Regular Expression bug: characters in a positive character group must redundantly be escaped

[KlausKi]

I've been struggling with the Regular Expression implementation of ExtendScript quite a bit, because it doesn't parse Regular Expressions correctly.

The following information may be helpful to others:

The ExtendScript Regular Expression parser fails to correctly parse positive character groups (characters enclosed in square brackets). Quantifiers and the dot (".") don't make sense in a positive character group, so characters representing these quantifiers are not escaped in positive character groups. However, for the ExtendScript Regular Expression parser to interpret positive character groups, you MUST escape those characters.

For example:

A valid Regular Expression like this:

/[?.*]/

… must be written like this in ExtendScript:

/[\?\.\*]/

… for the ExtendScript Regular Expression parser to accept the Regular Expression.

[K.Daube]

MHM, I don't get this problem in FM-15 (2019)

• Neiter with the standard FM Find/Change dialogue
• 
• Nor with my own FMfindReplace scipt.
• 
• FMfindRepl uses this code for replacement if something was found by the ES Find method:

  KLD_F.ReplaceInString = function (string, sSearch, sReplace, bCase, bWord, bRegEx) {
  /*            Find substring with options and replae it
  Arguments     string          String in which to search
                sSearch         String to be searched
                sReplace        String replacing the search
                bCase           true for case sentitive search
                bWord           true so search whole words only
                bRegEx          true to interprete sSearch as RegEx
  Returns       The string with the replacement
  Called by     ButtonReplace
  Calling       -
  History       2021-10-18
  */
  var eReg0, eReg1, eReg2, eReg2, eReg3, eRegR, sResult, sFind;
  
    if (string === undefined) { string = ""; }
    if (!bRegEx) { // escape the 12 syntactical chars
      sFind = sSearch.replace (/([\\\^\$\.\|\?\*\+\(\)\[\{])/g,"\\$1"); 
    }
  
    eReg0 = new RegExp (sFind, "i");          // case insesitive, no words
    eReg1 = new RegExp (sFind);               // case sensitive, no words
    eReg2 = new RegExp (sFind + "\\b", "i");  // case insesitive, word
    eReg3 = new RegExp (sFind + "\\b");       // case sensitive, word
  
    if (bRegEx) {
      eRegR = new RegExp (sSearch);   // interprete as RegEx
      sResult = string.replace (eRegR, sReplace);
    } else {
      if (!bCase && !bWord) {
        sResult = string.replace (eReg0, sReplace);
      } else if (bCase && !bWord) {
        sResult = string.replace (eReg1, sReplace);
      } else if (!bCase && bWord) {
        sResult = string.replace (eReg2, sReplace);
      } else {
        sResult = string.replace (eReg3, sReplace);
      }
    }
    return sResult;
  } //--- end ReplaceInString ​

[Manan Joshi]

Hi @KlausKi,

Which application are you targetting your code to. I have never faced this issue, though mainly I have worked with InDesign and a bit with Illustrator and PS. I just tried the following code on InDesign 2023 and it worked for me

var a = "hello? Nwo."
var r = a.search(/[?*.]/)
alert(r)

-Manan

[KlausKi]

I've been programming a script for PhotoShop CC 2022:

https://github.com/SetTrend/Raw-Image-Converter/blob/main/includes/_FileOperations.jsx#L111