Behdad Esfahbod
Behdad Esfahbod
For DWrite, also: https://github.com/microsoft/DWriteShapePy/blob/8b738ee03eccb4641f8de21a597bec17049774cf/src/cpp/DWriteShapeLib.cpp#L300-L314
Maybe @dscorbett can look into this? We already added a private script for Zawgyi. Is not too much work to do this.
> What’s the issue? You like the behavior of 'kana' but you would prefer to call it 'hrkt' instead? I don’t see the point. Since both scripts resolve to same...
We don't even have a shaper for those. So this sounds like just adding: ``` hb.hh:#define HB_SCRIPT_ KATAKANA_OR_HIRAGANA ((hb_script_t) HB_TAG ('H','r','k','t')) ``` and make sure that converts to OT script....
ISO 15924 has: Hrkt | 412 | Japanese syllabaries (alias for Hiragana + Katakana) -- | -- | --
Right. But `hb_script_t` declares itself as Unicode scripts. And we definitely don't want to diverge from that too much.
The pre/post-context don't care about clusters. But if you add main content in multiple calls, those need to have same base, yes, or you get wrong clusters. I thought a...
@khaledhosny do you know where we had that discussion? I can't find it.
> #1837? Yes. Thanks. So Matthias was involved in that one as well :)). Let's try to find a conclusion this time.
Right. The more I think about it, the more it makes sense that for vertical, it's relative to the vertical origin indeed; that is, the central baseline.