[HarfBuzz] Old Korean T jamo that don't conjoin?

[Richard Wordingham]

On Fri, 30 Mar 2018 18:15:13 -0500
Nathan Willis <nwillis at glyphography.com> wrote:

> If the rule was that only <Modern,Modern,Modern> would compose into
> something in the Syllables block, that would make total sense to me.
> But the comment makes it sound like <Old,Old,Modern> also maps to
> stuff in the Syllables block and that it's only the Old Ts that are
> excluded. That's what I don't understand.

The comment is explaining which <LV,T> combinations compose at the
character level.  I think you will find that LV combinations only exist
for <Modern,Modern> combinations.  It is at that level that
<Old,Old,Modern> is excluded.

What you seem to be after is a change to the text on line 1154 from

"Only the <L,V> sequences for the 11xx ranges combine."

to

"Only the <L,V> sequences for parts of the 11xx ranges combine."

or

"Only the <L,V> sequences for some L and some V combine."

The interesting bit of Unicode history is why the consortium caved in
to Korean demands for the full set of <modern, modern, modern> to be
included.  An uninteresting bit of HarfBuzz archaeology would be
whether someone briefly thought that the all L and V in the Hangul Jamo
block composed at the character level.

Richard.