std.unicode: add utf16 byte length and codepoints counting routines

[matu3ba]

• comptime and runtime tests are based on tests for counting utf8 code points

[matu3ba]

justification is existence of utf8CountCodepoints
personal goal: finish the basic stuff for rosettacode like http://www.rosettacode.org/wiki/String_length#Zig

[ifreund]

I notice that the doc comment for utf8CountCodepoints() says /// Asserts that the data is valid UTF-8. but it doesn't actually seem to assert this and rather returns an error.

It'd be worth considering whether that behavior is more desirable. At the very least the doc comment should be fixed.

[matu3ba]

Suggestions by @ifreund were applied. Note, that I did not modify that function.
No idea, why git was unable to keep commit history during rebase on top of master, which did not touch unicode.zig.