I've been grinding away at the MikuBox core lately, and honestly? Sing-box is such a pain in the ass. The versioning is way too aggressive—1.10 to 1.12 to 1.13 to 1.14, just constantly deprecating fields. Every time I bump the core, I’m hit with a mountain of breaking changes; it’s like fixing one bug just to uncover three more.
It got me thinking, though: since I’m already deep into refactoring the backend, why not just jump ship to Mihomo (formerly Clash.Meta, the Clash lineage)? The Mihomo ecosystem is massive—proxy subscriptions, rule-providers, all the various GUIs and dashboards—most of them are built around the Clash spec. It’s got an overwhelming edge in terms of maturity and user base.
So, the plan is to drop support for sing-box. I’m going to keep the current framework but rewrite the whole thing from scratch for Mihomo. It’s going to be a hell of a lot of work, but I’m aiming for a much better user experience. Stay tuned!
It got me thinking, though: since I’m already deep into refactoring the backend, why not just jump ship to Mihomo (formerly Clash.Meta, the Clash lineage)? The Mihomo ecosystem is massive—proxy subscriptions, rule-providers, all the various GUIs and dashboards—most of them are built around the Clash spec. It’s got an overwhelming edge in terms of maturity and user base.
So, the plan is to drop support for sing-box. I’m going to keep the current framework but rewrite the whole thing from scratch for Mihomo. It’s going to be a hell of a lot of work, but I’m aiming for a much better user experience. Stay tuned!
- 👍14