Truncate a string straight JavaScript

Use the substring method:

var length = 3;
var myString = "ABCDEFG";
var myTruncatedString = myString.substring(0,length);
// The value of myTruncatedString is "ABC"

So in your case:

var length = 3;  // set to the number of characters you want to keep
var pathname = document.referrer;
var trimmedPathname = pathname.substring(0, Math.min(length,pathname.length));

document.getElementById("foo").innerHTML =
     "<a href='" + pathname +"'>" + trimmedPathname + "</a>"

Here's one method you can use. This is the answer for one of FreeCodeCamp Challenges:

function truncateString(str, num) {
  if (str.length > num) {
    return str.slice(0, num) + "...";
  } else {
    return str;
  }
}

yes, substring. You don't need to do a Math.min; substring with a longer index than the length of the string ends at the original length.

But!

document.getElementById("foo").innerHTML = "<a href='" + pathname +"'>" + pathname +"</a>"

This is a mistake. What if document.referrer had an apostrophe in? Or various other characters that have special meaning in HTML. In the worst case, attacker code in the referrer could inject JavaScript into your page, which is a XSS security hole.

Whilst it's possible to escape the characters in pathname manually to stop this happening, it's a bit of a pain. You're better off using DOM methods than fiddling with innerHTML strings.

if (document.referrer) {
    var trimmed= document.referrer.substring(0, 64);
    var link= document.createElement('a');
    link.href= document.referrer;
    link.appendChild(document.createTextNode(trimmed));
    document.getElementById('foo').appendChild(link);
}

Following code truncates a string and will not split words up, and instead discard the word where the truncation occurred. Totally based on Sugar.js source.

function truncateOnWord(str, limit) {
        var trimmable = '\u0009\u000A\u000B\u000C\u000D\u0020\u00A0\u1680\u180E\u2000\u2001\u2002\u2003\u2004\u2005\u2006\u2007\u2008\u2009\u200A\u202F\u205F\u2028\u2029\u3000\uFEFF';
        var reg = new RegExp('(?=[' + trimmable + '])');
        var words = str.split(reg);
        var count = 0;
        return words.filter(function(word) {
            count += word.length;
            return count <= limit;
        }).join('');
    }

Updated ES6 version

const truncateString = (string = '', maxLength = 50) => 
  string.length > maxLength 
    ? `${string.substring(0, maxLength)}…`
    : string

// demo the above function
console.log(
  truncateString(prompt("", "This is a test"), 4)
)