替换所有字符串 int java 源代码

Question

I am trying to replace every declared String that is inside of a Java formatted source code (using java itself). A source code to remove the strings from may look like this (>...< indicates what should be matched):

package a.b;

import b.a.*;

@SomeAnnotation(>"abc"<)
public class XY extends YZ {

    String s = >"Hello \"World\""<;

    public static void main(String[] args) {
        System.out.println(>"Hello World"<);
    }
}

Imagine this all stored in one string, possibly without the newlines. I now need a regex that can find all of these. I am having trouble with the escaped quotation marks inside of the strings. My best bet is ".*[^\]" .

It does not matter whether it will detect these inside of comments, as long as it does not excede the comment.

Answer 1

If you read in the source code in another java application with a BufferedLineReader and check every line you could maybe use the following regex:

(?:\>\")(([\w\s]*)(\\"(?!\<)))*(?:\"\<)

However you would want to escape all of the backslashes when used inside the java replace method.

eg:

lineAsString.replace((?:\\>\\")(([\\w\\s]*)(\\\\"(?!\\<)))*(?:\\"\\<),"youReplacementString");

Maybe additional escape sequences might be necessary. Also, you might want to add further special chars to be allowed inside the string beside \w and \s .

I suggest to paste that regex into a regex checker like https://regex101.com/ and your sample code for more explanations.