[tex-k] dvips 5.92b: Missing "LIGKERN"s

[Nick Patavalis]

Hi,

I have noticed that when a glyph is encoded twice, afm2tfm misses
(drops) the LIGKERNs referencing this glyph. For example:

$ cat someenc.enc

/SomeEncoding [
  ...
  /hyphen % 2D
  ...
  /hyphen % 7F
  ...
]

% LIGKERN hyphen hyphen =: endash ; endash hyphen =: emdash ;

...

$ afm2tfm font.afm -t someenc.enc -v font rfont
$ cat font.vpl

...

(LIGTABLE
   (LABEL O 41) (comment exclam)
   (LIG O 140 O 275)
   (STOP)
   (LABEL O 47) (comment quoteright)
   (LIG O 47 O 21)
   (STOP)
   (LABEL O 54) (comment comma)
   (LIG O 54 O 22)
   (STOP)
   (LABEL O 74) (comment less)
   (LIG O 74 O 23)
   (STOP)
   (LABEL O 76) (comment greater)
   (LIG O 76 O 24)
   (STOP)
   (LABEL O 77) (comment question)
   (LIG O 140 O 276)
   (STOP)
   (LABEL O 140) (comment quoteleft)
   (LIG O 140 O 20)
   (STOP)
   (LABEL C f)
   (LIG C l O 35)
   (LIG C i O 34)
   (STOP)
   (LABEL C i)
   (LIG C j O 274)
   (STOP)
   )

...

Even if I remove the second encoding of "hyphen", e.g. by saying:

/SomeEncoding [
  ...
  /hyphen % 2D
  ...
  /.notdef % 7F
  ...
]

afm2tfm reencodes the hyphen at the empty spot and still messes things
up:

$ afm2tfm font.afm -t someenc.enc -v font rfont
$ cat font.vpl

(LIGTABLE
   (LABEL O 25) (comment endash)
   (LIG O 177 O 26)
   (STOP)
   (LABEL O 41) (comment exclam)
   (LIG O 140 O 275)
   (STOP)
   (LABEL O 47) (comment quoteright)
   (LIG O 47 O 21)
   (STOP)
   (LABEL O 54) (comment comma)
   (LIG O 54 O 22)
   (STOP)
   (LABEL O 55) (comment hyphen)
   (LIG O 177 O 25)
   (STOP)
   (LABEL O 74) (comment less)
   (LIG O 74 O 23)
   (STOP)
   (LABEL O 76) (comment greater)
   (LIG O 76 O 24)
   (STOP)
   (LABEL O 77) (comment question)
   (LIG O 140 O 276)
   (STOP)
   (LABEL O 140) (comment quoteleft)
   (LIG O 140 O 20)
   (STOP)
   (LABEL C f)
   (LIG C l O 35)
   (LIG C i O 34)
   (STOP)
   (LABEL C i)
   (LIG C j O 274)
   (STOP)
   )

Quite creative, huh? 

The only way to get a sensible behavior is to say:

$ afm2tfm font.afm -u -t someenc.enc -v font rfont

Is this considered normal?

Thanks in advance
/npat

-- 
flowchart, n.: The innumerate misleading the illiterate. "A
thousand pictures is worth ten lines of code.
  -- Stan Kelly-Bootle