Using Ant, replace all _ in a file but only within a specific string?

Question

This seems like it should be simple: Using an Ant task, can I use replaceregexp to replace all of certain repeated characters, only in certain strings in files?

File contents:

Blah blah blah <ac:link> words_with_underscores_to_turn_to_spaces</link>
Blah blah blah Blah blah blah Blah blah blah Blah blah blah
Words_with_underscores_that_I_want_to_keep. Blah blah blah Blah blah blah.

Result wanted is:

Blah blah blah <ac:link> words with underscores to turn to spaces</link> 
Blah blah blah Blah blah blah Blah blah blah Blah blah blah 
Words_with_underscores_that_I_want_to_keep. Blah blah blah Blah blah blah.

I can use replaceregexp to match <ac:link.*?/link> and limit the replacements to only within those strings, but in that case how do I tell it to replace all underscores that it finds inside that string, no matter where they fall? The lines with underscores aren't always the same number of words.

I also tried a copy task approach, like this:

  <copy todir=".\test_output">
   <filterchain>
   <tokenfilter>
     <containsregex pattern="(ac:link.*?link)" flags="gi"/>
    <replacestring from="_" to=" "/>
   </tokenfilter>
  </filterchain>
  <fileset dir=".\underscore_test_output" includes="**/*.txt"/>
 </copy>

That replaces the underscores with spaces in the links and moves the links into a new file, but it excludes the rest of the source file, since I only matched the links. Any ideas?

Chad Nouis · Accepted Answer · 2014-07-31 17:27:30Z

Using a <scriptfilter> is an excellent way to have conditional logic in a <filterchain>.

In the script below, a <filetokenizer/> treats the entire input file as a single token. This allows the JavaScript to match tags across newlines.

Ant script

<copy todir="${out.dir}">
  <fileset dir="${basedir}" includes="test.txt"/>
  <filterchain>
    <tokenfilter>
      <filetokenizer/>
      <scriptfilter language="javascript"><![CDATA[
        var originalFile = self.getToken();
        var originalFileIndex = 0;
        var transformedFile = '';
        var keepGoing = true;

        // The "ac:" vs no "ac:" discrepency between the opening and closing 
        // tags is in the sample text from the question.
        var openingTagFormat = '<ac:link>';
        var closingTagFormat = '</link>';

        while (keepGoing) {
          var openingAcLinkBeginIndex = originalFile.indexOf(openingTagFormat, originalFileIndex);
          keepGoing = openingAcLinkBeginIndex > -1;
          if (keepGoing) {
            var openingAcLinkEndIndex = openingAcLinkBeginIndex + openingTagFormat.length;
            var closingAcLinkBeginIndex = originalFile.indexOf(closingTagFormat, openingAcLinkEndIndex);
            keepGoing = closingAcLinkBeginIndex > -1;
            if (keepGoing) {
              transformedFile += originalFile.slice(originalFileIndex, openingAcLinkEndIndex);
              var closingAcLinkEndIndex = closingAcLinkBeginIndex + closingTagFormat.length;
              var stringBetweenAcLinkTags = originalFile.slice(openingAcLinkEndIndex, closingAcLinkBeginIndex);
              transformedFile += stringBetweenAcLinkTags.replace(/_/g, ' ');
              transformedFile += originalFile.slice(closingAcLinkBeginIndex, closingAcLinkEndIndex);
              originalFileIndex = closingAcLinkEndIndex;
            }
          }
        }

        transformedFile += originalFile.substring(originalFileIndex);

        self.setToken(transformedFile);
      ]]></scriptfilter>
    </tokenfilter>
  </filterchain>
</copy>

Output

Blah blah blah <ac:link> words with underscores to turn to spaces</link>
Blah blah blah Blah blah blah Blah blah blah Blah blah blah
Words_with_underscores_that_I_want_to_keep. Blah blah blah Blah blah blah.

Collectives™ on Stack Overflow

Using Ant, replace all _ in a file but only within a specific string?

1 Answer 1

Ant script

Output

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Ant script

Output

Comments

Your Answer

Sign up or log in

Post as a guest

Related