[ft-devel] Enhancements for Subpixel Rendering

[Infinality]

Here are the latest updates I've been working on for the subpixel 
hinting stuff.  This removes a few hard-coded tweaks in favor of 
automatic detection.  In addition it attempts to follow the MS 
whitepaper more by restricting the conditions of the FDEFs to those 
specified instead of "everything".  The ttfautohint detection is 
probably no longer needed, as it should behave like a modern font, given 
that it doesn't contain signatures that trigger other opcode pattern 
recognition.

Some variables have been moved into the face object, because they really 
belong there, and not in the execution context.  Also, this patch should 
bring back the high performance (close to native) that it had prior to 
the last commit I did.  I attempted to commit this with David's latest 
patch but ran into some issues, so this doesn't include his C++ 
compatibility stuff.

Some of the syntax may need to be checked... it's probably not perfect 
but it's close.  Should apply cleanly to the latest git. Comments / 
suggestions / compliments welcome.  :D

Thanks,
Erik

subpixel-enhancements-20130123.patch
Description: Text Data