Re: [ft-devel] Experimental: v38 interpreter with minimal backwards compatibility mode and linear advance widths

[Nikolaus Waxweiler]

> Seconded.  Even if we somehow suppress horizontal hinting we won't get
> good results, I believe, since virtually all TrueType fonts try to
> align to full pixels, something the auto-hinter doesn't do in many
> cases.

So that settles it, unless someone knows of a reliable way to weaken 
full-pixel hinting on the Y-axis. I'm doubtful here, given the explicit 
nature of bytecode instructing and how a lot of hinting is probably 
done unsystematically and by eye.

I'm limiting the scope of my v38, it should be just the default variant 
of the v35 interpreter to be used in FT_LOAD_TARGET_NORMAL. The 
suppression of point movement on the x-axis should make it suitable for 
x-axis-emboldening though, hopefully also for "Native ClearType" fonts.

It is probably a much better idea to heave the CFF grid-fitter into its' 
own module and rework the autohinter to use it. I guess that would also 
take care of the long-standing implementation bugs in the autohinter 
fitting code and replace them with newer bugs? ;)

PS: I just found that DFKai-SB doesn't work with my v38 but with v35. 
ifelse statements incoming :)