[Groff] mapping of glyphs to Unicode

[Bruno Haible]

Hi,

One should expect that the mapping from groff glyphs to Unicode is device
independent, right? And that glyphuni.cpp represents this mapping?

Comparing font/devhtml/R.proto
          font/devutf8/R.proto
          src/libs/libgroff/glyphuni.cpp
gives the following differences:

- devhtml maps "hy" and "-" to U+002D.
  devutf8 and glyphuni.cpp map "hy" and "-" to U+2010.

- devhtml maps "la" to U+2329, "ra" to U+232A.
  devutf8 and glyphuni.cpp map "la" to U+27E8, "ra" to U+27E9.
  U+2329..232A are LEFT/RIGHT-POINTING ANGLE BRACKET,
  U+27E8..27E9 are MATHEMATICAL LEFT/RIGHT ANGLE BRACKET and were introduced
  in Unicode 3.2.
  The glyphs are very similar (see http://www.unicode.org/charts/symbols.html
  under "Misc. Math Symbols A" and "Miscellaneous Technical").
  Why prefer one over the other?

- devhtml and devutf8 map "[" to U+005B.
  glyphuni.cpp doesn't.
  Why?

- devhtml and devutf8 map "\" to U+005C.
  glyphuni.cpp doesn't.
  Why?

- devhtml and devutf8 map "]" to U+005D.
  glyphuni.cpp doesn't.
  Why?

- devhtml and devutf8 map "\`" to U+0060.
  glyphuni.cpp doesn't.
  Why?

- devhtml and devutf8 map "\'" to U+00B4.
  glyphuni.cpp doesn't.
  Why?

- glyphuni.cpp maps "shc" to U+00AD.
  devhtml and devutf8 don't.
  Why?

- devhtml and devutf8 map "+f" to U+03C6 and "*f" to U+03D5.
  glyphuni.cpp does the opposite: maps "*f" to U+03C6 and "+f" to U+03D5.
  According to the info in the groff_char.7 man page,
    "symbol `\[*f]' always denotes the stroked  version of phi, and
     `\[+f]' the curly variant."
  this means that glyphuni.cpp is wrong. I vote for following current
  standards.

- devhtml and devutf8 map "\-" to U+2212.
  glyphuni.cpp doesn't.
  Why?

- devhtml and devutf8 map "sqrt" to U+221A.
  glyphuni.cpp doesn't.
  Why?

- devhtml and devutf8 map "<<" to U+226A and ">>" to U+226B.
  glyphuni.cpp does the opposite: maps ">>" to U+226A and "<<" to U+226B.
  IMO glyphuni.cpp is wrong.

Bruno