[Fontinst] Re: Bug in fontinstversion{1.927}

Peter Dyballa Peter_Dyballa at Web.DE
Mon Feb 7 13:39:54 CET 2005

Am 28.01.2005 um 16:50 schrieb Peter Dyballa:

> Am 28.01.2005 um 14:57 schrieb Ralf Stubner:
>> The Lucida Sans Oblique that came with Java (Sun's jvm 1.4.2) has an
>> error there:
>> ff  (U+FB00) is actually an fi
>> fi  (U+FB01) is actually an fl
>> fl  (U+FB02) is actually an ffi
>> ffi (U+FB03) is actually an ffl
>> ffl (U+FB04) is empty
>> Looks as if your Lucida Sans Oblique has a similar error. :-(

I could correct this quite easily last weekend (the *9z files provide 
dotlessj from t1dotlessj where needed):

%%%% This is the usual set:
   \transformfont{slsr8r}    {\reencodefont{8r} {\fromafm{slsr8a}}}
   \transformfont{slsro8r}   {\reencodefont{8r} {\fromafm{slsro8a}}}
   \transformfont{slsd8r}    {\reencodefont{8r} {\fromafm{slsd8a}}}
   \transformfont{slsdo8r}   {\reencodefont{8r} {\fromafm{slsdo8a}}}
% Special expert work on faulty Lucida Sans Oblique
   \killglyph{ffl}               % ffl at U+FB04 †
   \renameglyph{ffl}{ffi}        % ffi at U+FB03 -> ffl at U+FB04
   \renameglyph{ffi}{fl}         %  fl at U+FB02 -> ffi at U+FB03
   \renameglyph{fl}{fi}          %  fi at U+FB01 ->  fl at U+FB02
   \renameglyph{fi}{ff}          %  ff at U+FB00 ->  fi at U+FB01
   \installfont  {slsr7t}           {slsr8r,slsr8p,8p,newlatin}          
   {ot1} {OT1}{sls}{m}{n}{}
   \installfont  {slsrc7t}          {slsr8r,slsr8p,8p,newlatin}          
   \installfont  {slsd7t}           {slsd8r,slsd8p,8p,newlatin}          
   {ot1} {OT1}{sls}{b}{n}{}
   \installfont  {slsdc7t}          {slsd8r,slsd8p,8p,newlatin}          
   \installfont  {slsro7t}{slsro9r,slsro8p,slsro9z,8p,newlatin option 
   \installfont  {slsdo7t}{slsdo8r,slsdo8p,slsdo9z,8p,newlatin option 
  \installfamily{T1} {sls}{}
   \installfont  {slsr8t}           {slsr8r,slsr8p,8p,newlatin}          
    {t1}  {T1}{sls}{m}{n}{}
   \installfont  {slsrc8t}          {slsr8r,slsr8p,8p,newlatin}          
    {t1c} {T1}{sls}{m}{sc}{}
   \installfont  {slsd8t}           {slsd8r,slsd8p,8p,newlatin}          
    {t1}  {T1}{sls}{b}{n}{}
   \installfont  {slsdc8t}          {slsd8r,slsd8p,8p,newlatin}          
    {t1c} {T1}{sls}{b}{sc}{}
   \installfont  {slsro8t}{slsro9r,slsro8p,slsro9z,8p,newlatin option 
nosc} {t1}  {T1}{sls}{m}{sl}{}
   \installfont  {slsdo8t}{slsdo8r,slsdo8p,slsdo9z,8p,newlatin option 
nosc} {t1}  {T1}{sls}{b}{sl}{}
   \installfont  {slsr8c}   {slsr8r,slsr8p,8p,textcomp}                  
   \installfont  {slsd8c}   {slsd8r,slsd8p,8p,textcomp}                  
   \installfont  {slsro8c}{slsro8r,slsro8p,8p,textcomp}                  
   \installfont  {slsdo8c}{slsdo8r,slsdo8p,8p,textcomp}                  

Since it can happen with more fonts that they're fawlty and have to be 
'corrected' with \renameglyph I think of finally writing a Perl script 
that compares the *8p.afm files from ttf2pt1 with a reference and 
generates automatically the needed statements for the 8p.mtx file that 
'translate' deviating glyph names to those fontinst knows.

I do not use any more the 8p.etx file to reencode the MTX files from 
the already 8p encoded AFM files, so there is no need to teach fontinst 
any deviating new names in that early stage, only during \installfont 
fontinst needs to be instructed to accept a name like afii08941 as a 
synonym to lira. These instructions can be created with the Perl script 
as needed and can be restricted to the font family it is momentarily 
working on. So it shouldn't happen that names might clash (middot and 
periodcentered)? This accommodation is separated from any needed 
corrections -- a clean setup.



More information about the fontinst mailing list