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

[Nikolaus Waxweiler]

> AFAIK, Microsoft maintains a whitelist of common fonts that must be
> handled specially, that is, where the default settings for backwards
> compatibility don't work correctly.  Essentially, the Infinality stuff
> does something similar.  If a font is not contained in the whitelist
> we simply have to trust that it does the right thing.

I wanted to see how far I can get before I have to resort to lists 
and/or detection trickery. Throwing most, if not all, x-movement out the 
window in b.c. mode made this a lot easier. And no x-movement means no 
need to change advance widths.

What I've done so far is turn on backwards compatibility as soon as my 
v38 is used and only turn it off if the font says the magic INSTCTRL 
word. Basically, if a font is developed with ClearType in mind but 
without going "native", with the patch it will display pretty much like 
on Windows. If it goes "native", it can do whatever it wants and will 
look even more Windowsy. Both cases apply for all web fonts I have come 
across so far. Even the Ubuntu font renders fine, and it looks like it 
was developed for v35.

The classically hinted fonts I tested (Arial, Times, Courier New, 
Georgia, Verdana, Palatino) render mostly okay to my surprise and 
present the only problem cases I have encountered. I want to tweak that 
some more once I get a working visual debugger going. Maybe I can lie to 
FF about the interpreter version. Who knows, it this works out and 
you're okay with the results, maybe we can get around lists and bytecode 
sniffing. That is my goal at least.