Page name collision when they use a mix of unicode and ascii characters

[Stvad]

Current page naming logic normalizes page names to create readable URLs. Part of that normalization is removing unicode characters.

When it's fully comprized of unicode characters - the original transformation returns an empty string and there is an additional logic to handle that case. Unfortunately another case is not handled rn:

🙂1 and 🙀1 would resolve to the same page name - 1 (and special logic won't kick-in as it's not empty)

requirements:

• Map pages like this onto distinct URLs
• Minimize change in name-structure otherwise. Ideally avoid any changes not related to the affected pages. As changing the name structure generally would break already existing links to gardens.

---

Example garden: https://nicolejiang.roam.garden