be used to build and manage projects written in C#, Ruby, Scala, and other languages. It is similar in functionality to the Apache Ant tool, but is based on different concepts. Maven is hosted by the Apache Software Foundation, where it was formerly part of the Jakarta Project.

Maven uses a construct known as a Project Object Model (POM) to describe the software project being built, its dependencies on other external modules and components, and the build order. It comes with pre-defined targets for performing certain well defined tasks such as compilation of code and its packaging.

Maven dynamically downloads Java libraries and Maven plug-ins from one or more repositories. Maven provides built-in support for retrieving files from the Maven 2 Central Repository^[1] and other Maven repositories, and can upload artifacts to specific repositories after a successful build. A local cache of downloaded artifacts acts as the primary means of synchronizing the output of projects on a local system.

Maven is built using a plugin-based architecture that allows it to make use of any application controllable through standard input. Theoretically, this would allow anyone to write plugins to interface with build tools (compilers, unit test tools, etc.) for any other language. In reality, support and use for languages other than Java has been minimal. Currently a plugin for the .Net framework exists and is maintained ^[2], and a C/C++ native plugin was at one time maintained for Maven 1.^[3]

[hide]

[edit] Example

Maven projects are configured using a Project Object Model, which is stored in a pom.xml-file. Here's a minimal example:

<project>
  <!-- model version is always 4.0.0 for Maven 2.x POMs -->
  <modelVersion>4.0.0</modelVersion>
 
  <!-- project coordinates, i.e. a group of values which
       uniquely identify this project -->
 
  <groupId>com.mycompany.app</groupId>
  <artifactId>my-app</artifactId>
  <version>1.0</version>
 
  <!-- library dependencies -->
 
  <dependencies>
    <dependency>
 
      <!-- coordinates of the required library -->
 
      <groupId>junit</groupId>
      <artifactId>junit</artifactId>
      <version>3.8.1</version>
 
      <!-- this dependency is only used for running and compiling tests -->
 
      <scope>test</scope>
 
    </dependency>
  </dependencies>
</project>

This POM only defines a unique identifier for the project (coordinates) and its dependency on the JUnit framework. However, that is already enough for building the project and running the unit tests associated with the project. Maven accomplishes this by embracing the idea of Convention over Configuration, that is, Maven provides good default values for the project's configuration. The directory structure of a normal idiomatic Maven project has the following directory entries:

Directory name

Purpose

project home

Contains the pom.xml and all subdirectories.

src/main/java

Contains the deliverable Java sourcecode for the project.

src/main/resources

Contains the deliverable resources for the project, such as property files.

src/test/java

Contains the testing classes (JUnit or TestNG test cases, for example) for the project.

src/test/resources

Contains resources necessary for testing.

Then the command

mvn package

will compile all the Java files, run any tests, and package the deliverable code and resources into target/my-app-1.0.jar (assuming the artifactId is my-app and the version is 1.0.)

The main idea of Maven is that the user only provides configuration for the project, while the configurable plug-ins do the actual work of compiling the project, cleaning target directories, running unit tests, generating API documentation and so on. In general, users should not have to write plugins themselves. Contrast this with Ant and make in which one writes imperative procedures for doing the aforementioned tasks.

[edit] Concepts

[edit] Project Object Model

Project Object Model provides all the configuration for a single project. General configuration includes the project's name, its owner and its dependencies on other projects. One can also configure individual phases of the build process, which are implemented as plugins. For example, one can configure the compiler-plugin to use Java version 1.5 for compilation, or specify that project can be packaged even if some unit test fails.

Larger projects should be divided into several modules, or sub-projects, each with its own POM. One can then write a root POM through which one can compile all the modules with a single command. POMs can also inherit configuration from other POMs. All POMs inherit from the Super POM^[4] by default. Super POM provides default configuration, such as default source directories, default plugins and so on.

[edit] Plugins

Most of Maven's functionality is in plugins. A plugin provides a set of goals that can be executed using the following syntax:

mvn [plugin-name]:[goal-name]

For example, a Java project can be compiled with the compiler-plugin's compile-goal^[5] by running mvn compiler:compile.

There are Maven plugins for building, testing, source control management, running a web server, generating Eclipse project files, and much more^[6]. Plugins are introduced and configured in a <plugins>-section of a pom.xml file. Some basic plugins are included in every project by default, and they have sensible default settings.

However, it would be cumbersome if one would have to run several goals manually in order to build, test and package a project:

mvn compiler:compile
mvn surefire:test
mvn jar:jar

Maven's lifecycle-concept handles this issue.

[edit] Build Lifecycles

Build lifecycle is a list of named phases that can be used to give order to goal execution. One of Maven's standard lifecycles is the default lifecycle, which includes the following phases, in this order^[7]:

process-resources

compile

process-test-resources

test-compile

test

package

install

deploy

Goals provided by plugins can be associated with different phases of the lifecycle. For example, by default, the goal "compiler:compile" is associated with the compile-phase, while the goal "surefire:test" is associated with the test-phase. When the command

mvn test

is executed, Maven will run all the goals associated with each of the phases up to the test-phase. So it will run the "resources:resources"-goal associated with the process-resources-phase, then "compiler:compile", and so on until it finally runs the "surefire:test"-goal.

Maven also has standard lifecycles for cleaning the project and for generating a project site. If cleaning were part of the default lifecycle, the project would be cleaned every time it was built. This is clearly undesirable, so cleaning has been given its own lifecycle.

Thanks to standard lifecycles, one should be able to build, test and install every Maven-project using the mvn install-command.

[edit] Dependencies

The example-section hinted at Maven's dependency-handling mechanism. A project that needs the Hibernate-library simply has to declare Hibernate's project coordinates in its POM. Maven will automatically download the dependency and all the dependencies that Hibernate itself needs (called transitive dependencies) and store them in the user's local repository. Maven 2 Central Repository^[1] is used by default to search for libraries, but one can configure e.g. company-private repositories in POM.

There are search engines such as mvnrepository, which can be used to find out coordinates for different open-source libraries and frameworks.

Projects developed on a single machine can depend on each other through the local repository. The local repository is a simple folder structure which acts both as a cache for downloaded dependencies and as a centralized storage place for locally built artifacts. The Maven command mvn install builds a project and places its binaries in the local repository. Then other projects can utilize this project by specifying its coordinates in their POMs.

[edit] IDE Integration

Add-ons to several popular IDEs exist to provide integration of Maven with the IDE's build mechanism and source editing tools, allowing Maven to compile projects from within the IDE, and also to set the classpath for code completion, highlighting compiler errors, etc. Examples of popular IDEs supporting development with Maven include:

Eclipse

Netbeans

IntelliJ

JBuilder

These add-ons also provide the ability to edit the POM or use the POM to determine a project's complete set of dependencies directly within the IDE.

Some built-in features of IDEs are forfeited when the IDE no longer performs compilation. For example, Eclipse's JDT has the ability to recompile a single java source file after it has been edited. Many IDEs work with a flat set of projects instead of the hierarchy of folders preferred by Maven. This complicates the use of SCM systems in IDEs when using Maven. ^[8] ^[9] ^[10]

[edit] History

Maven, created by Sonatype's Jason van Zyl, began as a subproject of Apache Turbine in 2002. In 2003, it was voted on and accepted as a top level Apache Software Foundation project. In July of 2004, Maven was released as the critical first milestone, v1.0. Maven 2 was declared v1.0 in October of 2005 after about 6 months in beta cycles.

[edit] Future

Maven 3.0 information began trickling out in 2008. After eight alpha releases, the first beta version of Maven 3.0 was released in April 2010. Maven 3.0 has reworked the core Project Builder infrastructure such that the POMs file-based representation is now decoupled from its in-memory object representation. This has expanded the possibility for Maven 3.0 add-ons to leverage non-XML based project definition files. Languages suggested include Ruby (already in private prototype by Jason van Zyl), YAML, and Groovy. Experimental work for a YAML-based POM definition file (requires an external conversion script to be executed) has been piloted by Don Brown of Atlassian.

Special attention has been paid to ensuring compatibility between Maven 2 and 3. For most projects, an upgrade to Maven 3 won't require any adjustments of their project structure. The first beta of Maven 3 saw the introduction of a parallel build feature which leverages a configurable amount of cores on a multi-core machine and is especially suited for large multi-module projects.