Re: [Devel] Optimizatins to ttinterp.c

[David Turner]

Hi Nathan,

> I'm wondering whether it would be better to use a floating point
> version for most high end machines?  (once we have a glyph, the render
> would still be better in fixed point, but there is a lot of chance of
> bugs in complex code that performs scaling or sqrts using fixed point
> operations for things like glyph extent calculation.)
>
The problem is that TrueType bytecode decoding strongly relies on fixed
point arithmetic. More precisely, diagonal hinting is extremely fragile
to errors in dot products, and there are great chances that using
floating point would break some glyphs in unpleasant ways.

For example, I have already noticed some slight differences between
1.x and 2.0 because of the scaling difference explained in a different
post.

I have also played with sqaure root computation, and it _does_ indeed
change a lot of things.

We have no way than hoping to replicate the computational bugs that are
present in the original Windows and Mac hinters.

And of course, there is the problem of embedded systems. I'm currently
writing code that must run on an iPAQ, among other things, and floats
are simply not an option on this platform..

For example, running LibArt on this platform really, really sucks... :-)

Regards,

- David