从java中的源代码中删除注释

Question

I want to remove all type of comments statements from a java source code file. Example:

    String str1 = "SUM 10"      /*This is a Comments */ ;   
    String str2 = "SUM 10";     //This is a Comments"  
    String str3 = "http://google.com";   /*This is a Comments*/
    String str4 = "('file:///xghsghsh.html/')";  //Comments
    String str5 = "{\"temperature\": {\"type\"}}";  //comments

Expected Output:

    String str1 = "SUM 10"; 
    String str2 = "SUM 10";  
    String str3 = "http://google.com";
    String str4 = "('file:///xghsghsh.html/')";
    String str5 = "{\"temperature\": {\"type\"}}";

I am using the below regular expression to achieve :

    System.out.println(str1.replaceAll("[^:]//.*|/\\\\*((?!=*/)(?s:.))+\\\\*/", ""));

This gives me wrong result for str4 and str5. Please help me to resolve this issue.

Using Andreas solutions:

        final String regex = "//.*|/\\*(?s:.*?)\\*/|(\"(?:(?<!\\\\)(?:\\\\\\\\)*\\\\\"|[^\\r\\n\"])*\")";
        final String string = "    String str1 = \"SUM 10\"      /*This is a Comments */ ;   \n"
             + "    String str2 = \"SUM 10\";     //This is a Comments\"  \n"
             + "    String str3 = \"http://google.com\";   /*This is a Comments*/\n"
             + "    String str4 = \"('file:///xghsghsh.html/')\";  //Comments\n"
             + "    String str5 = \"{\"temperature\": {\"type\"}}";  //comments";
        final String subst = "$1";

        // The substituted value will be contained in the result variable
        final String result = string.replaceAll(regex,subst);

        System.out.println("Substitution result: " + result);

Its working except str5.

Answer 1

To make it work, you need to "skip" string literals. You can do that by matching string literals, capturing them so they can be retained.

The following regex will do that, using $1 as the substitution string:

//.*|/\\*(?s:.*?)\\*/|("(?:(?<!\\\\)(?:\\\\\\\\)*\\\\"|[^\\r\\n"])*")

See regex101 for demo.

Java code is then:

str1.replaceAll("//.*|/\\*(?s:.*?)\\*/|(\"(?:(?<!\\\\)(?:\\\\\\\\)*\\\\\"|[^\r\n\"])*\")", "$1")

Explanation

//.*                      Match // and rest of line
|                        or
/\*(?s:.*?)\*/            Match /* and */, with any characters in-between, incl. linebreaks
|                        or
("                        Start capture group and match "
  (?:                      Start repeating group:
     (?<!\\)(?:\\\\)*\\"     Match escaped " optionally prefixed by escaped \'s
     |                      or
     [^\r\n"]                Match any character except " and linebreak
  )*                       End of repeating group
")                        Match terminating ", and end of capture group

$1                        Keep captured string literal

Answer 2

As others said, regex is not a good option here. You could use a simple DFA for this task.
Here's an example that will get you intervals of multiple line comments ( /* */ ).
You can do the same way for single line comments ( // -- \\n ).

    String input = ...; //here's your input String

    //0 - source code, 
    //1 - multiple lines comment (start) (/ char)
    //2 - multiple lines comment (start) (* char)
    //3 - multiple lines comment (finish) (* char)
    //4 - multiple lines comment (finish) (/ char)
    byte state = 0; 
    int startPos = -1;
    int endPos = -1;
    for (int i = 0; i < input.length(); i++) {
        switch (state) {
        case 0:
            if (input.charAt(i) == '/') {
                   state = 1;
                   startPos = i;
            }
            break;
        case 1:
            if (input.charAt(i) == '*') {
                state = 2;
            }
            break;
        case 2:
            if (input.charAt(i) == '*') {
               state = 3;
            }
            break;
        case 3:
            if (input.charAt(i) == '/') {
                state = 0;
                endPos = i+1;

                //here you have the comment between startPos and endPos indices,
                //you can do whatever you want with it
            }

            break;
        default:
            break;
        }
    }

Answer 3

{...wishing I could comment...}

I recommend a two-pass process; one based upon end of line (//) the other not (/* */).

I like Pavel's idea; however, I don't see how it checks to make sure the star is the next character after a slash and vice versa on closing out.

I like Andreas' idea; however, I wasn't able to get it to work on multi-line comments.

https://docs.oracle.com/javase/specs/jls/se12/html/jls-3.html#jls-CommentTail

Answer 4

Maybe, it would be best to start with multiple simple expressions, step by step, such as:

.*(\s*\/\*.*|\s*\/\/.*)

to initially remove the inline comments.

Demo

Test

import java.util.regex.Matcher;
import java.util.regex.Pattern;

final String regex = "(.*)(\\s*\\/\\*.*|\\s*\\/\\/.*)";
final String string = "    String str1 = \"SUM 10\"      /*This is a Comments */ ;   \n"
     + "    String str2 = \"SUM 10\";     //This is a Comments\"  \n"
     + "    String str3 = \"http://google.com\";   /*This is a Comments*/\n"
     + "    String str4 = \"('file:///xghsghsh.html/')\";  //Comments\n"
     + "    String str5 = \"{\\\"temperature\\\": {\\\"type\\\"}}\";  //comments";
final String subst = "\\1";

final Pattern pattern = Pattern.compile(regex, Pattern.MULTILINE);
final Matcher matcher = pattern.matcher(string);

// The substituted value will be contained in the result variable
final String result = matcher.replaceAll(subst);

System.out.println("Substitution result: " + result);