假设文件的顺序并不重要,并且当您说没有自定义任务时,您并没有排除 ant脚本任务,那么由此派生的东西可能是合适的。
<fileset id="Fileset" dir="${basedir}" includes="Test*.java" />
<target name="scr">
<script language="javascript">
<![CDATA[
// Obtain a reference to fileset in the enclosing project
var fileSet = project.getReference( "Fileset" );
// Now get matching files.
var ds = fileSet.getDirectoryScanner( project );
var includes = ds.getIncludedFiles( );
var batchSize = 2;
var batch = 1;
for ( var i = 0; i < includes.length; i += batchSize )
{
// Create a new fileset to hold the sub-Fileset.
var filesetN = project.createDataType( "fileset" );
filesetN.setDir( fileSet.getDir( ) );
// Give the new Fileset an id and associate with the project.
project.addReference( "Fileset" + batch, filesetN );
// Populate the sub-Fileset.
for ( var j = 0; j < batchSize && ( i + j ) < includes.length; j++ )
{
filesetN.setIncludes( includes[i + j] );
}
batch++;
}
]]>
</script>
<!-- Just to illustrate. -->
<echo message="Set 1: ${toString:Fileset1}" />
<echo message="Set 2: ${toString:Fileset2}" />
<echo message="Set 3: ${toString:Fileset3}" />
</target>
鉴于您引用的六个文件,上面给出了:
scr:
[echo] Set 1: TestFive.java;TestFour.java
[echo] Set 2: TestOne.java;TestSix.java
[echo] Set 3: TestThree.java;TestTwo.java
我还没有想出一个纯粹的蚂蚁方式来实现这一点。您可能会考虑ant-contrib,它是一组相当广泛使用的扩展,可满足循环和可变属性的需求。
最后的评论,对于您的示例集,您可以获得所需的效果,但它不会超出三个文件集。使用first
和last
资源集合。
<target name="reso">
<first id="Fileset1" count="2">
<fileset refid="Fileset" />
</first>
<last id="Fileset3" count="2">
<fileset refid="Fileset" />
</last>
<difference id="Fileset2">
<resources refid="Fileset" />
<resources refid="Fileset1" />
<resources refid="Fileset3" />
</difference>
</target>