How to traverse string and replace ** with b / b

Using regex:

var str = "*Visit* jsduflsd \*fjkdsufdsjlkfds\* djfuldsifjdlsk *jflusdlkfjdsds* jfdisufkljds!";
var res = (" "+str).replace(/([^\])\*([^\*]*)([^\])\*/g, "$1<b>$2$3</b>");
res = res.trim();
res = res.replace(/\\*/g, "*");
document.write(res)

• ( Starts character capture
• [^\] Any character that is different from \
• ) Closes character capture
• \* Starts with * to \ to escape
• ( Starts character capture
• % with% all characters other than [^\*]* repeated 0 or many times
• * Closes character capture
• ) Starts character capture
• ( Any character that is different from [^\]
• \ Closes character capture
• ) Ends with \* a * to escape
• \ , $1 and $2 Used to "paste" what was captured

To get all $3 not preceded by * is to be able to escape them, however this requires a character before \ , if the string starts with bold markdown, it will fail because it will not there is a previous character. To solve this, a * is added before ( ) and it is then extracted with (" "+str) . It is also necessary to "reset" this previous character, so the trim and $1 After making the change to the HTML tags then the markdowns escaped ( $2 to \* )

Most Dynamic Usage Example:

String.prototype.markdown = function(caracter, tag) {
  return (" "+this).replace(new RegExp('([^\\])\${caracter.split('').join('\')}([^\${caracter.split('').join('\')}]*)([^\\])\${caracter.split('').join('\')}', 'g'), '$1<${tag}>$2$3</${tag}>').trim().split('\${caracter}').join(caracter);
}

document.querySelector('textarea').oninput = function() {
  document.querySelector('p').innerHTML = this.value.markdown('*', 'b').markdown('--', 'strike');
}

<textarea cols='75'>--riscado-- \--riscado\-- *negrito* \*negrito\*</textarea>

<p>

From a space in * to load with the example

Regex is certainly the most practical and straightforward way to solve the problem as @GuilheremCostamilam has already shown. The replace getting a string as a replacement works only for one substitution and not all, so you would not have to solve it without regex.

The MDN reinforces this point clearly

  

If pattern is a string, only the first occurrence will be replaced

However it would also be possible to resolve manually, albeit much more laborious. For this, it would have to have a boolean variable to know if it will add the beginning of the tag or end of the tag, and every time it finds a * it makes the change by the start / closing according to boolean. After each change you have to toggle the Boolean.

Example:

var str = "Visit *jsduflsd* fjkdsufdsjlkfds djfuldsifjdlsk *jflusdlkfjdsds* jfdisufkljds!";

let inicio = true, res = "";
for (let letra of str){
  if (letra === '*'){
    res += inicio ? "<b>": "</b>";
    inicio = !inicio;
  }
  else res += letra;
}

document.write(res);

If there is a possibility that there is a * lost without its pair and we do not want it to be bold, then some additional concerns are already needed, which will complicate and increase the code a bit.

Example:

var str = "Visit *jsduflsd* fjkdsufdsjlkfds djfuldsifjdlsk *jflusdlkfjdsds jfdisufkljds!";

let emBold = false, res = "", blocoBold ="";
for (let letra of str){
  if (letra === '*'){
    emBold = !emBold;
    if (emBold)
      blocoBold = "";
    else //só no fecho é que adiciona o bloco todo
      res += "<b>" + blocoBold + "</b>";
  }
  else {
    if (emBold)
      blocoBold += letra;
    else 
      res += letra;
  }
}

res += blocoBold;//adiciona o que ficou por fechar sem as tags
document.write(res);