Fontinst modifications

[Ulrik Vieth]

> Ulrik Vieth writes:
>> After having a quick look at it, there's one point that puzzles me:
>> fontinst 1.511 talks about installing expertized fonts (9t/9e) 
>> as well as expertized oldstyle fonts (9o/9d), but in fontinst 1.6
>> the \if_oldstyle switch seems to have disappeared.  What happened?

> gurk. this is the scenario I had hoped did *not* arise... guess this
> means a detailed comparison of 1.511 and  1.6

There's not that much which has actually changed: Some fixes to the
AFM parsing code, the addition of the \if_fake_narrow switch and some
changes to the oldstyle stuff.

As far as I can see:

- both versions build expertized fonts using 9t/9e instead of 7t/8t,
  if you call \latinfamily{<fam>x}

- fontinst 1.511 builds expertized oldstyle fonts (using oldstyle 
  digits taken come from expert fonts) using 9o/9d instead of 9t/9e, 
  if you call \latinfamily{<fam>9} or \latinfamily{<fam>j}

- fontinst 1.6 builds expertized oldstyle fonts (using oldstyle 
  digits taken come from expert fonts), calling them 9t/9e as well, 
  if you call \latinfamily{<fam>9}, but not \latinfamily{<fam>j}

The questions are:

- why was the 9o/9d desingation or oldstyle expertized fonts 
  dropped and is it appropriate to use 9t/9e in that case?

- why isn't \latinfamily{<fam>j} no longer supported in 1.6, 
  contrary to what the LGC claims?

Personally, I do not particularly care about oldstyle digits,
especially if we can get them from TS1 (9c) if they exists.
I'm just interested to find out how it all came about.

Cheers, Ulrik.