Sarah Groß

> Ok. @Connum currently I do not have pending items, unless you share your failing test cases (font names, content tested, etc) with me according [to this](https://github.com/opentypejs/opentype.js/pull/557#issuecomment-1837164457) so I can...

I don't really like the idea of having both approaches in the code base in parallel. The arab and latn features should work with the new approach as well, so...

Shouldn't I be able to change `applyLatinLigatures()` to ``` function applyLatinLigatures() { applySubstitutions.call(this, 'latn', ['liga']); } ``` and it should just work? But the ligature tests then fail.

opentype.js currently only supports writing in CFF format, and even that doesn't support all features yet. I think it's very unlikely the library will ever implement TTF format in writing...

`getBoundingBox()` will only give you the bounding box of the path itself. Right now, you will have to add ascender and descender yourself for the height, and calculate glyph spacing...

Thanks for the report! So the issue at heart is that the only supported substitution type inside chained substitutions is '12' right now. I tried to understand the lookup-related code...

Here are two people who touched those parts of the code in recent years and might be able to shed more light on this and at least lead us currently...

If you remove the if/else statement checking for type 12, the dz and uz ligatures look like they are supposed to according to your image, but d and u then...

What confuses me even more is the fact that we're not using the backtrack and lookahead at all, apart from checking their length for the `contextRulesMatch` constant, or am I...

Yeah, that's what I thought. However, it then always uses the replacement forms, because we're not actually checking the backtrack and lookahead context. Going on the commit message, I guess...