Cleanup and improvements in unit mapping

[Enchufa2]

Before tackling #289, I realized that unit mapping was not entirely correct: we were mapping symbols/names to units, but not units to symbol/names, and apparently both are required. This PR solves this.

I've also cleaned up the C interface by distinguishing two kinds of functions:

• Thin wrappers around udunits2 functions, following the name convention R_<udunits2_name>, e.g., R_ut_parse.
• More high-level functions, following the name convention ud_<something>, e.g., ud_convert.

Finally, I've removed deprecated functions and created a man page specifically to document all the defunct ones.

[codecov]

Codecov Report

Merging #290 (6e7d3ad) into master (c77471a) will decrease coverage by 1.30%.
The diff coverage is 91.56%.

@@            Coverage Diff             @@
##           master     #290      +/-   ##
==========================================
- Coverage   94.42%   93.12%   -1.31%     
==========================================
  Files          18       18              
  Lines         879      858      -21     
==========================================
- Hits          830      799      -31     
- Misses         49       59      +10     

Impacted Files	Coverage Δ
R/database.R	80.55% <0.00%> (ø)
R/make_units.R	89.83% <ø> (ø)
R/init.R	9.09% <50.00%> (ø)
src/udunits.cpp	90.19% <93.33%> (-9.81%)	⬇️
R/udunits.R	100.00% <100.00%> (ø)
R/user_conversion.R	87.50% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c77471a...6e7d3ad. Read the comment docs.

[Enchufa2]

Codecov complains, but the few lines without coverage are on purpose.

[Enchufa2]

@edzer Kind reminder. :) Did you have time to take a look?