nodejs code demonstrating how Unicode strings from the same source can get changed so they are not the same, and how to renormalize then for comparison.

nodejs unicode normalizaition.js

var unorm = require('unorm');
// These strings appear to be the same, but are NOT!
// String 1 is from a [track][location][path] extracted with Scripting Bridge from iTunes on Mac running an HFS+ filesystem
// String 2 is a path from an HFS+ filesystem mounted on Linux, then stored in an SQLite3 TEXT field.
// The file String 1 was derived from was copied from the file String 2 was derived from using rsync.

// Both strings then printed in Terminal.app and pasted into Emacs.

// Demonstration that two strings that appear the same might not be, even when they came from the same place originally.

// I have no idea if the normalization changed because Linux HFS+ is
// different from Mac HFS+, or if SQLite3 renormalizes Unicode, or if
// iTunes did it, or rsync, or Scripting Bridge.

var string1 = "Jan Lindblad/i en klosterträdgård/01 - Våren.mp3";
var string2 = "Jan Lindblad/i en klosterträdgård/01 - Våren.mp3";

    ['nfd','nfc','nfkd','nfkc'].forEach(function(normalization) {

        if (unorm[normalization](string1) === string2) {
            console.log('the strings match with string1 %s\'d', normalization);
        } else {
            console.log('strings not match with string1 %s\'d', normalization);
        }
        if (unorm[normalization](string2) === string1) {
            console.log('the strings match with string2 %s\'d', normalization);
        } else {
            console.log('strings not match with string2 %s\'d', normalization);
        }
    });

// Result: 
// strings not match with string1 nfd'd
// the strings match with string2 nfd'd
// the strings match with string1 nfc'd
// the strings match with string2 nfc'd
// strings not match with string1 nfkd'd
// the strings match with string2 nfkd'd
// the strings match with string1 nfkc'd
// the strings match with string2 nfkc'd