I've found another thing I'm unclear on. The MS docs say that Old Korean
text may use a sequence of individual jamo instead of the
combo-consonant/vowel jamo, so you'd encounter sequences like "L,L,V,T,T,T"
-- where the "L,L" and the "T,T,T" get turned into Old Hangul jamo
codepoints by the `ccmp` feature. (Then they all get shaped by the _jmo
features, of course)

How widespread is this? I haven't found a Korean font that actually does
this, although (as usual) I'm looking predominantly at FOSS fonts....

I'm just kind of wondering whether or not it's behavior worth documenting,
or is in the "Uniscribe oddities" or "MS documentation folklore"
categories. It seems like it would wreak havoc with syllable identification
as Harfbuzz does it, for sure.


HarfBuzz mailing list

Reply via email to