Chapter 3: Logback configuration & Joran

Joran stands for a cold north-west wind which, every now and then, blows forcefully on Lake Geneva. Located right in the middle of Europe, the Geneva lake happens to be the continent's largest sweet water reserve.

In the first part, we start by presenting ways for configuring logback, with many example configuration scripts. In the second part, we present Joran, a generic configuration framework, that you can put into use in order to configure your own applications.

Configuration in logback

Inserting log requests into the application code requires a fair amount of planning and effort. Observation shows that approximately four percent of code is dedicated to logging. Consequently, even moderately sized applications will contain thousands of logging statements embedded within its code. Given their number, we need powerful tools to manage these log statements.

Logback can be configured either programmatically or with configuration a script (expressed in XML format). By the way, existing log4j users can convert their log4j.properties files to logback.xml using our PropertiesTranslator web-application.

Let us begin by discussing the initialization steps that logback follows to try to configure itself:

1. Logback tries to find a file called logback-test.xml in the classpath.

2. If no such file is found, it checks for the file logback.xml.

3. In case neither file is found, logback configures itself automatically using the BasicConfigurator which will cause logging output to be directed on the console.

The third and last step is meant to provide a default (but very basic) logging functionnality in the absence of a configuration file.

Assuming the logback-test.xml file is placed under src/test/resources folder, Maven will ensure that it won't be included in the artifact produced. Thus, you can use a different configuration file, namely logback-test.xml during testing, and another file, namely, logback.xml, in production. The same principle could apply by analogy for Ant.

Automatically configuring logback

The simplest way to configure logback is by letting logback fallback to its default configuration. Let us give a taste of how this is done in an imaginary application called MyApp1.

package chapter3;

import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

public class MyApp1 {
  final static Logger logger = LoggerFactory.getLogger(MyApp1.class);

  public static void main(String[] args) {
    logger.info("Entering application.");

    Foo foo = new Foo();
    foo.doIt();
    logger.info("Exiting application.");
  }
}

This class defines a static logger variable. It then instantiates a Foo object. The Foo class is listed below:

package chapter3;
  
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
   
public class Foo {
  static final Logger logger = LoggerFactory.getLogger(Foo.class);
  
  public void doIt() {
    logger.debug("Did it again!");
  }
}

In order to run the examples in this chapter, you need to make sure that certain jar files are present on the classpath. Please refer to the setup page for further details.

Assuming the configuration files logback-test.xml or logback.xml could not be found, logback will default to a minimal configuration hardwired to attaching a ConsoleAppender to the root logger. The output is formatted using a PatternLayout set to the pattern %d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n. Moreover, by default the root logger is assigned to the DEBUG level.

Thus, the output of the command java chapter3.MyApp1 should be similar to:

16:06:09.031 [main] INFO  chapter3.MyApp1 - Entering application.
16:06:09.046 [main] DEBUG chapter3.Foo - Did it again!
16:06:09.046 [main] INFO  chapter3.MyApp1 - Exiting application.

The MyApp1 application links to logback through calls org.slf4j.LoggerFactory and org.slf4j.Logger classes, retrieve the loggers it wishes to use, and chugs on. Note that the only dependence of the Foo class on logback are through org.slf4j.LoggerFactory and org.slf4j.Logger imports. Except code that configures logback (if such code exists) client code does not need to depend on logback. Given that SLF4J permits the use of any implementation under its abstraction layer, it is rather easy to migrate large bodies of code from one logging system to another.

The same with JoranConfigurator

The previous example outputs logging information always in the same fixed manner. How boring! It takes few easy steps to modify MyApp1 so that the log output can be controlled at runtime. Here is a slightly modified application called MyApp2.

Example 3.2: Simple example of BasicConfigurator usage (logback-examples/src/main/java/chapter3/MyApp2.java)

package chapter3;

import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import ch.qos.logback.classic.LoggerContext;
import ch.qos.logback.classic.joran.JoranConfigurator;
import ch.qos.logback.core.joran.spi.JoranException;
import ch.qos.logback.core.util.StatusPrinter;

public class MyApp2 {
  final static Logger logger = LoggerFactory.getLogger(MyApp2.class);

  public static void main(String[] args) {
    // assume logback is in use
    LoggerContext lc = (LoggerContext) LoggerFactory.getILoggerFactory();
    
    try {
      JoranConfigurator configurator = new JoranConfigurator();
      configurator.setContext(lc);
      lc.shutdownAndReset();
      configurator.doConfigure(args[0]);
    } catch (JoranException je) {
      StatusPrinter.print(lc);
    }

    logger.info("Entering application.");

    Foo foo = new Foo();
    foo.doIt();
    logger.info("Exiting application.");
  }
}

MyApp2 fetches the LoggerContext currently in effect, creates a new JoranConfigurator, gives it the context and finally asks that the configurator parses a configuration file. A basic configuration file that creates the same logging environment as would the default configuration is listed below:

Example 3.3: Basic configuration file (logback-examples/src/main/java/chapter3/sample0.xml)

<configuration>

  <appender name="STDOUT"
    class="ch.qos.logback.core.ConsoleAppender">
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>%d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n</Pattern>
    </layout>
  </appender>

  <root>
    <level value="debug" />
    <appender-ref ref="STDOUT" />
  </root>
</configuration>

The above configuration file directs logback to create an ConsoleAppender named STDOUT whose output is format is managed by a PatternLayout. The root logger is configued by setting its level to DEBUG and attaching it the previously created ConsoleAppender.

Note we could have omitted setting the level of the root logger to DEBUG because that is its default value.

Assuming the current directory is logback-examples, try running the MyApp2 class by issuing the following command:

java chapter3.MyApp2 src/main/java/chapter3/sample0.xml

The ouput of this command is very similar to the output of the previous example, except that MyApp2 retrieves a logger called chapter3.MyApp2 instead of chapter3.MyApp1. The output will reflect the difference.

16:09:00.593 [main] INFO  chapter3.MyApp2 - Entering application.
16:09:00.593 [main] DEBUG chapter3.Foo - Did it again!
16:09:00.593 [main] INFO  chapter3.MyApp2 - Exiting application.

It is often very useful to define the logback debug configuration property in order to instruct logback to output internal configuration messages on the console. To achieve this, one only needs to add an attribute to the main configuration element in the configuration file, as shown above: wec

<configuration debug="true">

  <appender name="STDOUT"
    class="ch.qos.logback.core.ConsoleAppender">
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>%d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n</Pattern>
    </layout>
  </appender>

  <root>
    <level value="debug" />
    <appender-ref ref="STDOUT" />
  </root>
</configuration>

This should cause logback to print internal configuration messages in addition to the actual logs. Relaunching the MyApp2 application with this new configuration file will ouput the following lines:

|-INFO in ch.qos.logback.core.joran.action.AppenderAction - About to instantiate appender of type [ch. \
qos.logback.core.ConsoleAppender]
|-INFO in ch.qos.logback.core.joran.action.AppenderAction - Naming appender as [STDOUT]
|-INFO in ch.qos.logback.core.joran.action.AppenderAction - Popping appender named [STDOUT] from the \
object stack
|-INFO in ch.qos.logback.classic.joran.action.LevelAction - root level set to DEBUG
|-INFO in ch.qos.logback.core.joran.action.AppenderRefAction - Attaching appender named [STDOUT] to \
Logger[root]
|-INFO in ch.qos.logback.classic.joran.action.ConfigurationAction - End of configuration.
16:18:23.687 [main] INFO  chapter3.MyApp2 - Entering application.
16:18:23.687 [main] DEBUG chapter3.Foo - Did it again!
16:18:23.687 [main] INFO  chapter3.MyApp2 - Exiting application.

At the end of this output, you can recognize the lines that were printed in the previous example. You should also notive the logback's internal messages based on Status objects, logback internal reporting mechanism. Status objects allow easy access to logback's internal state.

XML Syntax

To obtain these different logging behaviors we do not need to recompile code. You can easily configure logging so that output is directed to a UNIX Syslog daemon, or to a log visualizer, or forward logging events to a remote logback server, which would log according to local server policy, for example by forwarding the log event to a second logback server.

The remainder of this chapter presents the syntax of the directives allowing all these variations.

Configuring Loggers

Loggers are configured using logger elements. A logger element takes exactly one mandatory name atttribute and an optional additivity attribute, which takes values true or false. The logger element admits at most one level element which is discussed next. It may also contain zero or more appender-ref elements; each appender thus referenced is added to the named logger. It is important to keep mind that each named logger that is declared with a logger element first has all its appenders removed and only then are the referenced appenders attached to it. In particular, if there are no appender references, then the named logger will lose all its appenders.

The level element is used to set logger levels. It admits two attributes value and class. The value attribute can be one of the strings DEBUG, INFO, WARN ERROR, ALL or OFF. The special case-insensitive value INHERITED, or its synonym NULL, will force the level of the logger to be inherited from higher up in the hierarchy. Note that the level of the root logger cannot be inherited. If you set the level of a logger and later decide that it should inherit its level, then you need to specify INHERITED or its synonym NULL as the level value. The class attribute allows you to specify a custom level where the value of the attribute is the fully qualified name of a custom level class. You may alternatively use the level#classname syntax within the value attribute. The level element has no children.

The root element configures the root logger. It does not admit any attributes because the additivity flag does not apply to the root logger. Moreover, since the root logger cannot be named, it does not admit a name attribute either. The root element admits at most one level element and zero or more appender-ref elements. Similar to the logger element, declaring a root element will have the effect of first closing and then detaching all its current appenders and only subsequently will referenced appenders, if any, will be added. In particular, if it has no appender references, then the root logger will lose all its appenders.

Setting the level of a logger is as simple as declaring it and setting its level, as the next example illustrates. Suppose we are no longer interested in seeing any DEBUG level logs from any component belonging to the chapter3 package. The following configuration file shows how to achieve that.

<configuration>

  <appender name="STDOUT"
    class="ch.qos.logback.core.ConsoleAppender">
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>%d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n</Pattern>
    </layout>
  </appender>

  <logger name="chapter3">
    <level value="INFO" />
  </logger>
  
  <root>		
    <!-- The following level element is not necessary since the -->
    <!-- level of the root level is set to DEBUG by default.    -->
    <level value="DEBUG" />
    <appender-ref ref="STDOUT" />
  </root>  
  
</configuration>

This new configuration will yield the following output, when used with the same chapter3.MyApp2 class.

17:34:07.578 [main] INFO  chapter3.MyApp2 - Entering application.
17:34:07.578 [main] INFO  chapter3.MyApp2 - Exiting application.

Obviously, you can configure the levels of as many loggers as you wish. In the next configuration file we set the level of the chapter3 logger to INFO but at the same time set the level of the chapter3.Foo logger to DEBUG.

<configuration>

  <appender name="STDOUT"
    class="ch.qos.logback.core.ConsoleAppender">
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>
        %d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n
     </Pattern>
    </layout>
  </appender>

  <logger name="chapter3">
    <level value="INFO" />
  </logger>

  <logger name="chapter3.Foo">
    <level value="DEBUG" />
  </logger>

  <root>
    <!-- The following level element is not necessary since the -->
    <!-- level of the root level is set to DEBUG by default.    -->
    <level value="DEBUG" />
    <appender-ref ref="STDOUT" />
  </root>

</configuration>

Running MyApp2 with this configuration file will result in the following output on the console:

17:39:27.593 [main] INFO  chapter3.MyApp2 - Entering application.
17:39:27.593 [main] DEBUG chapter3.Foo - Did it again!
17:39:27.593 [main] INFO  chapter3.MyApp2 - Exiting application.

After JoranConfigurator configures logback using the sample3.xml file, the logger settings, more specifically their levels, are summarized in the following table.

It follows that the two logging statements of level INFO in the MyApp2 class are enabled while the debug statement in Foo.doIt() method will also print without hindrance. Note that the level of the root logger is always set to a non-null value, which is DEBUG by default. One rather important point to remember is that the logger-level filter depends on the effective level of the logger being invoked, not the level of the logger where the appenders are attached. The configuration file sample4.xml is a case in point:

<configuration>

  <appender name="STDOUT"
   class="ch.qos.logback.core.ConsoleAppender">
   <layout class="ch.qos.logback.classic.PatternLayout">
     <Pattern>
        %d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n
      </Pattern>
    </layout>
  </appender>

  <logger name="chapter3">
    <level value="INFO" />
  </logger>

  <root>
    <level value="OFF" />
    <appender-ref ref="STDOUT" />
  </root>

</configuration>

The following table lists the loggers and their level setting after applying the sample4.xml configuration file.

The ConsoleAppender named STDOUT, the only configured appender in sample4.xml, is attached to the root logger whose level is set to OFF. However, running MyApp2 with configuration script sample4.xml will output:

17:52:23.609 [main] INFO  chapter3.MyApp2 - Entering application.
17:52:23.609 [main] INFO  chapter3.MyApp2 - Exiting application.

Thus, the level of the root logger has no apparent effect because the loggers in chapter3.MyApp2 and chapter3.Foo classes, namely chapter3.MyApp2 and chapter3.Foo, inherit their level from the chapter3 logger which has its level set to INFO. As noted previously, the chapter3 logger exists by virtue of its declaration in the configuration file - even if the Java source code does not directly refer to it.

Configuring Appenders

Appenders are configured using appender elements. These elements admit two attributes name and class both of which are mandatory. The name attribute specifies the name of the appender whereas the class attribute specifies the fully qualified name of the class of which the named appender will be an instance. The appender may contain zero or one layout elements and zero or more filter elements. Appart from these two basic elements, appender elements may contain any element that corresponds to a setter method of the appender class, to configure the appender's options.

The layout element takes a mandatory class attribute specifying the fully qualified name of the class of which the associated layout should be an instance. Like the appender element, it may contain other elements, referring to setter methods, to configure its options.

Logging to multiple appenders is as easy as defining the various appenders and referencing them in a logger, as the next configuration file illustrates:

<configuration>

  <appender name="FILE"
    class="ch.qos.logback.core.rolling.RollingFileAppender">
    <file>myApp.log</file>

    <rollingPolicy
      class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
      <FileNamePattern>
        myApp-%d{yyyy-MM-dd-HH-mm-ss}.log
      </FileNamePattern>
    </rollingPolicy>

    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>
        %date %level [%thread] %logger{10} [%file : %line] %msg%n
      </Pattern>
    </layout>
  </appender>

  <appender name="STDOUT"
    class="ch.qos.logback.core.ConsoleAppender">
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>
        %msg%n
      </Pattern>
    </layout>
  </appender>

  <root>
    <level value="debug" />
    <appender-ref ref="FILE" />
    <appender-ref ref="STDOUT" />
  </root>
</configuration>

This configuration scripts defines two appenders called FILE and STDOUT. The FILE appender logs to a file called myApp.log. The layout for this appender is a PatternLayout that outputs the date, level, thread name, logger name, file name and line number where the log request is located, the message and line separator character(s). The second appender called STDOUT outputs to the console. The layout for this appender outputs only the message string followed by a line separator.

The appenders are attached to the root logger by referencing them by name within an appender-ref element. Note that each appender has its own layout. Layouts are usually not designed to be shared by multiple appenders. XML configuration files do not provide any syntactical means for sharing layouts.

By default, appenders are cumulative: a logger will log to the appenders attached to itself (if any) as well as all the appenders attached to its ancestors. Thus, attaching the same appender to multiple loggers will cause logging output to be duplicated.

<configuration>

  <appender name="STDOUT"
    class="ch.qos.logback.core.ConsoleAppender">
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>
        %d{HH:mm:ss.SSS} [%thread] %-5level %logger{36} - %msg%n
      </Pattern>
    </layout>
  </appender>

  <logger name="chapter3">
    <appender-ref ref="STDOUT" />
  </logger>

  <root>
    <level value="debug" />
    <appender-ref ref="STDOUT" />
  </root>
</configuration>

Running MyApp2 with duplicate.xml will yield the following output:

14:25:36.343 [main] INFO  chapter3.MyApp2 - Entering application.
14:25:36.343 [main] INFO  chapter3.MyApp2 - Entering application.
14:25:36.359 [main] DEBUG chapter3.Foo - Did it again!
14:25:36.359 [main] DEBUG chapter3.Foo - Did it again!
14:25:36.359 [main] INFO  chapter3.MyApp2 - Exiting application.
14:25:36.359 [main] INFO  chapter3.MyApp2 - Exiting application.

Notice the duplicated output. The appender named STDOUT is attached to two loggers, to root and to chapter3. Since the root logger is the ancestor of all loggers and chapter3 is the parent of chapter3.MyApp2 and chapter3.Foo, logging request made with these two loggers will be output twice, once because STDOUT is attached to chapter3 and once because it is attached to root.

Appender additivity is not intended as a trap for new users. It is a quite convenient logback feature. For instance, you can configure logging such that log messages appear on the console (for all loggers in the system) while messages only from some specific set of loggers flow into a specific appender.

<configuration>

  <appender name="FILE"
    class="ch.qos.logback.core.rolling.RollingFileAppender">
    <file>myApp.log</file>
    <rollingPolicy
      class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
      <FileNamePattern>
        myApp-%d{yyyy-MM-dd-HH-mm-ss}.log
      </FileNamePattern>
    </rollingPolicy>
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>
        %date %level [%thread] %logger{10} [%file : %line] %msg%n
      </Pattern>
    </layout>
  </appender>

  <appender name="STDOUT"
    class="ch.qos.logback.core.ConsoleAppender">
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>%msg%n</Pattern>
    </layout>
  </appender>

  <logger name="chapter3">
    <appender-ref ref="FILE" />
  </logger>

  <root>
    <level value="debug" />
    <appender-ref ref="STDOUT" />
  </root>
</configuration>

In this example, the console appender will log all the messages (for all loggers in the system) whereas only logs under the chapter3 tree go into the myApp.log file.

Overriding the default cumulative behaviour

In case the default cumulative behavior turns out to be unsuitable for one's needs, one can override it by setting the additivity flag to false. Thus, a branch in your logger tree may direct output to a set of appenders different than those of the rest of the tree.

<configuration>

  <appender name="FILE"
    class="ch.qos.logback.core.rolling.RollingFileAppender">
    <file>foo.log</file>
    <rollingPolicy
      class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
      <FileNamePattern>
        foo-%d{yyyy-MM-dd-HH-mm-ss}.log
      </FileNamePattern>
    </rollingPolicy>
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>
        %date %level [%thread] %logger{10} [%file : %line] %msg%n
      </Pattern>
    </layout>
  </appender>

  <appender name="STDOUT"
    class="ch.qos.logback.core.ConsoleAppender">
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>%msg%n</Pattern>
    </layout>
  </appender>

  <logger name="chapter3.Foo" additivity="false">
    <appender-ref ref="FILE" />
  </logger>

  <root>
    <level value="debug" />
    <appender-ref ref="STDOUT" />
  </root>
</configuration>

This example, the appender named FILE is attached to the chapter3.Foo logger. Moreover, the chapter3.Foo logger has its additivity flag set to false such that its logging output will be sent to the appender named FILE but not to any appender attached higher in the hierarchy. Other loggers remain oblivious to the additivity setting of the chapter3.Foo logger. Running the MyApp2 application with the additivityFlag.xml configuration file will output results on the console from the chapter3.MyApp2 logger. However, output from the chapter3.Foo logger will appear in the foo.log file and only in that file.

Variable substitution

All option values admit variable substitution. The syntax of variable substitution is similar to that of Unix shells. The string between an opening ${ and closing } is interpreted as a key. The value of the substituted variable can be defined as a system property in the configuration file itself or in a separate file linked to the configuration file. The value of the key is first searched in configuration file or linked properties file, and if not found there, it is then searched in system properties. The corresponding value replaces ${aKey} sequence. For example, if java.home.dir system property is set to /home/xyz, then every occurrence of the sequence ${java.home.dir} will be interpreted as /home/xyz.

The first example shows a declared property at the beginning of the configuration file. It is then used further down the file to specify the place to create the output file.

<configuration>

  <substitutionProperty name="user.home.dir" value="/Users/seb" />

  <appender name="FILE"
    class="ch.qos.logback.core.rolling.RollingFileAppender">
    <file>${user.home.dir}/myApp.log</file>
    <rollingPolicy
      class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
      <FileNamePattern>
        myApp-%d{yyyy-MM-dd-HH}.log
      </FileNamePattern>
    </rollingPolicy>
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>
        %date %level [%thread] %logger{10} [%file : %line] %msg%n
      </Pattern>
    </layout>
  </appender>

  <root>
    <level value="debug" />
    <appender-ref ref="FILE" />
  </root>
</configuration>

The next example shows the use of a System property to achieve the same result. The property is not declared anywhere, thus logback will look for it in the System properties.

<configuration>

  <appender name="FILE"
    class="ch.qos.logback.core.rolling.RollingFileAppender">
    <file>${user.home.dir}/myApp.log</file>
    <rollingPolicy
      class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
      <FileNamePattern>
        myApp-%d{yyyy-MM-dd-HH}.log
      </FileNamePattern>
    </rollingPolicy>
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>
        %date %level [%thread] %logger{10} [%file : %line] %msg%n
      </Pattern>
    </layout>
  </appender>

  <root>
    <level value="debug" />
    <appender-ref ref="FILE" />
  </root>
</configuration>

When many variables are used, it is often more convenient to create a separate file that will contain all the variables. Here is how one can do such a setup.

<configuration>

  <substitutionProperty file="variables1.properties" />

  <appender name="FILE"
    class="ch.qos.logback.core.rolling.RollingFileAppender">
    <file>${user.home.dir}/myApp.log</file>
    <rollingPolicy
      class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
      <FileNamePattern>
        myApp-%d{yyyy-MM-dd-HH}.log
      </FileNamePattern>
    </rollingPolicy>
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>
        %date %level [%thread] %logger{10} [%file : %line] %msg%n
      </Pattern>
    </layout>
  </appender>

  <root>
    <level value="debug" />
    <appender-ref ref="FILE" />
  </root>
</configuration>

This configuration file contains a link to another file called variables1.properties. The variables contained in that other file will be read and will thus be available to the logback configuration file. Here is what the variable.properties file looks like.

user.home.dir=/Users/seb

Nothing more is needed to declare the variable.

Recursive subsitution is also available. If the user wants to use variables to specify not only the destination directory but also the file name, here is what she would write in her variables1.properties file.

user.home.dir=/Users/seb
fileName=myApp.log
destination=${user.home.dir}/${fileName}

In the configuration file, only the last variable, ${destination} will be used, as shown below:

<configuration>

  <substitutionProperty file="variables1.properties" />

  <appender name="FILE"
    class="ch.qos.logback.core.rolling.RollingFileAppender">
    <file>${destination}</file>
    <rollingPolicy
      class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
      <FileNamePattern>
        myApp-%d{yyyy-MM-dd-HH}.log
      </FileNamePattern>
    </rollingPolicy>
    <layout class="ch.qos.logback.classic.PatternLayout">
      <Pattern>
        %date %level [%thread] %logger{10} [%file : %line] %msg%n
      </Pattern>
    </layout>
  </appender>

  <root>
    <level value="debug" />
    <appender-ref ref="FILE" />
  </root>
</configuration>

File inclusion

It is possible to include configuration elements into a logback.xml configuration file. This is done by using a include element, as shown in the example below:

<configuration>
  <include file="path/to/configuration/file"/>
</configuration>

The target file must have its elements nested inside an included element. For example, a ConsoleAppender would be declared this way:

<included>

  <appender name="redirectConsole"
    class="ch.qos.logback.core.ConsoleAppender">
    <layout class="ch.qos.logback.classic.PatternLayout">
      <param name="Pattern" value="%d - %m%n" />
    </layout>
  </appender>

  <root>
    <level value="DEBUG" />
    <appender-ref ref="redirectConsole" />
  </root>

</included>

The file to be included can be referenced as a URL or as a resource. To reference a URL use the url attribute instead of the file attribute in the previous example. To reference a resource, use the resource attribute instead of the file attribute in the previous example.

Using Joran in your own application

As we've seen, logback relies on Joran, a mature, flexible and powerful configuration framework. Many of the capabilities offered by logback modules are possible with the help of Joran.

Joran is actually a generic configuration system which can be used independently of logging. To emphaises this point, we should mention that the logback-core module does not have a notion of loggers. In that spirit, many of the examples related to this tutorial, have nothing to do with loggers, appenders or layouts.

The examples for this chapter can be found under LOGBACK_HOME/logback-examples/src/main/java/chapter3.

To install joran, simply download logback and add logback-core-VERSION.jar to your classpath.

Historical perspective

One of the most powerful features of the Java language is reflection. Reflection makes it possible to configure software systems declaratively. For example, many important properties of an EJB are configured with the ejb.xml file. While EJBs are written in Java, many of their properties are specified within the ejb.xml file. Similarly, logback settings can be specified in a configuration file, expressed in XML format.

In log4j, logback's predecessor, DOMConfigurator that shipped with log4j version 1.2.x can parse configuration files written in XML. The DOMConfigurator was written in a way that forced to tweak it each time the structure of the configuration file changed. The modified code had to be recompiled and redeployed. Just as importantly, the code of the DOMConfigurator consists of loops dealing with children elements containing many interspersed if/else statements. One could not help but notice that that particular code reeked of redundancy. The digester project has shown that it is possible to parse XML files using pattern matching rules. At parse time, digester will apply the rules that match previously stated patterns. Rule classes are usually quite small and specialized. Consequently, they are relatively easy to understand and to maintain.

Joran is heavily inspired by the commons-digester project but uses a slightly different terminology. In commons-digester, a rule can be seen as consisting of a pattern and a rule, as shown by the Digester.addRule(String pattern, Rule rule) method. We find it unnecessarily confusing to have a rule to consist of itself, not recursively but with a different meaning. In Joran, a rule consists of a pattern and an action. An action is invoked when a match occurs for the corresponding pattern. This relation between patterns and actions lies at the core of Joran. Quite remarkably, one can deal with quite complex requirements by using simple patterns, or more precisely with exact matches and wildcard matches. For example, the pattern a/b will match a <b> element nested within an <a> element but not a <c> element, even if nested within a <b> element. It is also possible to match a particular XML element, regardless of its nesting level, by using the * wildcard character. For example, the pattern */a will match an <a> element at any nesting position within the document. Other types of patterns, for example a/*, are not currently supported by Joran.

SAX or DOM?

Due to the event-based architecture of the SAX API, a tool based on SAX cannot easily deal with forward references, that is, references to elements which are defined later than the current element being processed. Elements with cyclical references are equally problematic. More generally, the DOM API allows the user to perform searches on all the elements and make forward jumps.

This extra flexibility initially led us to choose the DOM API as the underlying parsing API for Joran. After some experimentation, it quickly became clear that dealing with jumps to distant elements while parsing the DOM tree did not make sense when the interpretation rules were expressed in the form of patterns and actions. Joran only needs to be given the elements in the XML document in a sequential, depth-first order.

Joran was first implemented in DOM. However, the author migrated to SAX in order to benefit form the location information provided to the user, that is, to an org.w3.sax.ContentHandler. With the help of location information, it becomes possible to display essential error reports to the user which include exact line and column. This extra information turns out to be handy in hunting down problems.

Actions

Actions extend the ch.qos.logback.core.joran.action.Action class which consists of the following abstract methods.

package ch.qos.logback.core.joran.action;

import org.xml.sax.Attributes;
import ch.qos.logback.core.joran.spi.ExecutionContext;

public abstract class Action {
  

 /**
  * Called when the parser first encounters an element.
  */
  public abstract void begin(ExecutionContext ec, 
                             String name, 
                             Attributes attributes);

 /**
  * Called when the parser encounters the element end. At 
  * this stage, we can assume that child elements, if any,
  * have been processed. 
  */
  public abstract void end(ExecutionContext ec, String name);
}

Thus, every action must implement the begin and end methods.

Execution context

To allow various actions to collaborate, the invocation of begin and end methods include an execution context as the first parameter. The execution context includes an object stack, an object map, an error list and a reference to the Joran interpreter invoking the action. Please see the ch.qos.logback.core.joran.spi.ExecutionContext class for the exact list of fields contained in the execution context.

Actions can collaborate together by fetching, pushing or popping objects from the common object stack, or by putting and fetching keyed objects on the common object map. Actions can report any error conditions by adding error items on the execution context's StatusManager.

A hello world example

The logback-examples/src/main/java/chapter3/helloWorld/ directory includes a trivial action and Joran interpreter setup which just displays Hello World when a <hello-world> element is encountered in an XML file. It also includes the basic steps which are necessary to set up and invoke a Joran interpreter.

The hello.xml file contains only one element, without any other nested elements. The HelloWorldAction class is a trivial implementation: it only prints "Hello World" in the console when it's begin() method is called.

HelloWorld is a class that sets up the Joran interpreter, with the minimal steps necessary:

It creates a RuleStore and a Context

It adds the hello-world pattern, with it's corresponding action

It creates a Joran interpreter, and passes the RuleStore

It creates a SAX parser and parses the given file, specifying the newly created Joran interpreter as the ContentHandler

It's last step is to print the content of the Context. Since Joran uses logback's powerfull Status objects for error reporting, one can have a good feedback on what happened during the parsing.

In this example, the parsing is rather simple. The hello-world element will activate HelloWorldAction's begin() and end() methods. In the first method, a simple call to System.out.println() will be issued, displaying Hello World in the console.

Collaborating actions

The logback-examples/src/main/java/joran/calculator/ directory includes several actions which collaborate together through the common object stack in order to accomplish simple computations.

The calculator1.xml file contains a computation element, with a nested literal element.

In the Calculator1 class, we declare various patterns and actions, that will collaborate and calculate a result based on the xml file. The simple calculator1.xml file only creates a computation and declares a literal value. The resulting parsing is pretty simple:

The ComputationAction1 class' begin() method is called

The LiteralAction class' begin() and end() methods are called

The ComputationAction1 class' end() method is called

What is interesting here is the way that the Actions collaborate. The LiteralAction reads a literal value and pushes it in the object stack maintained by the ExecutionContext. Once done, any other action can pop the value to read or modify it. Here, the end() method of the ComputationAction1 class pops the value from the stack and prints it.

The calculator2.xml file is a bit more complex, but much more interesting.

It contains the following elements:

<computation name="toto">
  <literal value="7"/>
  <literal value="3"/>
  <add/>
  <literal value="3"/>
  <multiply/>
</computation>

Here, there are obviously more actions that will be part of the computation.

When called, the AddAction class will remove the two integers at the bottom of the stack, add them and push the resulting integer at the top of the stack, for further use.

Later in the computation, the MultiplyAction class will be called. It will take the last two integers from the stack, multiply them and push the result in the stack.

We have here two examples of action whose begin() method behaves in a certain, predictable way, but whose end() methods are empty.

Finally, a calculator3.xml is also provided, to demonstrate the possibility elements that contain instances of the same element. Here's the content of calculator3.xml:

<computation name="toto">
  <computation>
    <literal value="7"/>
    <literal value="3"/>
    <add/>
  </computation>   
 
  <literal value="3"/>
  <multiply/>
</computation>

Much like the use of parentheses in an algebrical equation, the presence of a computation element nested in another is managed by the ComputationAction2 class using an internal stack. The well-formedness of XML will guarantee that a value saved by one begin() will be consumed only by the matching end() method.

New-rule action

Joran includes an action which allows the Joran interpreter to lean new rules on the fly while interpreting the XML file containing the new rules. See the logback-examples/src/main/java/joran/newRule/ directory for sample code.

In this package, the NewRuleCalculator class contains the same setup as we have seen so far, but for one line:

By adding this line, we ask Joran to allow new rules to be learnt at parsing time. It works pretty much like the other rules: it has a begin() and end() method, and is called each time the parser finds a new-rule element.

When called, the begin() method looks for a pattern and a actionClass attribute. The action class is then instanciated and added to the RuleStore, along with its corresponding pattern.

Here is how new rules can be declared in an xml file:

<new-rule pattern="*/computation/literal" actionClass="chapter3.calculator.LiteralAction"/>

Using new rule declarations, the preceding example, involving the calculation, could be expressed this way:

<computation name="toto">
  <new-rule pattern="*/computation/literal" 
            actionClass="chapter3.calculator.LiteralAction"/>
  <new-rule pattern="*/computation/add" 
            actionClass="chapter3.calculator.AddAction"/>
  <new-rule pattern="*/computation/multiply" 
            actionClass="chapter3.calculator.MultiplyAction"/>

  <computation>
    <literal value="7"/>
    <literal value="3"/>
    <add/>
  </computation>   
 
  <literal value="3"/>
  <multiply/>
</computation>

Implicit actions

The rules defined thus far are called explicit rules because they require an explicit pattern, hence fixing the tag name of the elements for which they apply.

In highly extensible systems, the number and type of components to handle are innumerable so that it would become very tedious or even impossible to list all the applicable patterns by name.

At the same time, even in highly extensible systems one can observe well-defined patterns linking the various parts together. Implicit rules come in very handy when processing components composed of sub-components unknown ahead of time. For example, Apache Ant is capable of handling tasks which contain tags unknown at compile time by looking at methods whose names start with add, as in addFile, or addClassPath. When Ant encounters an embedded tag within a task, it simply instantiates an object that matches the signature of the task class' add method and attaches the resulting object to the parent.

Joran includes similar capability in the form of implicit actions. Joran keeps a list of implicit actions which can be applied if no explicit pattern matches the current XML element. However, applying an implicit action may not be always appropriate. Before executing the implicit action, Joran asks an implicit action whether it is appropriate in the current context. Only if the action replies affirmatively does Joran interpreter invoke the (implicit) action. This extra step makes it possible to support multiple implicit actions or obviously none, if no implicit action is appropriate for a given situation.

For example, the NestedComponentIA extending ImplicitAction , will instantiate the class specified in a nested component and attach it to the parent component by using setter method of the parent component and the nested element's name. Under certain circumstances, a nested action needs to be applied to an element say <a> and also to another element <b> nested within <a>. The current implementation of NestedComponentIA is capable of handling multiply nested elements requiring intervention by the same implicit action.

Both ImplicitAction and NestedComponentIA are located in the ch.qos.logback.core.joran.action package.

Refer to the logback-examples/src/main/java/joran/implicit directory for an example of an implicit action.

In that directory, you will find two actions classes, one xml file and one class containing the setup of Joran.

The NOPAction class does nothing. It is used to set the context of the foo element, using this line:

After that, the implicit action, namely PrintMeImplicitAction, is added to the RuleStore. This is done by simply adding a new instance of the action to the Joran interpreter

When called, the isApplicable() method of PrintMeImplicitAction checks the value of the printme attribute. If the value is true, the implicit action is applicable: its begin() method will be called.

The implicit1.xml file contains the following lines:

<foo>

  <xyz printme="true">
    <abc printme="true"/>
  </xyz>

  <xyz/>

  <foo printme="true"/>

</foo>

As one can see, the first element will be printed, since it has a printme attribute, which bears the value true.

The second element will not be printed, because no printme attibute is present.

The last element will not be printed, although the required attribute is present. This is because implicit rules are called only if no explicit rules are defined. Since we added a NOPAction with the */foo pattern, it will be used instead of the PrintMeImplicitAction.

Running the example yields the following output:

Element <xyz> asked to be printed.
Element <abc> asked to be printed.
ERROR in ch.qos.logback.core.joran.spi.ExecutionContext@1c5c1 - no applicable action \
for <xyz>, current pattern is [/foo/xyz]

The last line was printed because of a call to StatusPrinter at the end of the main class.

Non goals

The Joran API is not intended to be used to parse documents with thousands of elements.