rounding problems

[Thierry Bouche]

» For anyone who's wondering why the code does this in the first place, it's
» because otherwise you get overflow errors by multiplying a dimen by a
» scalefactor > 1000 (TeX doesn't support fp multiplication, you have to
» multiply by an int then divide by another int).

That's one of TeX's features that make me sometimes wonder whether TeX was
really the language of choice for a software like fontinst...

Thierry Bouche, Grenoble.