kelszo / personnummer.js Goto Github PK

The 2.2.0 version includes breaking changes compared to 2.1.0 which is unexpected if it's supposed to adhere to semantic versioning.

As an example, look at the normalise function. In 2.1.0 it returns an empty string if passed invalid input. In 2.2.0 it returns undefined. This means that consuming code that assumes the returned value will always be a string will now suddenly break if the input has not been validated before calling the function.

The changelog does not mention this change at all. It does mention other changes that sound breaking, like changing default options.

When a library does not follow semantic versioning it complicates things for consumers who want to keep their dependencies up to date. It is much easier if library authors try to be strict about breaking changes - then non-major version bumps can mostly be merged right away.

It seems like you are using an arrow function in one place in the code.
In the getBirthplace function: personalNumber/parse.js@156

I tried to replace it with a normal function and then your code works in IE11. I would be nice if you could fix that or maybe run it through babel before publish it on npm?

Thank you!

The documentation states that:

normaliseCIN(str) | Normalises the input to the following format. E.g. 168025216220 -> 802521-6220

But it's the other way around, i.e.

802521-6220 -> 168025216220

According to the documentation, normaliseCIN() or personnummer.parseCIN() will add the digits "16" to a normalised version of an aktiebolag CIN. For example "556007-3495" will become "165560073495".

Is that a standard of some sorts?