Emitting or deriving trait implementations

[fitzgen]

Let's implement all the traits!!

All of the "Derive Blah when possible" issues have code in tree already that they can crib from, so shouldn't be too difficult to write.

Once we start deriving lots of different traits, I expect we may find that we can unify their implementations, or at least make a generic CanDeriveTrait<T: DerivableTrait> kind of generic thing.

---

• "manually" implement Debug when we cannot derive it: Emit a "manual" implementation of Debug when it cannot be derived #875

• Derive Hash when possible: Derive Hash when we can #876

  • "manually" implement Hash when we cannot derive it: "manually" implement Hash when it cannot be derived #877

• Derive PartialEq when possible: Derive PartialEq when possible #878

  • "manually" implement PartialEq when we cannot derive it: "manually" implement PartialEq when it cannot be derived #879

  • Derive Eq when possible: Derive Eq when possible #880

    • "manually" implement Eq when we cannot derive it: "manually" implement Eq when we cannot derive it #881

• Derive PartialOrd when possible: Derive PartialOrd when possible #882

  • "manually" implement PartialOrd when we cannot derive it: "manually" implement PartialOrd when we cannot derive it #883

  • Derive Ord when possible: Derive Ord when possible #884

    • "manually" implement Ord when we cannot derive it: "manually" implement Ord when we cannot derive it #885

[dimbleby]

Should Copy be included in this meta-issue (as discussed in #934)? It would be a real odd-man-out if not.

[fitzgen]

Should Copy be included in this meta-issue (as discussed in #934)? It would be a real odd-man-out if not.

We already derive/impl Clone and Copy when possible, it's just that there's that caveat discussed in #934. This meta issue is for tracking traits we don't derive/impl yet.

[photoszzt]

@fitzgen derive Hash(#876) is implemented. But miss the mark there.

[fitzgen]

@fitzgen derive Hash(#876) is implemented. But miss the mark there.

Fixed, thanks!

[dimbleby]

We already derive/impl Clone and Copy when possible, it's just that there's that caveat discussed in #934. This meta issue is for tracking traits we don't derive/impl yet.

My point is that #934 says that WIBNI to derive Clone where possible and (only) otherwise manually implement it; and that this issue says pretty much the same thing for every other trait. In particular, bindgen often does not derive Clone when it would have been possible to do so.

But I don't suppose it really matters! Tracking that in one place is certainly enough; I was just trying to tie some loose ends together.