RFR: 8190907: Windows 10 default Korean Font Malgun Gothic available not used by GUI elements even though available without Korean Language Pack [v3]

[Phil Race]

On Sat, 5 Mar 2022 10:35:12 GMT, Dmitry Batrak <dbatrak at openjdk.org> wrote:

>> But serif.plain.korean isn't used unless it is a Korean locale .. so what's the point ?
>> 
>> And this one line already added 
>> allfonts.korean=Malgun Gothic
>> will provide it as a fall back so I don't see the need.
>> If I am an English language user running in a US / UK locale I'll get my Korean glyphs from there with the only difference being it isn't in the same place in the precedence order.
>> This also ensures that any "new" code points added in that font not present in the older ones will always be available.
>
>> But serif.plain.korean isn't used unless it is a Korean locale
> 
> That's not the case, and is easily verifiable for English locale - whatever font is specified there will become a component of Serif composite. 'korean' here is a script id, which is referenced from `sequence.fallback`, used for all locales.
> 
>> And this one line already added
> allfonts.korean=Malgun Gothic
> will provide it as a fall back so I don't see the need.
> 
> That's also not the case. With the patch applied, Serif font still doesn't have Malgun Gothic as a component.
> That line in the current patch has no effect at all - `allfonts.<script>` lines only have an effect for scripts, which don't have assigned fonts in `<composite>.<style>.<script>` properties. 
> 
>> BTW does anyone know if this is a problem for Chinese or Japanese too ?
> I think Indic might have needs a font pack too .. maybe not in this bug but should we be looking wider than Korean ?
> 
> Didn't hear about related problems for Chinese or Japanese, but the same problem exists for Thai script - DokChampa font isn't available by default on Windows 10, and we use Tahoma as a replacement currently.
> 
> BTW, maybe it's worth to create a test case for this problem? As Korean support is explicitly coded in `windows.fontconfig.properties`, it might be good to check that all 20 logical fonts can display Korean on Windows. If a newer Windows version will change the bundled font again, the problem will become apparent at once.

And @JB-Dmitry what did you actually test ? The current PR state ? 
I assume that @prsadhuk tested that before updating the PR.
But the current PR state looks like what you say doesn't work so I am getting confused.

-------------

PR: https://git.openjdk.java.net/jdk/pull/7643