How do you remove double quotes from the image of a JavaCC token?

Question

In JavaCC, I'm accepting strings that are under the condition:

   < STRING : "\"" ("\\" ~[] | ~["\"","\\"] )* "\"" >

So the image ends up printing anything that is a string but with another set of double quotes.

For example, I'll input: "This is a sentence."

And the value will result : ""This is a sentence."" stored in a String variable.

Is there a way in Java to remove the extra set of double quotes so that it only prints: "This is a sentence."?

What do you do with a string that extra quotes get stored? You also talk about printing - give us a [mcve] for what you get. — M. Prokhorov, Oct 06 '20 at 19:34

score 2 · Answer 1 · answered Oct 13 '20 at 21:35

If the input matched by your token is "Hello" then the value of the image field of the token will be a 7 character string whose first and last characters are double quote characters. They're not really extra they were they in the input. Say you write

void foo() : {
    Token t ; }
{
    t = <STRING>
    { System.out.println( t.image ) ; }
}

That'll print 7 characters and then a newline.

Now if you don't want those characters, well, @Bryan's answer will do it.

void foo() : {
    Token t ; }
{
    t = <STRING>
    { { String nakedImage = t.image.substring(1,t.image.length()-1) ;
        System.out.println( nakedImage ) ; } }
}

It should be noted that no quotes are removed. String objects in Java are immutable, meaning they can't be changed. What really happens is that a new String object gets created and a reference to it is assign to the nakedImage variable. The String object that t.image is a reference to remains the same.

Now you still have the problem of dealing with the back slashes. If the input is "Hello\tWorld", then t.image will be 14 characters long and nakedImage will be 12 characters long. What I do at this point is to run the string through a function builds a new string that has single characters where the nakedImage has escape sequences. So the result of that function on this example would be 11 characters long.

void foo() : {
    Token t ; }
{
    t = <STRING>
    { { String nakedImage = t.image.substring(1,t.image.length()-1) ;
        String unescapedImage = unescape( nakedImage ) ;
        System.out.println( unescapedImage ) ; } }
}

Here's such a function, based on one I wrote for a Java compiler.

private static String unescape( String str ) {
    StringBuffer result = new StringBuffer() ;
    for( int i=0, len = str.length() ; i<len ; ) {
        char ch = str.charAt(i) ;

        // Set ch and increment i ;
        if( ch == '\\' )  {
            ch = str.charAt(i+1) ;
            switch( ch ) {
                case 'b' : ch = '\b' ; i += 2 ; break ;
                case 't' : ch = '\t' ; i += 2 ; break ;
                case 'n' : ch = '\n' ; i += 2 ; break ;
                case 'f' : ch = '\f' ; i += 2 ; break ;
                case 'r' : ch = '\r' ; i += 2 ; break ;
                case '"' : case '\'' : case '\\' : i+= 2 ; break ;
                default: 
                    /*TODO Deal with errors. */ } }
        else {
            i += 1 ; }
        result.append( ch ) ; }
    return result.toString() ;
}

Bryan · Accepted Answer · 2020-10-06T19:31:37.893

0

str = str.substring(1,str.length()-1)

alternate for Javacc

Parsing Strings with JavaCC

edited Oct 06 '20 at 19:31

answered Oct 06 '20 at 19:16

Bryan

159
10

3

I'm not familiar with javacc but such a curt answer lets me think you might not be either and are just answering with some basic java string manipulation. OP might not have any `str` to run java code on – Aaron Oct 06 '20 at 19:19
Is there a way in **Java** to remove the extra set of double quotes so that it only prints: "This is a sentence."? But yes if solely using Javacc this probably isn't the answer. – Bryan Oct 06 '20 at 19:21
No worries, I was just mentioning it probably isn't that easy. The question definitely could use some improvement, the title alone will certainly attract the kind of answer you gave. – Aaron Oct 06 '20 at 19:25
Agreed, lets never let this kind of stuff come between our friendship again. ;) – Bryan Oct 06 '20 at 19:55
1

Of course, our intercontinental friendship shall not suffer from such trivial matters ! – Aaron Oct 06 '20 at 20:04
1

This worked for me thank you! And idk why ppl downvoted this. I clearly asked how to do it in Java even though I was explaining the way I obtained that token from JavaCC. – peach Oct 12 '20 at 18:29
1

I agree. It's the right answer. I think the problem was more that the question was unclear about whether it's a Java problem or a JavaCC problem. Anyway, the up and down votes are now at least equal. – Theodore Norvell Oct 13 '20 at 21:58

How do you remove double quotes from the image of a JavaCC token?

2 Answers2