Open simoncozens opened 10 months ago
@simoncozens putting 'mplusrounded' in GitHub search shows that M PLUS Rounded 1c lives under ofl/roundedmplus1c. ~Maybe that path should be changed?~
Nope, that was silly of me. The 'Rounded M+ Project' site linked in the description seems to imply that it's named 'Rounded M PLUS' and not 'M PLUS Rounded'.
Some of our Japanese fonts display kana-only (which might be OK because they are kana only fonts, but also might indicate a problem with kanji coverage). Some of them display Latin sample text. I'm proposing to restrict the
ja_Jpan
exemplar kanji to Joyo kanji only, but I'm going to use this issue to document whether or not that would actually fix the problem.subset
filter is set to Japanese. This is wrong, and I'm not sure why this is the case especially given that Noto Serif Japanese does not. METADATA.pb languages and primary script are the same in both fonts.subset=japanese
is not in the URL and a Japanese specimens if it is. They probably need a primary-script metadata field. Most of these fonts support all of Joyo kanji except the three Joyo kanji: 剝, 塡, 頰. These kanji are kind of outliers anyway. They are 2010 additions to the Joyo list, they are not part of the JIS character set, and are usually replaced by 填, 剥, and 頬, but they're in there now. Maybe we want to take these out of the exemplars.git grep -i "M PLUS Rounded 1c"
produces no results at all. Weird, weird, weird.