feat: generate avatars for display name with chinese characters

[kesselb]

• Resolves: https://help.nextcloud.com/t/the-problem-of-automatically-generating-an-avatar-for-a-chinese-user-name/177750

Summary

We are using NotoSans-Regular for generating avatars.

If the character is not supported by NotoSans-Regular the users end up with an avatar like below:

The pull request here adds another font and uses a poor regex to flip out the font when chinese characters are detected. The font actually should cover even more characters.

@danxuliu did a deep dive into the topic a while ago and wrote down everything you need to know: #25529

TODO

• Testing Testing
• Do we want to ship another font
• Extend regex to match supported characters for NotoSC font

Checklist

• Code is properly formatted
• Sign-off message is added to all commits
• Tests (unit, integration, api and/or acceptance) are included
• Screenshots before/after for front-end changes
• Documentation (manuals or wiki) has been updated or is not required
• Backports requested where applicable (ex: critical bugfixes)

[AndyScherzinger]

@kesselb can you tell a bit about the state of the PR? 😊

[kesselb]

@AndyScherzinger Added some context to the pull request.

I never got feedback in the forums if the suggested patch works, and therefore did not continue.

This font topic is a rather complicated one, and I have no idea if that's a sane approach.

[Phreeman33]

Tested and worked fine.

As traditional Chinese users, we've been overriding fonts with Noto Sans TC to display Chinese characters properly. Here's the catch: Noto Sans SC uses some Simplified Chinese glyphs (probably GB18030 standard with "List of Commonly Used Standard Chinese Characters"), but the same character actually has different strokes in:

1. Traditional Chinese (Taiwan)
2. Traditional Chinese (Hong Kong/Macau)
3. Japanese
4. Korean

e.g.: for \u89d2
In TC:

In SC:

I recommend we just merge this commit. Since:

1. This PR already fixes avatar display for most cases in Chinese characters.
2. The glyph differences only affect a small number of characters.

Maybe we could handle the glyph differences with languange/location setting for CJK fonts matching later.

Trivia: Lots of websites/devices secretly use LCSCC glyphs for all Chinese text. The stroke differences are super subtle (mostly like 1-2 pen strokes!), which is why most traditional Chinese users never even notice! 😉

[kesselb]

Hi @Phreeman33,

Many thanks for your feedback—it's much appreciated! 🙏

Thanks for taking the initiative and sending a reworked version of the PR to improve the avatar situation for users with Chinese characters.

[skjnldsv]

Overrules by #51855