/clojure-maven-plugin

Apache Maven Mojo for compiling clojure scripts to class files

Primary LanguageJava

Welcome to the clojure-maven-plugin plugin for Apache Maven 2.

This plugin has been designed to make working with clojure as easy as possible, when working in a mixed language, enterprise project.

Available goals

  • clojure:add-source
  • clojure:add-test-source
  • clojure:compile
  • clojure:test
  • clojure:test-with-junit
  • clojure:run
  • clojure:repl
  • clojure:nrepl
  • clojure:swank
  • clojure:nailgun
  • clojure:gendoc
  • clojure:autodoc
  • clojure:marginalia

Getting started with Clojure and Maven

To use this plugin and start working with clojure, start with a blank maven project and declare the plugin and add a dependency on clojure:

<packaging>clojure</packaging>
....
<build>
  <plugins>
    <plugin>
      <groupId>com.theoryinpractise</groupId>
      <artifactId>clojure-maven-plugin</artifactId>
      <version>1.8.3</version>
      <extensions>true</extensions>
    </plugin>
  </plugins>
</build>
....
<dependencies>
  <dependency>
    <groupId>org.clojure</groupId>
    <artifactId>clojure</artifactId>
    <version>1.12.0</version>
  </dependency>
</dependencies>

By changing your projects type to clojure, the plugin will automatically bind itself to the compile, test-compile, and test maven phases.

Without any additional configuration, the clojure-maven-plugin will compile any namespaces in ./src/main/clojure/.clj (or .cljc) and ./src/test/clojure/.clj (or .cljc).

Adding additional source directories

To change, or add additional source directories you can add the following configuration:

<configuration>
  <sourceDirectories>
    <sourceDirectory>src/main/clojure</sourceDirectory>
  </sourceDirectories>
  <testSourceDirectories>
    <testSourceDirectory>src/test/clojure</testSourceDirectory>
  </testSourceDirectories>
</configuration>

NOTE: The plugin will prepend the project's ${basedir} before each source/testSource directory specified.

Temporary Compile Paths

If you wish to take advantage of the compilers syntax checking, but wish to prevent any AOT classes from appearing in the maven generated JAR file, you can tell the plugin to compile to a temporary directory:

<configuration>
  <temporaryOutputDirectory>true</temporaryOutputDirectory>
</configuration>

Namespace configuration

If you wish to limit or filter out namespaces during your compile/test, simply add a <namespaces> or <testNamespaces> configuration section:

<configuration>
  <namespaces>
    <namespace>com.foo</namespace>
    <namespace>net.*</namespace>
    <namespace>!testing.*</namespace>
  </namespaces>
</configuration>

The namespace declaration is actually a regex match against discovered namespaces, and can also be prepended with an ! to filter the matching namespace.

If you wish to further limit test/compile usage to only the namespaces you define, you can enable this with the configuration block:

<configuration>
  <compileDeclaredNamespaceOnly>true</compileDeclaredNamespaceOnly>
  <testDeclaredNamespaceOnly>true</testDeclaredNamespaceOnly>
</configuration>

If you want that only compiled artifacts related to the above mentioned <namespaces> and <compileDeclaredNamespaceOnly> be kept, then add a cleanAOTNamespaces config param and set it to true.

For instance (1/2), with the following configuration ...

<configuration>
  <cleanAOTNamespaces>true</cleanAOTNamespaces>
  <namespaces>
    <namespace>!some.annoying.namespace</namespace>
  </namespaces>
<configuration>

... all aot-compiled classes created in the output directory will be kept as is, but the ones of the some.annoying.namespace namespace which will be deleted.

For instance (2/2), with the following configuration ...

<configuration>
  <cleanAOTNamespaces>true</cleanAOTNamespaces>
  <namespaces>
    <namespace>some.namespace.with.a.gen-class</namespace>
  </namespaces>
  <compileDeclaredNamespaceOnly>true</compileDeclaredNamespaceOnly>
<configuration>

... all aot-compiled classes will be deleted, but the ones of the some.annoying.namespace namespace.

Interactive Coding

The plugin supports several goals intended to make it easier for developers to run interactive clojure shells in the context of maven projects. This means that all dependencies in a project's runtime and test scopes will be automatically added to the classpath and available for experimentation.

By default these goals will use the test classpath, if you wish to only use the compile classpath/dependencies, you can disable this with:

<configuration>
  <runWithTests>false</runWithTests>
</configuration>

or by running maven with:

-Dclojure.runwith.test=false

Goals

Goal Description
clojure:repl

Starts an interactive clojure REPL right on the command line.

Property Variable Default Description
replScript An initialization script can be specified in the pom using the replScript configuration element.
clojure:nrepl

Starts a nREPL server that accepts connections.

Property Variable Default Description
replScript The clojure script to run before starting the repl
port clojure.nrepl.port 4005 The nREPL server port
nreplHost clojure.nrepl.host localhost The host to bind the nREPL server to/td>
nreplHandler clojure.nrepl.handler The nREPL Handler to use. i.e.
cider.nrepl/cider-nrepl-handler
from the
cider-nepl
project./td>
clojure:swank

Starts a Swank server that accepts connections.

Property Variable Default Description
replScript The clojure script to run before starting the repl
port clojure.swank.port 4005 The swank server port
protocolVersion clojure.swank.protocolVersion 2009-09-14 The swank protocol version
encoding clojure.swank.encoding iso-8859-1 The swank encoding to use
swankHost clojure.swank.host localhost The host to bind the swank server to/td>
clojure:nailgun

Starts a nailgun server.

Property Variable Default Description
replScript The clojure script to run before starting the repl
port clojure.nailgun.port 2113 The nailgun server port
clojure:run

Runs a clojure script.

Property Variable Default Description
script clojure.script The clojure script to run
scripts A list of clojure scripts to run
mainClass clojure.mainClass A java class to run
args clojure.args Arguments to the clojure script(s)
clojure:add-source Includes clojure source directory in -sources.jar.
clojure:add-test-source Includes clojure test source directory in -testsources.jar.

Testing Clojure Code

Whilst you could easily launch your tests from the clojure:run goal, the plugin provides two goals targeted specifically to testing: clojure:test and clojure:test-with-junit

Without any additional configuration the plugin will run a temporary clojure "test launcher" script:

The script runs all discovered test namespaces, and fails the build when any FAIL or ERROR cases are found.

If you require different test behavior, you can provide your own test script with the following configuration:

<configuration>
  <testScript>src/test/clojure/com/jobsheet/test.clj</testScript>
</configuration>

The first argument to the script is the name of a properties file that has in it a config for the user selected. These configs can be parsed out using the following code

(def props (Properties.))
(.load props (FileInputStream. (first *command-line-args*)))

;;namespaces to run tests for
(def namespaces  (into []
                       (for [[key val] props
                             :when (.startsWith key "ns.")]
                               (symbol val))))

;; should there be junit compatible output or not
(def junit (Boolean/valueOf (.get props "junit")))
;; what is the output directory that results should be written to
(def output-dir (.get props "outputDir"))
;; should we xml-escape *out* while the tests are running
(def xml-escape (Boolean/valueOf (.get props "xmlEscape")))

We reserve the right to add new configs in the future, and possibly new command line arguments as well.

Configuring your clojure session

If you want to provide additional arguments to all spawned java/clojure processes, the plugin provides several configuration properties:

Property Variable Default Description
vmargs clojure.vmargs JVM Arguments
clojureOptions Additional JVM Options such as system property definitions
warnOnReflection false Enable reflection warnings
prependClasses A list of classnames to prepend to the command line before the mainClass

The plugin can also copy source files to the output directory, filtered using the namespace mechanism that is used to control compilation. If you want to copy all compiled source files to the output:

<configuration>
  <copyAllCompiledNamespaces>true</copyAllCompiledNamespaces>
<configuration>

If you want to copy only a subset:

<configuration>
  <copiedNamespaces>
    <namespace>com.foo</namespace>
    <namespace>!com.foo.private.*</namespace>
  </copiedNamespaces>
  <copyDeclaredNamespaceOnly>true</copyDeclaredNamespaceOnly>
<configuration>

If you want to do no compilation at all, but copy all source files:

<configuration>
  <copyDeclaredNamespaceOnly>true</copyDeclaredNamespaceOnly>
  <namespaces>
    <namespace>!.*</namespace>
  </namespaces>
  <compileDeclaredNamespaceOnly>true</compileDeclaredNamespaceOnly>
<configuration>

Note that it will only copy clojure source files, which must a) end in .clj or .cljc and b) contain a namespace declaration.

Enjoy.

Dependencies

In order to run clojure:repl, clojure:swank or clojure:nailgun, your project needs to have a recent (1.0 or later) version of clojure as a dependency in pom.xml.

In order to run clojure:autodoc, your project needs to have autodoc as a dependency in pom.xml.

In order to run clojure:nrepl, your project needs to have org.clojure/tools.nrepl as a dependency in pom.xml.

JLine/IClojure/REPL-y

If JLine is detected in the classpath, it will be used to provide the clojure:repl goal with history, tab completion, etc. A simple way of enabling this is to put the following in your pom.xml:

    <dependency>
       <groupId>jline</groupId>
       <artifactId>jline</artifactId>
       <version>0.9.94</version>
    </dependency>

If you prefer IClojure you can add:

    <dependency>
       <groupId>com.offbytwo.iclojure</groupId>
       <artifactId>iclojure</artifactId>
       <version>1.1.0</version>
    </dependency>

Or REPL-y:

    <dependency>
       <groupId>reply</groupId>
       <artifactId>reply</artifactId>
       <version>0.1.0-beta9</version>
    </dependency>

Swank

The clojure:swank goal requires swank-clojure as a project dependency. Unfortunately, this library is currently not available in the central maven repository, but is available from clojars by first declaring the repository:

<repositories>
  <repository>
    <id>clojars</id>
    <url>https://clojars.org/repo/</url>
  </repository>
</repositories>

and then declaring the dependency itself:

<dependency>
  <groupId>swank-clojure</groupId>
  <artifactId>swank-clojure</artifactId>
  <version>1.3.0-SNAPSHOT</version>
</dependency>

By default the swank process will run against the local loopback device, if you wish to change the host your swank server runs against, you can configure it via:

<configuration>
  <swankHost>localhost</swankHost>
</configuration>

or by defining the clojure.swank.host system property.

nREPL

The clojure:nrepl goal requires nrepl as a project dependency as:

<dependency>
  <groupId>nrepl</groupId>
  <artifactId>nrepl</artifactId>
  <version>0.6.0</version>
</dependency>

By default the nREPL process will run against the local loopback device on port 4005, if you wish to change the host your nREPL server runs against or the port, you can configure it via:

<configuration>
  <nreplHost>localhost</nreplHost>
  <port>9001</port>
</configuration>

or by defining the clojure.nrepl.host and clojure.nrepl.port system properties.

It is also possible to specify a custom handler or server-side middleware to be added to the nREPL stack. This may be necessary for integrating with Clojure IDEs, such as LightTable or CIDER. These IDEs require custom nREPL middleware for best results or may not work at all with the default nREPL stack. nREPL middleware can be specified as follows:

<configuration>
    <nreplMiddlewares>
        <middleware>my.nrepl.middleware/my-middleware</middleware>
    </nreplMiddlewares>
</configuration>

Either a custom handler or as many middleware as necessary can be specified. Each middleware must be specified as a fully qualified symbol - namespace/name. Thy symbol must resolve to a var referencing a middleware function. If the middleware is not part of the project itself, it must be specified as a dependency. The same is true for custom nRepl handlers

LightTable configuration example:

<dependency>
    <groupId>lein-light-nrepl</groupId>
    <artifactId>lein-light-nrepl</artifactId>
    <version>0.3.3</version>
    <scope>test</scope>
</dependency>
...
<configuration>
    <nreplMiddlewares>
        <middleware>lighttable.nrepl.handler/lighttable-ops</middleware>
    </nreplMiddlewares>
</configuration>

CIDER configuration example:

<dependency>
    <groupId>cider</groupId>
    <artifactId>cider-nrepl</artifactId>
    <version>0.12.0</version>
    <scope>test</scope>
</dependency>
.......
<configuration>
        <nreplMiddlewares>
            <middleware>cider.nrepl/wrap-apropos</middleware>
            <middleware>cider.nrepl/wrap-classpath</middleware>
            <middleware>cider.nrepl/wrap-complete</middleware>
            <middleware>cider.nrepl/wrap-debug</middleware>
            <middleware>cider.nrepl/wrap-format</middleware>
            <middleware>cider.nrepl/wrap-info</middleware>
            <middleware>cider.nrepl/wrap-inspect</middleware>
            <middleware>cider.nrepl/wrap-macroexpand</middleware>
            <middleware>cider.nrepl/wrap-ns</middleware>
            <middleware>cider.nrepl/wrap-spec</middleware>
            <middleware>cider.nrepl/wrap-profile</middleware>
            <middleware>cider.nrepl/wrap-refresh</middleware>
            <middleware>cider.nrepl/wrap-resource</middleware>
            <middleware>cider.nrepl/wrap-stacktrace</middleware>
            <middleware>cider.nrepl/wrap-test</middleware>
            <middleware>cider.nrepl/wrap-trace</middleware>
            <middleware>cider.nrepl/wrap-out</middleware>
            <middleware>cider.nrepl/wrap-undef</middleware>
            <middleware>cider.nrepl/wrap-version</middleware>
            <middleware>cider.nrepl/wrap-xref</middleware>
        </nreplMiddlewares>
</configuration>

Nailgun for Vimclojure < 2.2.0

The clojure:nailgun goal requires a recent version of vimclojure as a dependency. Unfortunately, this library is currently not available in the central maven repository, and has to be downloaded and installed manually:

  1. Download vimclojure source code from http://cloud.github.com/downloads/jochu/swank-clojure/swank-clojure-1.0-SNAPSHOT-distribution.zip.
  2. Follow the README to compile and install vimclojure.
  • Locate vimclojure.jar and run the following command to install it to your local repository (replace X.X.X with your version of vimclojure):

    mvn install:install-file -DgroupId=de.kotka -DartifactId=vimclojure -Dversion=X.X.X -Dpackaging=jar -Dfile=/path/to/jarfile
    
  • Put the following in your pom.xml (replace X.X.X with your version of vimclojure)

    <dependency>
    <groupId>de.kotka</groupId>
    <artifactId>vimclojure</artifactId>
    <version>X.X.X</version>
    </dependency>
    
  • You will need to run mvn clojure:nailgun -Dclojure.nailgun.server=com.martiansoftware.nailgun.NGServer in order to

    work with the old version (pre 2.2.0) of vimclojure.

Nailgun for Vimclojure >= 2.2.0

To use clojure 1.2.0 comfortably, you will need to upgrade to Vimclojure 2.2.0 which isn't backwards compatible with previous vimclojure versions. Now you will need a dependency on the vimclojure:server:2.2.0 which contains the modified Nailgun server.

<dependency>
    <groupId>vimclojure</groupId>
    <artifactId>server</artifactId>
    <version>2.2.0</version>
</dependency>

The jar can be found in clojars maven repo (you'll have to add it to the repositories section)

<repository>
    <id>clojars</id>
    <name>Clojars</name>
    <url>https://clojars.org/repo/</url>
</repository>

The installation process for vimclojure remains the same (except for the vimclojure.jar which you don't need to install anymore). Just get the vimclojure package from http://kotka.de/projects/clojure/vimclojure.html and follow the README.

Notes for migration from the previous version of vimclojure:

  • clj_highlight_builtins was deprecated in favor of vimclojure#HighlightBuiltins
  • clj_highlight_contrib was removed
  • g:clj_paren_rainbow was deprecated in favor of vimclojure#ParenRainbow
  • g:clj_want_gorilla was deprecated in favor of vimclojure#WantNailgun

Windows configuration

As the default Windows console doesn't allow to easily copy and paste code, you can use the windowsConsole configuration option to specify which console command to run in Windows. For example if you are using http://code.google.com/p/conemu-maximus5/, you can configure the plugin with:

<windowsConsole>"C:\\Program Files\\ConEmu\\ConEmu64.exe" /cmd</windowsConsole>

which will give you a sane Windows console

Configuration

The following options that can be configured as system properties:

Property Default value Description
clojure.nailgun.port 4005 Only applicable for the clojure:nailgun goal. The port number that the Nailgun server should listen to.
clojure.swank.port 4005 Only applicable for the clojure:swank goal. The port number that the Swank server should listen to.
clojure.swank.protocolVersion 2009-09-14 Only applicable for the clojure:swank goal. Specifies the version of the swank protocol.
clojure.swank.encoding iso-8859-1 Only applicable for the clojure:swank goal. Specifies the encoding used by the swank protocol.
clojure.nrepl.port 4005 Only applicable for the clojure:nrepl goal. The port number that the nREPL should listen to.
clojure.nrepl.host 4005 Only applicable for the clojure:nrepl goal. The host that the nREPL should listen to.

Support

Join the discussion mailing list at:

http://groups.google.com/group/clojure-maven-plugin