Behdad Esfahbod
Behdad Esfahbod
> $ hb-shape XBRoya34.ttf " الأ" --direction=r --script=arab --language=ar --show-flags -V > [uniFEF7=2+1235|afii57415=1+460|space=0+600] With the patch: ``` $ hb-shape XBRoya34.ttf " الأ" --direction=r --script=arab --language=ar --show-flags [uniFEF7=2+1235#1|afii57415=1+460|space=0+600] ```
I don't think there's anything actionable here. Feel free to disagree.
FWIW defcon master if almost as fast as robofab. It's the ufo3 branch of defcon that is twice slower.
Can we merge ufo3 branch first, then start optimization work?
For me I just need to know which branch to spend my time on :). A local benchmark, as well as a strong test suite help. Not much else needed...
It all will be okay! Let's not go for a 5-year UFO-4 then. Gradual improvements FTW...
Thanks Ben. I see Tal's comment about this being intentional. If, say, group1 has a particular glyph A in a group, but other side doesn't have A in any group,...
cc @letterror @anthrotype @typemytype
> If a pair is not available in one of the sets, it assumes 0, whereas it should instead use the effective kerning value for that pair, which might come...
Ok, my bad. The first issue does not exist. I didn't note that the code is using "self.get()" which properly does fallback. So, the only issue outstanding is the fact...