Regex to find first comment in a JS file

Question

I am trying to write a regex (in JavaScript) that will match a multi line comment at the beginning of a JS file.

So far, I came up with this: /^(\/\*[^\*\/]*\*\/)/g

It works for a single line comment: https://regex101.com/r/ZS5PVI/1

But my problem is that it does not work for a multi line comment: https://regex101.com/r/ZS5PVI/2

Do you have any ideas how to solve it?

@SamIam, that's why you need to use [a proper parser](http://stackoverflow.com/a/11141960/201952). — josh3736, Jun 21 '12 at 15:56

score 4 · Accepted Answer · edited May 23 '17 at 12:07

Like HTML, JavaScript cannot be parsed by regular expressions. Attempting to do so correctly is futile.

Instead, you must use a parser that will correctly transform JavaScript source code into an AST, which you may inspect programmatically. Fortunately, there's libraries that do the parsing for you.

Here's a working example that outputs the AST of this code:

/* this is a
multi-line
comment */

var test = "this is a string, /* and this is not a comment! */";

// ..but this is

Which gets us:

[
  "toplevel",
  [
    [
      {
        "name": "var",
        "start": {
          "type": "keyword",
          "value": "var",
          "line": 5,
          "col": 4,
          "pos": 57,
          "endpos": 60,
          "nlb": true,
          "comments_before": [
            {
              "type": "comment2",
              "value": " this is a\n    multi-line\n    comment ",
              "line": 1,
              "col": 4,
              "pos": 5,
              "endpos": 47,
              "nlb": true
            }
          ]
        },
        "end": {
          "type": "punc",
          "value": ";",
          "line": 5,
          "col": 67,
          "pos": 120,
          "endpos": 121,
          "nlb": false,
          "comments_before": []
        }
      },
      [
        [
          "test",
          [
            {
              "name": "string",
              "start": {
                "type": "string",
                "value": "this is a string, /* and this is not a comment! */",
                "line": 5,
                "col": 15,
                "pos": 68,
                "endpos": 120,
                "nlb": false,
                "comments_before": []
              },
              "end": {
                "type": "string",
                "value": "this is a string, /* and this is not a comment! */",
                "line": 5,
                "col": 15,
                "pos": 68,
                "endpos": 120,
                "nlb": false,
                "comments_before": []
              }
            },
            "this is a string, /* and this is not a comment! */"
          ]
        ]
      ]
    ]
  ]
]

Now it's just a matter of looping over the AST and extracting what you need.

Niet the Dark Absol · Answer 2 · 2012-06-21T15:25:35.163

2

Your suggested Regex desn't work because there is a * in the comment. Additionally, it will only look for comments that are right at the beginning of the file.

Try using this instead:

/\/\*[\s\S]*?\*\//

edited Jun 21 '12 at 15:25

answered Jun 21 '12 at 15:19

Niet the Dark Absol

320,036
81
464
592

That matches everything between the first `/*` to the last `*/`. – sachleen Jun 21 '12 at 15:22

score 2 · Answer 3 · answered Jun 21 '12 at 15:23

2

There's a pretty good discussion of this problem at this link. Does that help you?

His solution was:

/\*([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\*+/

answered Jun 21 '12 at 15:23

score 1 · Answer 4 · answered Jun 21 '12 at 15:21

1

Try

/\*([^*]|[\r\n]|(\*+([^*/]|[\r\n])))*\*+/

This page goes into some detail on how to find multi-line comments.

answered Jun 21 '12 at 15:21

sachleen

30,730
8
78
73

LOL I have to +1 this since we posted the same link. :) – Jun 21 '12 at 15:23

score 0 · Answer 5 · answered Jun 21 '12 at 15:21

0

Here's one that will match any multi-line or single-line comments:

/(\/\*.*?\*\/|\/\/[^\n]+)/

If you just want multi-line matches, ditch the second half:

/\/\*.*?\*\//

For both of these, make sure you have s flag set so the . matches new lines.

answered Jun 21 '12 at 15:21

Cecchi

1,525
9
9

score 0 · Answer 6 · answered Jun 21 '12 at 22:50

I'm no javascript expert but it seems C/C++ comments have to be taken into account.
Properly done means quotes have to be accounted for in the process (escapes and all that).

Below are two regex methods that work. Regex 1 finds the first C-style comment directly, as soon as it matches, it is found. Regex 2 is a general case. It finds either C style, C++ style, or non-comments, is global, and allows you to break when you find what you want.

Tested here http://ideone.com/i1UWr

Code

var js = '\
// /* C++ comment  */      \\\n\
   /* C++ comment (cont) */  \n\
/* t "h /* is"               \n\
 is first C-style /*         \n\
//  comment */               \n\
and /*second C-style*/       \n\
then /*last C-style*/        \n\
';

var cmtrx1 = /^(?:\/\/(?:[^\\]|\\\n?)*?\n|(?:"(?:\\[\S\s]|[^"\\])*"|'(?:\\[\S\s]|[^'\\])*'|[^\/"'\\]*))+(\/\*[^*]*\*+(?:[^\/*][^*]*\*+)*\/)/;

var cmtrx2 = /(\/\*[^*]*\*+(?:[^\/*][^*]*\*+)*\/)|(\/\/(?:[^\\]|\\\n?)*?)\n|(?:"(?:\\[\S\s]|[^"\\])*"|'(?:\\[\S\s]|[^'\\])*'|[\S\s][^\/"'\\]*)/g;

//
print ('Script\n===========\n'+js+'\n===========\n\n');

var match;
//
print ("Using Regex 1\n---------------\n");
if ((match=cmtrx1.exec( js )) != null)
    print ("Found C style comment:\n'" + match[1] + "'\n\n");
//
print ("Using Regex 2\n---------------\n");
while ((match=cmtrx2.exec( js )) != null)
{
   if (match[1] != undefined)
   {
        print ("- C style :\n'" + match[1] + "'\n");
        // break;     // uncomment to stop after first c-style match
   }
   // comment this to not print it
   if (match[2] != undefined)
   {
        print ("- C++ style :\n'" + match[2] + "'\n");
   }
}

Output

Script
===========
// /* C++ comment  */      \
   /* C++ comment (cont) */  
/* t "h /* is"               
 is first C-style /*         
//  comment */               
and /*second C-style*/       
then /*last C-style*/        

===========


Using Regex 1
---------------

Found C style comment:
'/* t "h /* is"               
 is first C-style /*         
//  comment */'


Using Regex 2
---------------

- C++ style :
'// /* C++ comment  */      \
   /* C++ comment (cont) */  '

- C style :
'/* t "h /* is"               
 is first C-style /*         
//  comment */'

- C style :
'/*second C-style*/'

- C style :
'/*last C-style*/'

Expanded Regex's

Regex 1:

/^(?:\/\/(?:[^\\]|\\\n?)*?\n|(?:"(?:\\[\S\s]|[^"\\])*"|'(?:\\[\S\s]|[^'\\])*'|[^\/"'\\]*))+(\/\*[^*]*\*+(?:[^\/*][^*]*\*+)*\/)/

     /^
     (?:
          \/\/
          (?: [^\\] | \\\n? )*?
          \n
       |
          (?:
               "
               (?: \\[\S\s] | [^"\\] )*
               "
            |  '
               (?: \\[\S\s] | [^'\\] )*
               '
            |  [^\/"'\\]*
          )
     )+
1    (
          \/\* [^*]* \*+
          (?: [^\/*] [^*]* \*+ )*
          \/
1    )
     /


Regex 2:

/(\/\*[^*]*\*+(?:[^\/*][^*]*\*+)*\/)|(\/\/(?:[^\\]|\\\n?)*?)\n|(?:"(?:\\[\S\s]|[^"\\])*"|'(?:\\[\S\s]|[^'\\])*'|[\S\s][^\/"'\\]*)/g

     /
1    (
          \/\* [^*]* \*+
          (?: [^\/*] [^*]* \*+ )*
          \/
1    )
  |
2    (
          \/\/
          (?: [^\\] | \\\n? )*?
2    )
     \n
  |
     (?:
          "
          (?: \\[\S\s] | [^"\\] )*
          "
       |  '
          (?: \\[\S\s] | [^'\\] )*
          '
       |  [\S\s][^\/"'\\]*
     )
     /g

Regex to find first comment in a JS file

6 Answers6