jfkthame

I suppose when valueFormat2 is non-zero, we should record it as "unsafe" to break between the pair even if the values applied were zero?

Yeah, I'm pretty sure I remember seeing that sort of effect in a font. Whether it's OK to deviate here.... I'm not sure. I guess it'll be interesting to see...

One issue with the Unicode Script property is the characters that have Script=Inherited (generally diacritics) or Script=Common (mostly punctuation)... authors might be surprised at things that *don't* get included by...

This was just implemented in https://github.com/khaledhosny/ots/pull/247.

AFAIK, such fonts are not supported on (most?) other platforms, so a page that calls for (e.g.) MS Sans Serif will fall back to some other font -- either the...

> though [the lack of interop' here](https://jsfiddle.net/vyw9po7L/) is quite problematic, I admit Just FYI, the Firefox bug pointed out in your jsfiddle example (exposing an illegal "shorthand" that includes `all-small-caps`)...

> I understand that the Harfbuzz shaping of the sequence with ZWNJ between Ra and virama can mask an incorrectly encoded text string, whereas the Apple and Microsoft shaping makes...

I'm not sure I favor that. The hack in #435 was done to match deliberate Windows behavior (and an earlier Unicode recommendation), whereas AFAIK this case doesn't have that kind...

Isn't that what `` produces? (Sorry, I should've called it an explicit-halant form, not a half.) ISTM that `` should logically prevent the halant combining with the Ra, so we'd...

In the absence of more specific behavior, the default effect of ZWNJ is to request a less-ligated/joined rendering than would occur without it. Repha is the result of combining the...