Closed hfhchan closed 7 years ago
hfhchan
commented
7 years ago The issue is caused in Regular/OTC/features.OTC.J
Under lookup jp2tw useExtension {
substitute \27526 by \27527;
which is unexpected.
Meanwhile, in UniSourceHanSerifTW-UTF32-H,
<000074e5> <000074e6> 27525
which is expected.
hfhchan
commented
7 years ago In Regular/OTC/features.OTC.SC
Under lookup cn2tw useExtension {
substitute \27527 by \27526;
which is expected.
In Regular/OTC/features.OTC.K
Under lookup kr2tw useExtension {
substitute \27526 by \27527;
which is unexpected.
kenlunde
commented
7 years ago Thank you for reporting this.
Whew! I was worried that my script for generating the lookups for the 'locl' GSUB feature had a bug. The bug is elsewhere, and the fix is to map a completely different code point to the correct glyph.
U+74E6 瓦 correctly maps to uni74E6-JP in the TW CMap resource, but U+2F61 ⽡ maps to uni74E6-CN, but should instead map to uni74E6-JP. I am closing this and consolidating with Issue #37.
For some reason, when specifying zh-TW and Source Han Serif (the Japanese language-specific version), 27527 is not substituted with 27526. However, when the language is set to zh-HK, there is no problem.
The error does not occur in Source Han Serif TW (the Taiwanese regional-specific subset version).
Demo: https://codepen.io/anon/pen/Emjaxq