<html>

<head>
<meta http-equiv="Content-Language" content="en-us">
<title>Ant User Manual</title>
</head>

<body>

<h2><a name="directorybasedtasks">Directory-based Tasks</a></h2>
<p>Some tasks use directory trees for the task they perform.
For example, the <a
href="CoreTasks/javac.html">javac</a> task, which works upon a directory tree
with <code>.java</code> files.
Sometimes it can be very useful to work on a subset of that directory tree. This
section describes how you can select a subset of such a directory tree.</p>
<p>Ant gives you two ways to create a subset, both of which can be used at the same
time:</p>
<ul>
  <li>Only include files/directories that match at least one pattern of a set of
    patterns.</li>
  <li>Exclude files/directories that match at least one pattern a set of
    patterns.</li>
</ul>
<p>When both inclusion and exclusion are used, only files/directories that match
the include patterns, and don't match the exclude patterns, are used.</p>
<p>Patterns can be specified inside the buildfile via task attributes or
nested elements and via external files. Each line of the external file
is taken as a pattern that is added to the list of include or exclude 
patterns.</p>
<h3>Patterns</h3>
<p>As described earlier, patterns are used for the inclusion and exclusion.
These patterns look very much like the patterns used in DOS and UNIX:</p>
<p>'*' matches zero or more characters, '?' matches one character.</p>
<p><b>Examples:</b></p>
<p>
<code>*.java</code>&nbsp;&nbsp;matches&nbsp;&nbsp;<code>.java</code>,
<code>x.java</code> and <code>FooBar.java</code>, but
not <code>FooBar.xml</code> (does not end with <code>.java</code>).</p>
<p>
<code>?.java</code>&nbsp;&nbsp;matches&nbsp;&nbsp;<code>x.java</code>,
<code>A.java</code>, but not <code>.java</code> or <code>xyz.java</code>
(both don't have one character before <code>.java</code>).</p>
<p>
Combinations of <code>*</code>'s and <code>?</code>'s are allowed.</p>
<p>Matching is done per-directory. This means that first the first directory in
the pattern is matched against the first directory in the path to match. Then
the second directory is matched, and so on. For example, when we have the pattern <code>/?abc/*/*.java</code>
and the path <code>/xabc/foobar/test.java</code>,
the first <code>?abc</code> is matched with <code>xabc</code>,
then <code>*</code> is matched with <code>foobar</code>,
and finally <code>*.java</code> is matched with <code>test.java</code>.
They all match, so the path matches the pattern.</p>
<p>To make things a bit more flexible, we add one extra feature, which makes it
possible to match multiple directory levels. This can be used to match a
complete directory tree, or a file anywhere in the directory tree.
To do this, <code>**</code>
must be used as the name of a directory.
When <code>**</code> is used as the name of a
directory in the pattern, it matches zero or more directories.
For example:
<code>/test/**</code> matches all files/directories under <code>/test/</code>,
such as <code>/test/x.java</code>,
or <code>/test/foo/bar/xyz.html</code>, but not <code>/xyz.xml</code>.</p>
<p>There is one &quot;shorthand&quot; &#150; if a pattern ends
with <code>/</code>
or <code>\</code>, then <code>**</code>
is appended.
For example, <code>mypackage/test/</code> is interpreted as if it were
<code>mypackage/test/**</code>.</p>
<p><b>Example patterns:</b></p>
<table border="1" cellpadding="2" cellspacing="0">
  <tr>
    <td valign="top"><code>**/CVS/*</code></td>
    <td valign="top">Matches all files in <code>CVS</code> directories that can be located
      anywhere in the directory tree.<br>
      Matches:
      <pre>
      CVS/Repository
      org/apache/CVS/Entries
      org/apache/jakarta/tools/ant/CVS/Entries
      </pre>
      But not:
      <pre>
      org/apache/CVS/foo/bar/Entries (<code>foo/bar/</code> part does not match)</td>
      </pre>
  </tr>
  <tr>
    <td valign="top"><code>org/apache/jakarta/**</code></td>
    <td valign="top">Matches all files in the <code>org/apache/jakarta</code> directory tree.<br>
      Matches:
      <pre>
      org/apache/jakarta/tools/ant/docs/index.html
      org/apache/jakarta/test.xml
      </pre>
      But not:
      <pre>
      org/apache/xyz.java
      </pre>
      (<code>jakarta/</code> part is missing).</td>
  </tr>
  <tr>
    <td valign="top"><code>org/apache/**/CVS/*</code></td>
    <td valign="top">Matches all files in <code>CVS</code> directories
      that are located anywhere in the directory tree under
      <code>org/apache</code>.<br>
      Matches:
      <pre>
      org/apache/CVS/Entries
      org/apache/jakarta/tools/ant/CVS/Entries
      </pre>
      But not:
      <pre>
      org/apache/CVS/foo/bar/Entries
      </pre>
      (<code>foo/bar/</code> part does not match)</td>
  </tr>
  <tr>
    <td valign="top"><code>**/test/**</code></td>
    <td valign="top">Matches all files that have a <code>test</code>
        element in their path, including <code>test</code> as a filename.</td>
  </tr>
</table>
<p>When these patterns are used in inclusion and exclusion, you have a powerful
way to select just the files you want.</p>

<h3>Examples</h3>
<pre>  
&lt;copy todir=&quot;${dist}&quot;>
  &lt;fileset dir=&quot;${src}&quot; 
           includes=&quot;**/images/*&quot; 
           excludes=&quot;**/*.gif&quot; 
  /&gt;
&lt;/copy&gt;</pre>
<p>This copies all files in directories called <code>images</code> that are
located in the directory tree defined by <code>${src}</code> to the
destination directory defined by <code>${dist}</code>,
but excludes all <code>*.gif</code> files from the copy.</p>
<p> This example can also be expressed using nested elements:</p>
<pre>
&lt;copy todir=&quot;${dist}&quot;>
  &lt;fileset dir=&quot;${src}&quot;/&gt;
    &lt;include name=&quot;**/images/*&quot;/&gt;
    &lt;exclude name=&quot;**/*.gif&quot;/&gt;
  &lt;/fileset&gt;
&lt;/copy&gt;
</pre>

<h3>Default Excludes</h3>
<p>There are a set of definitions that are excluded by default from all directory-based tasks.
They are:</p>
<pre>
     **/*~
     **/#*#
     **/%*%
     **/CVS
     **/CVS/**
     **/.cvsignore
</pre>
<p>If you do not want these default excludes applied, you may disable them with the 
<code>defaultexcludes=&quot;no&quot;</code> attribute.</p>
<hr>
<p align="center">Copyright &copy; 2000,2001 Apache Software Foundation. All rights
Reserved.</p>

</body>
</html>