ptr_eq-first PartialEq implementation for Mrc/Irc?

[EndilWayfare]

I am working on a project that works with rather beefy data parsed from files. It's mostly heap allocated types -- Vecs and Strings -- so Mrc/Irc is a natural fit to pass that state down into child components. Avoiding unnecessary clones is fantastic, but profiling reveals a lot of time spent equality checking. The state from files is mostly immutable, changing only when a different file is loaded, so I pass around the app state that changes more often in a separate Irc. The unchanging data still goes through a full equality check very frequently, though.

I assumed that Mrc/Irc used the fact that they are pointers to optimize equality checks, using a ptr_eq to short circuit comparisons. Probably, I assumed this because yew.rs encourages this pattern in optimizations. This kind of "shallow equality" is leveraged heavily in React and the broader JS ecosystem for persistent(ish) structures. Neither smart pointer does this, however, and Irc doesn't even have a ptr_eq method.

I propose changing the PartialEq implementation (at least for Irc, since it's the one that is compared most frequently in Component::change) in addition to adding a ptr_eq method. This would play most nicely with NeqAssign.

The existing PartialEq implementation is purely a function of underlying data as it is: two pointers will compare equal if their data compares equal, even if it lives at two different addresses, so this proposal shouldn't change the semantics of existing programs.

If we agree this would be a good plan, I can draft a pull request.