sagudev

@jdm why script crate still have build.rs and https://github.com/jdm/servo/blob/ce640c6730d1379dbd2bc3688b79875f274aaf50/components/script/dom/bindings/mod.rs#L188?

Crown would be problematic as currently this does not raise error: ```rust #[crown::unrooted_must_root_lint::must_root] struct Foo(i32); struct Bar(TH::F); //~^ ERROR: Type must be rooted, use #[crown::unrooted_must_root_lint::must_root] on the struct definition to...

I think we can only get MIR for bodies (function, closures); maybe the fastest option would be to make crown enforce something like this: ```rust #[crown::unrooted_must_root_lint::must_root] struct Foo(i32); struct Bar(TH::F);...

> > @jdm why script crate still have build.rs and https://github.com/jdm/servo/blob/ce640c6730d1379dbd2bc3688b79875f274aaf50/components/script/dom/bindings/mod.rs#L188? > > We still generate some code via script's build script when the actual concrete DOM types are available....

Actually specs clearly defines that for caching we use `_cached` slots: https://w3c.github.io/webcrypto/#concept-cached-object; and [cryptokey already contains such slots: ..., [[algorithm]], [[algorithm_cached]] ,...](https://w3c.github.io/webcrypto/#dfn-CryptoKey-slot-algorithm).

> As far as I know, the slots terminology is just way to represent a persistent value associated with a JS object. It doesn't imply a particular implementation strategy. Indeed,...

I always wanted to see engine-servo as mozilla android component: https://github.com/mozilla-mobile/firefox-android/tree/main/android-components/components/browser/engine-system, that would allow embeddeders to easily swap out engines (even in runtime I think).

Then we should probably set `MACOSX_DEPLOYMENT_TARGET` in mozjs-sys and remove `__builtin_available` check.