Selectors

Selectors are a mechanism whereby the files that make up a <fileset> can be selected based on criteria other than filename as provided by the <include> and <exclude> tags.

How to use a Selector

A selector is an element of FileSet, and appears within it. It can also be defined outside of any target by using the <selector> tag and then using it as a reference.

Different selectors have different attributes. Some selectors can contain other selectors, and these are called Selector Containers. There is also a category of selectors that allow user-defined extensions, called Custom Selectors. The ones built in to Apache Ant are called Core Selectors.

Core Selectors

Core selectors are the ones that come standard with Ant. They can be used within a fileset and can be contained within Selector Containers.

The core selectors are:

Contains Selector

The <contains> tag in a FileSet limits the files defined by that fileset to only those which contain the string specified by the text attribute. .

The <contains> selector can be used as a ResourceSelector (see the <restrict> ResourceCollection).

Attribute Description Required
text Specifies the text that every file must contain Yes
casesensitive Whether to pay attention to case when looking for the string in the text attribute. Default is true. No
ignorewhitespace Whether to eliminate whitespace before checking for the string in the text attribute. Default is false. No
encoding Encoding of the resources being selected. Required in practice if the encoding of the files being selected is different from the default encoding of the JVM where Ant is running. Since Ant 1.9.0 No

Here is an example of how to use the Contains Selector:

<fileset dir="${doc.path}" includes="**/*.html">
    <contains text="script" casesensitive="no"/>
</fileset>

Selects all the HTML files that contain the string script.

Date Selector

The <date> tag in a FileSet will put a limit on the files specified by the include tag, so that tags whose last modified date does not meet the date limits specified by the selector will not end up being selected.

Attribute Description Required
datetime Specifies the date and time to test for. Should be in the format MM/DD/YYYY HH:MM AM_or_PM, or an alternative pattern specified via the pattern attribute. At least one of the two.
millis The number of milliseconds since 1970 that should be tested for. It is usually much easier to use the datetime attribute.
when Indicates how to interpret the date, whether the files to be selected are those whose last modified times should be before, after, or equal to the specified value. Acceptable values for this attribute are:
  • before - select files whose last modified date is before the indicated date
  • after - select files whose last modified date is after the indicated date
  • equal - select files whose last modified date is this exact date
The default is equal.
No
granularity The number of milliseconds leeway to use when comparing file modification times. This is needed because not every file system supports tracking the last modified time to the millisecond level. Default is 0 milliseconds, or 2 seconds on DOS systems. No
pattern The SimpleDateFormat-compatible pattern to use when interpreting the datetime attribute. Since Ant 1.6.2 No
checkdirs Indicates whether or not to check dates on directories. No, defaults to false

Here is an example of how to use the Date Selector:

<fileset dir="${jar.path}" includes="**/*.jar">
    <date datetime="01/01/2001 12:00 AM" when="before"/>
</fileset>

Selects all JAR files which were last modified before midnight January 1, 2001.

Depend Selector

The <depend> tag selects files whose last modified date is later than another, equivalent file in another location.

The <depend> tag supports the use of a contained <mapper> element to define the location of the file to be compared against. If no <mapper> element is specified, the identity type mapper is used.

The <depend> selector is case-sensitive.

Attribute Description Required
targetdir The base directory to look for the files to compare against. The precise location depends on a combination of this attribute and the <mapper> element, if any. Yes
granularity The number of milliseconds leeway to give before deciding a file is out of date. This is needed because not every file system supports tracking the last modified time to the millisecond level. Default is 0 milliseconds, or 2 seconds on DOS systems. No

Here is an example of how to use the Depend Selector:

<fileset dir="${ant.1.5}/src/main" includes="**/*.java">
    <depend targetdir="${ant.1.4.1}/src/main"/>
</fileset>

Selects all the Java source files which were modified in the 1.5 release.

Depth Selector

The <depth> tag selects files based on how many directory levels deep they are in relation to the base directory of the fileset.

Attribute Description Required
min The minimum number of directory levels below the base directory that a file must be in order to be selected. Default is no limit. At least one of the two.
max The maximum number of directory levels below the base directory that a file can be and still be selected. Default is no limit.

Here is an example of how to use the Depth Selector:

<fileset dir="${doc.path}" includes="**/*">
    <depth max="1"/>
</fileset>

Selects all files in the base directory and one directory below that.

Different Selector

The <different> selector will select a file if it is deemed to be 'different' from an equivalent file in another location. The rules for determining difference between the two files are as follows:

  1. If a file is only present in the resource collection you apply the selector to but not in targetdir (or after applying the mapper) the file is selected.
  2. If a file is only present in targetdir (or after applying the mapper) it is ignored.
  3. Files with different lengths are different.
  4. If ignoreFileTimes is turned off, then differing file timestamps will cause files to be regarded as different.
  5. Unless ignoreContents is set to true, a byte-for-byte check is run against the two files.
This is a useful selector to work with programs and tasks that don't handle dependency checking properly; even if a predecessor task always creates its output files, followup tasks can be driven off copies made with a different selector, so their dependencies are driven on the absolute state of the files, not just a timestamp. For example: anything fetched from a web site, or the output of some program. To reduce the amount of checking, when using this task inside a <copy> task, set preservelastmodified to true to propagate the timestamp from the source file to the destination file.

The <different> selector supports the use of a contained <mapper> element to define the location of the file to be compared against. If no <mapper> element is specified, the identity type mapper is used.

Attribute Description Required
targetdir The base directory to look for the files to compare against. The precise location depends on a combination of this attribute and the <mapper> element, if any. Yes
ignoreFileTimes Whether to use file times in the comparison or not. Default is true (time differences are ignored). No
ignoreContents Whether to do a byte per byte compare. Default is false (contents are compared). Since Ant 1.6.3 No
granularity The number of milliseconds leeway to give before deciding a file is out of date. This is needed because not every file system supports tracking the last modified time to the millisecond level. Default is 0 milliseconds, or 2 seconds on DOS systems. No

Here is an example of how to use the Different Selector:

<fileset dir="${ant.1.5}/src/main" includes="**/*.java">
    <different targetdir="${ant.1.4.1}/src/main"
        ignoreFileTimes="true"/>
</fileset>

Compares all the Java source files between the 1.4.1 and the 1.5 release and selects those who are different, disregarding file times.

Filename Selector

The <filename> tag acts like the <include> and <exclude> tags within a fileset. By using a selector instead, however, one can combine it with all the other selectors using whatever selector container is desired.

The <filename> selector is case-sensitive.

Attribute Description Required
name The name of files to select. The name parameter can contain the standard Ant wildcard characters. Exactly one of the two
regex The regular expression matching files to select.
casesensitive Whether to pay attention to case when looking at file names. Default is "true". No
negate Whether to reverse the effects of this filename selection, therefore emulating an exclude rather than include tag. Default is "false". No

Here is an example of how to use the Filename Selector:

<fileset dir="${doc.path}" includes="**/*">
    <filename name="**/*.css"/>
</fileset>

Selects all the cascading style sheet files.

Present Selector

The <present> tag selects files that have an equivalent file in another directory tree.

The <present> tag supports the use of a contained <mapper> element to define the location of the file to be tested against. If no <mapper> element is specified, the identity type mapper is used.

The <present> selector is case-sensitive.

Attribute Description Required
targetdir The base directory to look for the files to compare against. The precise location depends on a combination of this attribute and the <mapper> element, if any. Yes
present Whether we are requiring that a file is present in the src directory tree only, or in both the src and the target directory tree. Valid values are:
  • srconly - select files only if they are in the src directory tree but not in the target directory tree
  • both - select files only if they are present both in the src and target directory trees
Default is both. Setting this attribute to "srconly" is equivalent to wrapping the selector in the <not> selector container.
No

Here is an example of how to use the Present Selector:

<fileset dir="${ant.1.5}/src/main" includes="**/*.java">
    <present present="srconly" targetdir="${ant.1.4.1}/src/main"/>
</fileset>

Selects all the Java source files which are new in the 1.5 release.

Regular Expression Selector

The <containsregexp> tag in a FileSet limits the files defined by that fileset to only those which contents contain a match to the regular expression specified by the expression attribute.

The <containsregexp> selector can be used as a ResourceSelector (see the <restrict> ResourceCollection).

Attribute Description Required
expression Specifies the regular expression that must match true in every file Yes
casesensitive Perform a case sensitive match. Default is true. since Ant 1.8.2 No
multiline Perform a multi line match. Default is false. since Ant 1.8.2 No
singleline This allows '.' to match new lines. SingleLine is not to be confused with multiline, SingleLine is a perl regex term, it corresponds to dotall in java regex. Default is false. since Ant 1.8.2 No

Here is an example of how to use the regular expression Selector:

<fileset dir="${doc.path}" includes="*.txt">
    <containsregexp expression="[4-6]\.[0-9]"/>
</fileset>

Selects all the text files that match the regular expression (have a 4,5 or 6 followed by a period and a number from 0 to 9).

Size Selector

The <size> tag in a FileSet will put a limit on the files specified by the include tag, so that tags which do not meet the size limits specified by the selector will not end up being selected.

Attribute Description Required
value The size of the file which should be tested for. Yes
units The units that the value attribute is expressed in. When using the standard single letter SI designations, such as "k","M", or "G", multiples of 1000 are used. If you want to use power of 2 units, use the IEC standard: "Ki" for 1024, "Mi" for 1048576, and so on. The default is no units, which means the value attribute expresses the exact number of bytes. No
when Indicates how to interpret the size, whether the files to be selected should be larger, smaller, or equal to that value. Acceptable values for this attribute are:
  • less - select files less than the indicated size
  • more - select files greater than the indicated size
  • equal - select files this exact size
The default is equal.
No

Here is an example of how to use the Size Selector:

<fileset dir="${jar.path}">
  <patternset>
    <include name="**/*.jar"/>
  </patternset>
  <size value="4" units="Ki" when="more"/>
</fileset>

Selects all JAR files that are larger than 4096 bytes.

Type Selector

The <type> tag selects files of a certain type: directory or regular.

Attribute Description Required
type The type of file which should be tested for. Acceptable values are:
  • file - regular files
  • dir - directories
Yes

Here is an example of how to use the Type Selector to select only directories in ${src}

<fileset dir="${src}">
  <type type="dir"/>
</fileset>

The Type Selector is often used in conjunction with other selectors. For example, to select files that also exist in a template directory, but avoid selecting empty directories, use:

<fileset dir="${src}">
    <and>
        <present targetdir="template"/>
        <type type="file"/>
    </and>
</fileset>

Modified Selector

The <modified> selector computes a value for a file, compares that to the value stored in a cache and select the file, if these two values differ.

Because this selector is highly configurable the order in which the selection is done is:

  1. get the absolute path for the file
  2. get the cached value from the configured cache (absolute path as key)
  3. get the new value from the configured algorithm
  4. compare these two values with the configured comparator
  5. update the cache if needed and requested
  6. do the selection according to the comparison result

The comparison, computing of the hashvalue and the store is done by implementation of special interfaces. Therefore they may provide additional parameters.

The <modified> selector can be used as a ResourceSelector (see the <restrict> ResourceCollection). In that case it maps simple file resources to files and does its job. If the resource is from another type, the <modified> selector tries to (attention!) copy the content into a local file for computing the hashvalue.

Attribute Description Required
algorithm The type of algorithm should be used. Acceptable values are (further information see later):
  • hashvalue - HashvalueAlgorithm
  • digest - DigestAlgorithm
  • checksum - ChecksumAlgorithm
No, defaults to digest
cache The type of cache should be used. Acceptable values are (further information see later):
  • propertyfile - PropertyfileCache
No, defaults to propertyfile
comparator The type of comparator should be used. Acceptable values are (further information see later):
  • equal - EqualComparator
  • rule - java.text.RuleBasedCollator (see note for restrictions)
No, defaults to equal
algorithmclass Classname of custom algorithm implementation. Lower priority than algorithm. No
cacheclass Classname of custom cache implementation. Lower priority than cache. No
comparatorclass Classname of custom comparator implementation. Lower priority than comparator. No
update Should the cache be updated when values differ? (boolean) No, defaults to true
seldirs Should directories be selected? (boolean) No, defaults to true
selres Should Resources without an InputStream, and therefore without checking, be selected? (boolean) No, defaults to true. Only relevant when used as ResourceSelector.
delayupdate If set to true, the storage of the cache will be delayed until the next finished BuildEvent; task finished, target finished or build finished, whichever comes first. This is provided for increased performance. If set to false, the storage of the cache will happen with each change. This attribute depends upon the update attribute. (boolean) No, defaults to true

These attributes can be set with nested <param/> tags. With <param/> tags you can set other values too - as long as they are named according to the following rules:

Algorithm options
Name Description
hashvalue Reads the content of a file into a java.lang.String and use thats hashValue(). No additional configuration required.
digest Uses java.security.MessageDigest. This Algorithm supports the following attributes:
  • algorithm.algorithm (optional): Name of the Digest algorithm (e.g. 'MD5' or 'SHA', default = MD5)
  • algorithm.provider (optional): Name of the Digest provider (default = null)
checksum Uses java.util.zip.Checksum. This Algorithm supports the following attributes:
  • algorithm.algorithm (optional): Name of the algorithm (e.g. 'CRC' or 'ADLER', default = CRC)
Cache options
propertyfile Use the java.util.Properties class and its possibility to load and store to file. This Cache implementation supports the following attributes:
  • cache.cachefile (optional): Name of the properties-file (default = cache.properties)
Comparator options
equal Very simple object comparison.
rule Uses java.text.RuleBasedCollator for Object comparison. (see note for restrictions)

The <modified> selector supports a nested <classpath> element that represents a PATH like structure for finding custom interface implementations.

Here are some examples of how to use the Modified Selector:

    <copy todir="dest">
        <fileset dir="src">
            <modified/>
        </fileset>
    </copy>
    

This will copy all files from src to dest which content has changed. Using an updating PropertyfileCache with cache.properties and MD5-DigestAlgorithm.

    <copy todir="dest">
        <fileset dir="src">
            <modified update="true"
                      seldirs="true"
                      cache="propertyfile"
                      algorithm="digest"
                      comparator="equal">
                <param name="cache.cachefile"     value="cache.properties"/>
                <param name="algorithm.algorithm" value="MD5"/>
            </modified>
        </fileset>
    </copy>
    

This is the same example rewritten as CoreSelector with setting the all the values (same as defaults are).

    <copy todir="dest">
        <fileset dir="src">
            <custom class="org.apache.tools.ant.types.selectors.modifiedselector.ModifiedSelector">
                <param name="update"     value="true"/>
                <param name="seldirs"    value="true"/>
                <param name="cache"      value="propertyfile"/>
                <param name="algorithm"  value="digest"/>
                <param name="comparator" value="equal"/>
                <param name="cache.cachefile"     value="cache.properties"/>
                <param name="algorithm.algorithm" value="MD5"/>
            </custom>
        </fileset>
    </copy>
    

And this is the same rewritten as CustomSelector.

  <target name="generate-and-upload-site">
      <echo> generate the site using forrest </echo>
      <antcall target="site"/>

      <echo> upload the changed file </echo>
      <ftp server="${ftp.server}" userid="${ftp.user}" password="${ftp.pwd}">
          <fileset dir="htdocs/manual">
              <modified/>
          </fileset>
      </ftp>
  </target>
    

A useful scenario for this selector inside a build environment for homepage generation (e.g. with Apache Forrest). Here all changed files are uploaded to the server. The CacheSelector saves therefore much upload time.

  <modified cacheclassname="com.mycompany.MyCache">
      <classpath>
          <pathelement location="lib/mycompany-antutil.jar"/>
      </classpath>
  </modified>
    

Uses com.mycompany.MyCache from a jar outside of Ants own classpath as cache implementation

Note on RuleBasedCollator

The RuleBasedCollator needs a format for its work, but its needed while instantiation. There is a problem in the initialization algorithm for this case. Therefore you should not use this (or tell me the workaround :-).

Signed Selector

The <signedselector> tag selects signed files and optionally signed with a certain name.

This selector has been added in Apache Ant 1.7.

Attribute Description Required
name The signature name to check for. no

Readable Selector

The <readable> selector selects only files that are readable. Ant only invokes java.io.File#canRead so if a file is unreadable but the Java VM cannot detect this state, this selector will still select the file.

Writable Selector

The <writable> selector selects only files that are writable. Ant only invokes java.io.File#canWrite so if a file is unwritable but the Java VM cannot detect this state, this selector will still sele