Replace different characters using Regex

Question

I would like to remove all single comments // and block comments /* */ of a line inside a string. However, if those comments are inside the characters '', "", [] and/or `` it should not be removed.

Examples:

='a' &
'//b/*aasdsa //dsfds*/'& //comment
'//d' /*aasdsa //dsfds*/ & '//e' & /*aasdsa //dsfds*/ 
& 'c'
[// /* [] */] & h
`// /* [] */` & p

Should be:

='a' &
'//b'& 
'//d'  & '//e' &  
& 'c'
 & h
 & p

I have tried different solutions but I haven't got too far.

let text = "='a' &
'//b'& //comment expression 
'//d' /*aasdsa //dsfds*/ & '//e' & /*aasdsa //dsfds*/ 
& 'c'";
let arrayText = text.split('\n');
arrayText = arrayText.filter(a => a.indexOf('//') !== 0);
arrayContent = arrayContent.map(x =>
x.replace(/[^\(?<=\').*(?=\'$)|\(?<=\[).*(?=\]$](\/\*[\s\S]*?\*\/|\/\/.*)/gm, ''));
text = arrayContent.join(' ');

With the solution that I tried I just got the following:

Text:

='a' &
'//b/*aasdsa //dsfds*/'& //comment
'//d' /*aasdsa //dsfds*/ & '//e' & /*aasdsa //dsfds*/ 
& 'c'
[// /* [] */] & h
`// /* [] */` & p

Result with my solution (that doesn't work)

='a' &
'//b'&
'//d' & '//e' & 
& 'c'
[//]

Expected result:

='a' &
'//b/*aasdsa //dsfds*/'& 
'//d'  & '//e' &  
& 'c'
[// /* [] */] & h
`// /* [] */` & p

I would appreciate if somebody can point me what I am missing or any other hint.

Your 2nd sample input contains a block comment inside of the '//b' string that gets removed. Is that intentional? — Klaycon
– Klaycon, Commented Feb 13, 2020 at 20:06
IMO regex isn't the proper tool for this. I would consider using a tokenizer/parser instead. The regex you have is already heinously complex and difficult to understand, and I fully expect you will continually be running into inputs that breaks the regex. — user47589
– user47589, Commented Feb 13, 2020 at 20:08
See stackoverflow.com/questions/42287216/… for an example of how much more complex your regex can get. — user47589
– user47589, Commented Feb 13, 2020 at 20:09

Klaycon · Accepted Answer · 2020-02-13 21:05:46Z

As mentioned in the comments, regex isn't a proper tool for something normally done by tokenizers. For this particular use case, you could write a simple rule-based parser like this:

const rules = [
  { start: '[', end: ']', remove: false },
  { start: "'", end: "'", remove: false },
  { start: '"', end: '"', remove: false },
  { start: '`', end: '`', remove: false },
  { start: '//', end: {EOL:true}, remove: true },
  { start: '/*', end: '*/', remove: true },
];

function removeComments(str) {
  let start = -1, rule = null;
  //iterate over the input, character by character
  for(let i=0; i<str.length; i++) {
    if(!rule) {
      //if not currently in a 'group' (either string or comment) search for one
      let test = rules.find(r => str.slice(i).startsWith(r.start));
      if(test) {
        rule = test;
        start = i;
      }
    } else {
      //currently in a string or comment, check if it ended
      let end = -1;
      if(str.slice(i).startsWith(rule.end)) {
        end = i + rule.end.length;
      } else if(rule.end.EOL && (str.slice(i).startsWith('\n') || str.slice(i).startsWith('\r\n') || i == str.length - 1)) {
        //special handling for line comments which can end on many conditions
        end = i + 1; 
      }
      if(end > -1) {
        if(rule.remove) {
          //modify str if it was a comment rule - cut out the comment
          str = str.slice(0,start) + str.slice(end);
          i -= end - start;
        }
        rule = null;
      }
    }
  }
  return str;
}

["='a' &",
"'//b/*aasdsa //dsfds*/'& //comment",
"'//d' /*aasdsa //dsfds*/ & '//e' & /*aasdsa //dsfds*/",
"& 'c'",
"[// /* [] */] & h",
"`// /* [] */` & p"].forEach(str => console.log(removeComments(str)));

Note that the output differs from your expected output, because your expected output does things that defy the rules you laid out in the question - it removes everything bordered by [] and ````. as well as block comments contained in strings '/* */'.

@Joe This is already what is done by my solution. I don't understand what other requirements need to be added?

Collectives™ on Stack Overflow

Replace different characters using Regex

1 Answer 1

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related