The string to be encoded contains characters outside of the Latin1 range.

[gkkirilov]

As I am using some Cyrillic symbols I had to add this inside the btoa method this code
.replace(/[\u00A0-\u2666]/g, function(c) { return '&#' + c.charCodeAt(0) + ';'; })

[jekuer]

@gkkirilov , does this cover all relevant cyrillic symbols? Do you want to push a Pull Request for that?

[gkkirilov]

I think it does, but don't want to add more complexity to the library, maybe lets just leave it as a reference point if anyone needs it.

[jekuer]

I included your fix in the latest version. It is not ideal, but at least removes the console error. Left a comment in the code for future optimization.