Behdad Esfahbod
Behdad Esfahbod
Subsetting `PaintGlyphDelta` is tricky. To be determined.
I agree with you.
@qxliu76 for `*=drop`. In the mean time, if partial-instancing is requested we should just err instead of producing broken fonts.
This is a good start. But I think you need to tell the shaper which mode it's operating under. If a font has morx+GPOS, right now the `ot` shaper still...
Also, I think both those shapers should NOT fail if their layout tables are missing. We just need to communicate to the main shaper what layout tables to prefer /...
Maybe lets sketch out what exactly the two should do first.
> @behdad does any of what I wrote above make sense? I've been travelling. Will respond soon.
Okay the main thing I'm uncomfortable with is suddenly making "ot" shaper fail at all. This wasn't the case before. I'm fine if it starts ignoring "aat" tables, but failing...
The way I see it is rather: - "ot" would never apply morx/kerx, - "aat" would never apply GSUB, - "hb" is as is today. shaper order will have "hb"...
My reasoning is that "no layout table" is valid OpenType / AAT shaping. So we shouldn't fail.