Not many tests for multibyte Unicode characters

It's great to find a library like this, string manipulation in JS often leaves much to be desired. However, the limited scope of testing with strings containing multibyte Unicode characters is concerning. Multibyte characters often create quite a few headaches. (These's a [nice writeup about the challenges here](https://mathiasbynens.be/notes/javascript-unicode))

There are already a few tests that check the processing of multibyte characters, e.g. [decencode.test.js](https://github.com/dleitee/strman/blob/3f85db8997a0867f1d9a6f4785adbab219bfaf3b/test/decencode.test.js#L5). Having more tests to ensue safe handling of strings containing multibyte characters would be a wonderful addition!

Understandably, it might not be possible for all functions to safely support multibyte characters. Perhaps this could be documented as a limitation on these functions?

## Simple example of current behavior:

```
$ pwd
~/strman
$ node -v
v7.9.0

$ node -e 'console.log("A".length);'
1
$ node -e 'console.log("漢".length);'
1
$ node -e 'console.log("\u200B".length);'
1
$ node -e 'console.log("🐨".length);'
2

$ node -e 'console.log(require("./lib/strman.js").last("A"));'
A
$ node -e 'console.log(require("./lib/strman.js").last("漢"));'
漢
$ node -e 'console.log(require("./lib/strman.js").last("🐨"));'
🐨

$ node -e 'console.log(require("./lib/strman.js").reverse("A"));'
A
$ node -e 'console.log(require("./lib/strman.js").reverse("漢"));'
漢
$ node -e 'console.log(require("./lib/strman.js").reverse("🐨"));'
��
```