David Corbett

We should make this change. I checked all the Indic-shaper scripts in Notepad with strings analogous to ⟨ल्ँल⟩. Only Devanagari supports it, and only for certain marks: * U+0900 DEVANAGARI...

Currently, fallback composition happens if there is no 'mark' feature, and fallback mark positioning happens if there is no 'hebr' script in GPOS. It does not makes sense to do...

Yes. The first step is to decide the general approach: targeted exceptions for compatibility vs. broad changes for consistency. The next step is to find and document all the discrepancies...

There’s a similar problem at the top of U+0024 DOLLAR SIGN too.

For reference, Behdad’s first point’s issue was MicrosoftDocs/typography-issues#407 and the second’s are MicrosoftDocs/typography-issues#105 and MicrosoftDocs/typography-issues#184.

[Nishiki-teki](https://umihotaru.work/) has partial support.

It supports some sequences like ⟨a᪺᪽᪷᪽⟩. It is not full support but it is more than most fonts.

It can be any anchor, but I would use a middle anchor like `overlay`. Other anchors might cause unexpected problems for letters with unusual ascenders or descenders like U+046E CYRILLIC...

https://en.wikipedia.org/wiki/Cyrillic_numerals covers almost everything. There is also ⟨Ы꙲⟩ which is a special case. Sources are inconsistent about whether to use uppercase or lowercase letters.

[The Kannada script development spec](https://learn.microsoft.com/en-us/typography/script-development/kannada) allows one virama after the vowel signs: `{M}+[N]+[H]`. For example, here is ⟨ಕೕ್⟩ in Noto Sans Kannada in HarfBuzz:  It’s not the most obvious...