add new utf8proc_codepoint_units function

[markand]

I propose a simple function to determine the number of UTF-8 bytes that a codepoint require. This mostly addresses the issue #170

Example:

utf8proc_codepoint_units(233)    // é -> 2
utf8proc_codepoint_units(0x20ac) // € -> 3

[stevengj]

Maybe utf8proc_sizeof_char?

[markand]

Hmm, the term sizeof in C has special meaning, here the functions is meant to obtain a variable number between 0 and 4 which to me is the opposite of what "sizeof" sounds to me which always evaluates to a constant object size.

Other alternatives:

• utf8proc_codepoint_utf8len
• utf8proc_codepoint_mblength (multibyte length)
• utf8proc_codepoint_length (pretty name but still confusing at a glance IMHO)

[stevengj]

sizeof refers to a size in bytes, which is why it seems appropriate here.

[markand]

Updated.

[stevengj]

It would be good to check consistency with utf8proc_encode_char, i.e. check utf8proc_sizeof_char(c) == utf8proc_encode_char(c, buf) where you declare utf8proc_uint8_t buf[4].

[stevengj]

Not sure what the Travis failure is here?