update documentation for <fixcrlf>

PR: 1053
Submitted by:	Peter B. West <pbwest@powerup.com.au>


git-svn-id: https://svn.apache.org/repos/asf/ant/core/trunk@269468 13f79535-47bb-0310-9956-ffa450edef68

docs/manual/CoreTasks/fixcrlf.html

@@ -9,21 +9,38 @@
 
 <h2><a name="fixcrlf">FixCRLF</a></h2>
 <h3>Description</h3>
-<p>Adjusts a text file to local.</p>
-<p>It is possible to refine the set of files that are being adjusted. This can be
-done with the <i>includes</i>, <i>includesfile</i>, <i>excludes</i>, <i>excludesfile</i> and <i>defaultexcludes</i>
-attributes. With the <i>includes</i> or <i>includesfile</i> attribute you specify the files you want to
-have included by using patterns. The <i>exclude</i> or <i>excludesfile</i> attribute is used to specify
-the files you want to have excluded. This is also done with patterns. And
-finally with the <i>defaultexcludes</i> attribute, you can specify whether you
-want to use default exclusions or not. See the section on <a
-href="../dirtasks.html#directorybasedtasks">directory based tasks</a>, on how the
-inclusion/exclusion of files works, and how to write patterns.</p>
+      <p>
+	Adjusts a text file to local conventions.
+      </p>
+      
+      <p>
+	The set of files to be adjusted can be refined with the
+	<i>includes</i>, <i>includesfile</i>, <i>excludes</i>,
+	<i>excludesfile</i> and <i>defaultexcludes</i>
+	attributes. Patterns provided through the <i>includes</i> or
+	<i>includesfile</i> attributes specify files to be
+	included. Patterns provided through the <i>exclude</i> or
+	<i>excludesfile</i> attribute specify files to be
+	excluded. Additionally, default exclusions can be specified with
+	the <i>defaultexcludes</i> attribute. See the section on <a
+	href="../dirtasks.html#directorybasedtasks">directory based
+	tasks</a>, for details of file inclusion/exclusion patterns
+	and their usage.
+      </p>
+      
 <p>This task forms an implicit <a href="../CoreTypes/fileset.html">FileSet</a> and
 supports all attributes of <code>&lt;fileset&gt;</code>
 (<code>dir</code> becomes <code>srcdir</code>) as well as the nested
 <code>&lt;include&gt;</code>, <code>&lt;exclude&gt;</code> and
 <code>&lt;patternset&gt;</code> elements.</p>
+
+      <p>
+	The output file is only written if it is a new file, or if it
+	differs from the existing file.  This prevents spurious
+	rebuilds based on unchanged files which have been regenerated
+	by this task.
+      </p>
+
 <h3>Parameters</h3>
 <table border="1" cellpadding="2" cellspacing="0">
   <tr>
@@ -72,24 +89,82 @@ supports all attributes of <code>&lt;fileset&gt;</code>
       (&quot;yes&quot;/&quot;no&quot;). Default excludes are used when omitted.</td>
     <td valign="top" align="center">No</td>
   </tr>
-  <tr>
-    <td valign="top">cr</td>
-    <td valign="top">Specifies how carriage return (CR) characters are to
-      be handled.  Valid values for this property are:
-      <ul>
-      <li>add: ensure that there is a CR before every LF</li>
-      <li>asis: leave CR characters alone</li>
-      <li>remove: remove all CR characters</li>
-      </ul>
-      Default is based on the platform on which you are running this task.
-      For Unix platforms, the default is remove.  For DOS based systems
-      (including Windows), the default is add.
-      <p>
-      Note: Unless this property is specified as &quot;asis&quot;, extra CR characters
-      which do not precede a LF will be removed.</p>
-      </td>
-    <td valign="top" align="center">No</td>
-  </tr>
+	<tr>
+	  <td valign="top">eol</td>
+	  <td valign="top">
+	    Specifies how end-of-line (EOL) characters are to be
+	    handled.  The EOL characters are CR, LF and the pair CRLF.
+	    Valid values for this property are:
+	    <ul>
+	      <li>asis: leave EOL characters alone</li>
+	      <li>cr: convert all EOLs to a single CR</li>
+	      <li>lf: convert all EOLs to a single LF</li>
+	      <li>crlf: convert all EOLs to the pair CRLF</li>
+	    </ul>
+	    Default is based on the platform on which you are running
+	    this task.  For Unix platforms, the default is "lf".
+	    For DOS based systems (including Windows), the default is
+	    "crlf".  For Mac OS, the default is "cr".
+	    <p>
+	      This is the preferred method for specifying EOL.  The
+	      &quot;<i><b>cr</b></i>&quot; attribute (see below) is
+	      now deprecated.  If both are specified, &quot;eol&quot;
+	      takes precedence.
+	    </p>
+	    <p>
+	      <i>N.B.</i>: One special case is recognized. The three
+	      characters CR-CR-LF are regarded as a single EOL.
+	      Unless this property is specified as &quot;asis&quot;,
+	      this sequence will be converted into the specified EOL
+	      type.
+	    </p>
+	  </td>
+	  <td valign="top" align="center">No</td>
+	</tr>
+	<tr>
+	  <td valign="top">cr</td>
+	  <td valign="top">
+	    <i><b>Deprecated.</b></i> Specifies how CR characters are
+	    to be handled at end-of-line (EOL).  Valid values for this
+	    property are:
+	    <ul>
+	      <li>asis: leave EOL characters alone.</li>
+	      <li>
+		add: add a CR before any single LF characters. The
+		intent is to convert all EOLs to the pair CRLF.
+	      </li>
+	      <li>
+		remove: remove all CRs from the file.  The intent is
+		to convert all EOLs to a single LF.
+	      </li>
+	    </ul>
+	    Default is based on the platform on which you are running
+	    this task.  For Unix platforms, the default is "remove".
+	    For DOS based systems (including Windows), the default is
+	    "add".
+	    <p>
+	      <i>N.B.</i>: One special case is recognized. The three
+	      characters CR-CR-LF are regarded as a single EOL.
+	      Unless this property is specified as &quot;asis&quot;,
+	      this sequence will be converted into the specified EOL
+	      type.
+	    </p>
+	  </td>
+	  <td valign="top" align="center">No</td>
+	</tr>
+	<tr>
+	  <td valign="top">javafiles</td>
+	  <td valign="top">
+	    Used only in association with the
+	    &quot;<i><b>tab</b></i>&quot; attribute (see below), this
+	    boolean attribute indicates whether the fileset is a set
+	    of java source files
+	    (&quot;yes&quot;/&quot;no&quot;). Defaults to
+	    &quot;no&quot;.  See notes in section on &quot;tab&quot;.
+	  </td>
+	  <td valign="top" align="center">No</td>
+	</tr>
+	<tr>
   <tr>
     <td valign="top">tab</td>
     <td valign="top">Specifies how tab characters are to be handled.  Valid
@@ -100,16 +175,27 @@ supports all attributes of <code>&lt;fileset&gt;</code>
       <li>remove: convert tabs to spaces</li>
       </ul>
       Default for this parameter is &quot;asis&quot;.
+	    <p>
+	      <i>N.B.</i>: When the attribute
+	      &quot;<i><b>javafiles</b></i>&quot; (see above) is
+	      &quot;true&quot;, literal TAB characters occurring
+	      within Java string or character constants are never
+	      modified.  This functionality also requires the
+	      recognition of Java-style comments.
+	    </p>
       <p>
-      Note: Unless this property is specified as &quot;asis&quot;, extra spaces and
-      tabs after the last non-whitespace character on the line will be removed.</p>
+		<i>N.B.</i>: There is an incompatibility between this
+		and the previous version in the handling of white
+		space at the end of lines.  This version does
+		<i><b>not</b></i> remove trailing whitespace on lines.
+</p>
       </td>
     <td valign="top" align="center">No</td>
   </tr>
   <tr>
     <td valign="top">tablength</td>
-    <td valign="top">The number of characters a TAB stop corresponds to. 
-      Must be a positive power of 2, default for this parameter is 8.</td>
+    <td valign="top">TAB character interval. Valid values are between
+	      2 and 80 inclusive.  The default for this parameter is 8.</td>
     <td valign="top" align="center">No</td>
   </tr>
   <tr>
@@ -130,16 +216,16 @@ supports all attributes of <code>&lt;fileset&gt;</code>
 </table>
 <h3>Examples</h3>
 <pre>  &lt;fixcrlf srcdir=&quot;${src}&quot;
-       cr=&quot;remove&quot; eof=&quot;remove&quot;
+       eol=&quot;lf&quot; eof=&quot;remove&quot;
        includes=&quot;**/*.sh&quot;
   /&gt;</pre>
-<p>Removes carriage return and eof characters from the shell scripts.  Tabs and
-spaces are left as is.</p>
+<p>Replaces EOLs with LF characters and removes eof characters from
+	the shell scripts.  Tabs and spaces are left as is.</p>
 <pre>  &lt;fixcrlf srcdir=&quot;${src}&quot;
-       cr=&quot;add&quot;
+       eol=&quot;crlf&quot;
        includes=&quot;**/*.bat&quot;
   /&gt;</pre>
-<p>Ensures that there are carriage return characters prior to evey line feed.
+<p>Replaces all EOLs with cr-lf pairs in the batch files.
 Tabs and spaces are left as is.
 EOF characters are left alone if run on
 DOS systems, and are removed if run on Unix systems.</p>
@@ -147,16 +233,34 @@ DOS systems, and are removed if run on Unix systems.</p>
        tabs=&quot;add&quot;
        includes=&quot;**/Makefile&quot;
   /&gt;</pre>
-<p>Adds or removes CR characters to match local OS conventions, and
-converts spaces to tabs when appropriate.  EOF characters are left alone if
+<p>Sets EOLs according to local OS conventions, and
+converts sequences of spaces and tabs to the minimal set of spaces and
+	tabs which will maintain spacing within the line.  Tabs are
+	set at 8 character intervals.  EOF characters are left alone if
 run on DOS systems, and are removed if run on Unix systems.
 Many versions of make require tabs prior to commands.</p>
+      <pre>  &lt;fixcrlf srcdir=&quot;${src}&quot;
+       tabs=&quot;remove&quot;
+       tablength=&quot;3&quot;
+       eol=&quot;lf&quot;
+       javafiles=&quot;yes&quot;
+       includes=&quot;**/*.java&quot;
+  /&gt;</pre>
+      <p>
+	Converts all EOLs in the included java source files to a
+	single LF.  Replace all TAB characters except those in string
+	or character constants with spaces, assuming a tab width of 3.
+	If run on a unix system, any CTRL-Z EOF characters at the end
+	of the file are removed.  On DOS/Windows, any such EOF
+	characters will be left untouched.
+      </p>
 <pre>  &lt;fixcrlf srcdir=&quot;${src}&quot;
        tabs=&quot;remove&quot;
        includes=&quot;**/README*&quot;
   /&gt;</pre>
-<p>Adds or removes CR characters to match local OS conventions, and
-converts all tabs to spaces.  EOF characters are left alone if run on
+<p>Sets EOLs according to local OS conventions, and
+converts all tabs to spaces, assuming a tab width of 8.
+EOF characters are left alone if run on
 DOS systems, and are removed if run on Unix systems.
 You never know what editor a user will use to browse README's.</p>
 <hr>