Adrian Wong

@brendanzab I've attempted a fix for this. Due to parsing ambiguities arising from adding a second repeating pattern, I took the approach of separating the panic message args from the...

> I've been re-examining this, and I cannot see a difference between REPH_POS_AFTER_SUBJOINED and REPH_POS_BEFORE_POST, according to the Microsoft docs, other than the fact that at step (e) you have...

Apologies for not responding sooner, @n8willis! > | Reph Position | a | b | c | d | e | mH | > | ----------------------------- | - | -...

> the difference between (b) and (e) is that (b) checks for ZWJ/ZWNJ, which (e) does not do Could you expound on this? I'm struggling to see how (b) and...

> I'm certainly easily confused, so I defer to your tracing through here. But it looked to me like new_reph_pos would always be less than end. Ah! You may very...

Apologies for the delay here. Give me a week or two to re-familiarise myself with this; it's clear from Behdad's response on the other thread that there are some things...

> non-threatening ping If someone says that they'll get back to you in "a week or two" but hasn't responded in 1.5 years, I think it's well within your right...

> Was the example in https://github.com/n8willis/opentype-shaping-documents/issues/48#issuecomment-536838375 where it actually came up? Gosh. My memory is very hazy at this point, but it appears to be _somewhat_ related to both issues...

> It does seem like U+0DBA takes on post-base form via `post`, regardless of the base search. Are we sure that the difference seen in this test string is about...

> But explicitly saying "skip consonants that take on post-base form" would be in line with what we say in the Indic2 docs, and that is an easier solution than...