CODEWITHSUNDEEP
regexant
nolleh
nolleh

Reputation: 107

ant's replaceregexp , to remove a specific space

I can't find the way to remove a space using ant. (I don't mean trim, but removing from ~ to )

I want to make one project a little difference from another project.

For example:

I have below code. 

.... doSomeThing;

//#removeStart

NotReleasedObject nro = new NotReleasedObject();
nro.doSomeThing();

//removeEnd#

....doAnyThing..

as you can see easily, I want to remove from //#removeStart to //removeEnd#. I'm not sure how much code will be written, but all code inside these 2 tags need to be gone.And there can be multiple //#removeStart ... //removeEnd# in one java file.

Here is my test replaceregexp code:

<replaceregexp flags="s" replace="" byline="false" encoding="utf-8">
        <regexp pattern="//#removeStart.*//removeEnd#"/>
        <fileset dir="${extern-source-path}"/>
</replaceregexp>

------ source for victim.

//#removeStart
space 1
//removeEnd#

something

//#removeStart
space 2
//removeEnd#

Now my code is not working properly.The flag -s works, but if one file has more than 1 match, then everything from space 1 to space 2 is gone. ( including something!) What I actually want is space 1,2 gone, but something should be left as it is.

How can I fix it so that it replaces each occurrence 1 by 1 ?

Upvotes: 2

Views: 2300

Answers (2)

David W.
David W.

Reputation: 107040

I've successfully did it with this:

foo.txt

This is a test
--->
###start remove
Blah blah blah
Yadda yadda yadda
###end remove
--->
End of my test

build.xml

<project>
    <replaceregexp file="foo.txt"
        flags="s"
        match="###start remove.+###end remove"
        replace=""/>
</project>

Resulting foo.txt

This is a test
--->

--->
End of my test

It looks like your regex are not quite right. You're regex is saying ##ExternRemove.*, but your file has //#removeStart instead.

I was seeing if I could do this in a copy or concat, because I hate changing files directly. Unfortunately, I hadn't figured out a way to do this.

Upvotes: 0

Jeff Tratner
Jeff Tratner

Reputation: 17076

Your explicit characters don't match: your regex says //##ExternRemove whereas your actual text is either //##removeStart or //#removeStart. If that's an actual regex, the explicit characters need to be exactly the same. Also . doesn't match newlines, you'd need to include that specifically (e.g. [\s\S]*).

Finally, you should make the matching non-greedy if you might have multiple pairs of this match (often this is by adding ?). Taken together, you'd probably want a regular expression like:

//##?removeStart[\s\S]*?//removeEnd##?

That's assuming you want to allow 1-2 # to start a line and that the phrase will always be removeStart/removeEnd. Otherwise you could add options, e.g.

//##?(?:removeStart|ExternRemove)[\s\S]*?//(?:removeEnd|ExternEnd)##?

again, this depends on the ant terminology for non-capturing matches.

Upvotes: 2

Related Questions