Jeremias Maerki wrote:
> > Most people will want to point the auto-registration to
> > Mine has 581 files in it, each of which would have to be opened and
> > partially parsed. I actually use no more than 10-15 for FO work. I
> > think this might actually open up a bunch of potentially
> ugly support
> > issues, and I would think twice about it. Nevertheless, it
> can be done.
> Not necessarily. I thought about that. We would have to
> create some sort of font cache file which holds the most
> important info to avoid parsing every font before it's really
> used. I've seen that even the old Acrobat Reader 4.05 on
> Linux did something like that.
It is an interesting idea. I would probably tend to implement it, at least
in the beginning, as a batch job that simply reads a directory of font files
and spits out a font-configuration file for the contents of that directory.
That saves the user some trouble, but still gives him control (and
responsibility) over the output. If your application tries to take
responsibility for this, then you will end up needing to reconcile the cache
with reality every time anyway, which defeats the purpose of the cache.
There is a good reason that fonts tend to evolve into an o/s service.