Re: 8r fonts

[Thierry Bouche <Thierry.Bouche@ujf-grenoble.fr>]

there is a pb with compwordmark in T1 fonts:
here is what you find in ptmr8t.vpl

(CHARACTER O 27
   (CHARWD R 0.0)
   (MAP
      )
   )

also I thought cwm should be 1ex high?

apart from this, found no bug. Regarding 8r.mtx, it's obvious in times
for fè that suppressing the automatic kern is an
improvement. Unfortunately, it's not enough for fï (BTW I'm building a
gs35nfss based on urw fonts: they have much more kern pairs for
accented letters but forget fï as well, on the other hand they have
_positive_ kerns fi rather than faking the lig so that the automatic
ff ligs by fontinst are completely different and are not width
compatible with those of ptm (although individual chars are...) and
_in this case_ using the kern pair f i would improve the fï problem as
opposed to what happens with adobe's afm). Maybe one day foundries
will provide good afms and all this discussion will loose its object.

[here is what I got with my tools:
(CHARACTER D 23 (COMMENT compwordmark)
   (CHARWD R 0.0)
   (CHARHT R 0.0)
   (CHARDP R 0.0)
   (MAP
      )
   )
]

I propose to do a dotlessj gs35nfss archive so that everybody can test
it. Fot this purpose, the only technical detail involved would be the
following: how can I have fontinst believe that the font ptmr8r.mtx
has a raw glyph whose metrics are defined like this

\setglyph{dotlessj}
   \resetitalic{\italic{dotlessi}}
         \resetdepth{\depth{j}}
         \resetwidth{\width{j}}
         \resetitalic{\italic{dotlessi}}
         \resetheight{\height{dotlessi}}
\endsetglyph

if it's not in the AFM. would it be easy to add an \adddotlessj to the
\transformfont sugar? or is there a better way?