Re: [Groff] new backend for UTF-8 output

[Bruno Haible]

Markus Kuhn wrote:
>     Ideally, the characters that it would use in postscript output
>     should be mapped to Unicode precisely according to
> 
>        http://partners.adobe.com/asn/developer/typeforum/unicodegn.html

OK, I've double-checked the devutf8 "font" file against the devps one.
Here is a patch with the minor corrections that came out.

Bruno


2000-01-12  Bruno Haible  <address@hidden>

        * devutf8/R.proto: Add mappings for ti, Fn, st, an. Change mappings
          of Im, Re.

*** groff-1.15/devutf8/R.proto.bak      Sat Jan  8 20:53:31 2000
--- groff-1.15/devutf8/R.proto  Wed Jan 12 23:42:57 2000
***************
*** 115,120 ****
--- 115,121 ----
  }     "
  a~    24      0       0x007E
  ~     "
+ ti    "
  r!    24      0       0x00A1
  char161       "
  ct    24      0       0x00A2
***************
*** 307,312 ****
--- 308,314 ----
  /l    24      0       0x0142
  OE    24      0       0x0152
  oe    24      0       0x0153
+ Fn    24      0       0x0192
  ah    24      0       0x02C7
  ab    24      0       0x02D8
  a.    24      0       0x02D9
***************
*** 386,393 ****
  fc    24      0       0x203A
  rn    24      0       0x203E
  f/    24      0       0x2044
! Im    24      0       0x2110
! Re    24      0       0x211B
  tm    24      0       0x2122
  Ah    24      0       0x2135
  <-    24      0       0x2190
--- 388,395 ----
  fc    24      0       0x203A
  rn    24      0       0x203E
  f/    24      0       0x2044
! Im    24      0       0x2111
! Re    24      0       0x211C
  tm    24      0       0x2122
  Ah    24      0       0x2135
  <-    24      0       0x2190
***************
*** 407,412 ****
--- 409,415 ----
  gr    24      0       0x2207
  mo    24      0       0x2208
  nm    24      0       0x2209
+ st    24      0       0x220B
  \-    24      0       0x2212
  mi    "
  **    24      0       0x2217
***************
*** 443,448 ****
--- 446,452 ----
  la    24      0       0x2329
  ra    24      0       0x232A
  CR    24      0       0x240D
+ an    24      0       0x2500
  br    24      0       0x2502
  bv    "
  rk    24      0       0x251D
*** groff-1.15/devutf8/NOTES.bak        Sat Jan  8 20:53:12 2000
--- groff-1.15/devutf8/NOTES    Wed Jan 12 23:51:35 2000
***************
*** 1,11 ****
  The following from groff_char.7 are unmapped:
  
-        f        \(Fn            florin
         <=       \(lh            handleft
         =>       \(rh            handright
-        -)       \(st            suchthat
         p        \(wp            weierstrass      Weierstrass p
!        -        \(an            arrowhorizex     horizontal arrow extension
  
  The following from the original troff manual (by Ossanna and Kernighan) are
  unmapped:
--- 1,11 ----
  The following from groff_char.7 are unmapped:
  
         <=       \(lh            handleft
         =>       \(rh            handright
         p        \(wp            weierstrass      Weierstrass p
! 
! (The latter is mapped to U+2118 in the Adobe glyph list, but this is wrong:
! U+2118 is a "SCRIPT CAPITAL P", not a Weierstrass p).
  
  The following from the original troff manual (by Ossanna and Kernighan) are
  unmapped:
***************
*** 31,36 ****
--- 31,77 ----
  
  \(bv has been mapped to 0x2502, not 0x2759, because it appears to mean
  "bar vertical", not "bold vertical".
+ 
+ The following lines could be added, to define names which are known to
+ "devps" but are not documented and not known to "devdvi".
+ space 24      0       0x0020
+ vS    24      0       0x0160
+ vs    24      0       0x0161
+ :Y    24      0       0x0178
+ vZ    24      0       0x017D
+ vz    24      0       0x017E
+ 3d    24      0       0x2234
+ nb    24      0       0x2284
+ product       24      0       0x220F
+ sum   24      0       0x2211
+ SP    24      0       0x2660
+ CL    24      0       0x2663
+ HE    24      0       0x2665
+ DI    24      0       0x2666
+ 
+ "devps" maps \(a~ and ~ to asciitilde, which is equivalent to 0x02DC, but
+ this glyph is just too small. We map them to 0x007E instead.
+ 
+ "devps" maps \(a^ and ^ to circumflex, which is equivalent to 0x02C6, but
+ this glyph is just too small. We map them to 0x005E instead.
+ 
+ "devps" maps *m to mu, which is equivalent to either 0x00B5 or 0x03BC. We
+ map it to 0x03BC.
+ 
+ "devps" maps *U to Upsilon1, which is equivalent to 0x03D2. We map it to
+ 0x03A5 instead.
+ 
+ "devps" maps *W to Omega, which is equivalent to either 0x2126 or 0x03A9. We
+ map it to 0x03A9.
+ 
+ "devps" maps *D to Delta, which is equivalent to either 0x2206 or 0x0394. We
+ map it to 0x0394.
+ 
+ "devps" maps CR to carriagereturn, which is equivalent to 0x21B5. We map it
+ to 0x240D instead.
+ 
+ "devps" maps md to dotmath, which is equivalent to 0x22C5. We map it to 0x00B7
+ instead.
  
  The line
  char173       24      0       0x00AD