need some guru's help

Lars Hellström
Tue, 2 Jul 2002 20:19:26 +0200

At 10.16 +0200 2002-07-02, Sebastian Ahrens wrote:
>Hi Lars,
>Well after fiddling around for the last 6 days I am about to give up; even
>though I followed your tip and got me the original afm-files where the
>ligatures have their charposition set correctly in the end I'll never get
>them in the vpl's. Could you please have a look at the code-snipplets just
>to see whether I made a stupid mistake there?
>Thanks a lot for your patience and best regards
>                Sebastian
>------------------------ installpad.tex---------------------------

These \transformfont's may be causing you problems. The thing is that
\transformfont will call \mtxtomtx which expects to read one MTX file and
write another. In these cases the source and destination MTXs are the same,
so it may well be that \mtxtomtx wipes out the MTX created by \fromafm (or
actually by \afmtomtx) before it starts to read what is in it. Exactly what
happens probably depends on the OS, but I can't imagine it is any good.
Remove these \transformfont commands and the MTX and PL files they have
created. As long as you don't need to do any reencoding, \installfont will
see to that the MTX it needs is generated from the AFM (or PL).

>  \input csc2x

I'd use csckrn2x instead. padr8r and padr8x together contains all useful
glyphs found in padrc8r, hence there is no need to include any glyphs from
that font. padrc8a.afm does however contain useful metric information
(upper case--small caps kerns) which you do want to include.

>  \reglyphfont{padrcx8r}{padrc8r}

I'd say


instead. The file name padrc.mtx is as usual irrelevant, but it needs to be


The r:s in the Berry names tell you the weights, which in the NFSS scheme
becomes part of the series. Hence there is no need to include that in the
family name. The j is sometimes included in the family name because there
is no other place in the NFSS2 scheme where it is convenient to express
this property of the font. (There are suggestions that NFSS3 should have a
"case" axis which could handle this, but NFSS3 is currently about as
distant as LaTeX3.)


No need to poke around with unsetnum, kernoff, and kernon here. You do have
a proper set of expert fonts, so there are no headaches with distinct
glyphs having identical names to work around. Just do





>/* It doesn't matter where I place the padr8x in the lines above */

Probably because padr8x.mtx was empty. It should appear before latin, since
latin.mtx fakes the ligatures. Once a glyph has been set, any further
command that try to set it will be ignored.

Lars Hellström