On Fri, 11 Mar 2022 17:53:12 GMT, Phil Race <p...@openjdk.org> wrote:
> Yes, the sequence.fallback but I am relaying my observation that adding korean.allfonts=.. did not work even though sequence.fallback lists korean whereas adding korean-foo.allfonts=.. and adding korean-foo into sequence.fallback DID work. It's just there are (only) two ways to define Korean (or any other script) font to be used for fallback - either use only one of them for all 20 logical Java fonts (like e.g. it's currently done for Thai language), or configure a different Korean font for each of those 20 fonts separately (like it's currently done in OpenJDK for Korean language). It's not possible to 'mix' those two approaches - by using 'allfonts' entry as a kind of 'fallback' for the case when a specific font doesn't exist. When both ways (forms) are provided in the configuration file, current code picks the 'second' form, ignoring `allfonts` entry. That's why adding `allfonts.korean` in the patch didn't have any effect. But there are no existing entries like `serif.plain.korean-foo`, so adding `allfonts.korean-foo` does have an effect. ------------- PR: https://git.openjdk.java.net/jdk/pull/7643
