Update data tables to Unicode 7.0.0

[jiahao]

Updates:

1. Updates the data_generator.rb script. This script now runs on a modern version of ruby (>1.8) and has the hard-coded data tables replaced with file reads from the appropriate Unicode data (UNIDATA) files.
2. Provides a new Makefile target, update, which automatically downloads the relevant UNIDATA and runs data_generator.rb to produce the file utf8proc_data.c.new.
3. Updates utf8proc_data.c to the output generated by running make update against UNIDATA v7.0.0

Observations:

1. There are #defined constants in utf8proc.c which may in principle have changed from v5.0 to v7.0, such as the constants marking the location of Hangul, Unihan, etc. I haven't checked them and it's probably not worth recomputing for each new Unicode version.
2. It looks like utf8proc implements an internal processing mode called LUMP, which is briefly described in lump.txt. As far as I can tell, this is a custom normalization mode which is separate from the Unicode standard, but I think we'll want to use these.

Closes #1

[stevengj]

This is great!

As a sanity check, if you run on the Unicode 5.0.0 files then does it reproduce the old utf8proc_data.c?

[stevengj]

And yes, LUMP is a custom normalization of utf8proc, which we should keep as-is for API compatibility.

[jiahao]

As a sanity check, if you run on the Unicode 5.0.0 files then does it reproduce the old utf8proc_data.c?

Yes, see #8