On recursive ADT, provide indirection structured suggestion

[estebank]

No description provided.

[rust-highfive]

r? @matthewjasper

(rust_highfive has picked a reviewer for you, use r? to override)

[estebank]

This suggestion isn't entirely correct, but if used the compiler will correctly tell you to write

enum MList<'a> { Cons(isize, &'a MList), Nil }

so I think it is ok.

[lcnr]

These suggestions are now fairly long. While I don't have a strong opinion here, it may be better to only mention Box and & here.

[lcnr]

Or thinking more about it, I would prefer if we keep

help: insert indirection (e.g., a Box, Rc, or &) at some point to make Foo representable

and only add an example using Box.

[matthewjasper]

I'm also a bit concerned by the length of this suggestion.

[matthewjasper]

@bors r+

[bors]

📌 Commit c209040 has been approved by matthewjasper

[RalfJung]

Failed in https://github.com/rust-lang/rust/runs/744866295:

---- [ui] ui/issues/issue-72554.rs stdout ----
diff of stderr:

6	LL |     A(ElemDerived)
7	   |       ----------- recursive without indirection
8	   |
-	   = help: insert indirection (e.g., a `Box`, `Rc`, or `&`) at some point to make `ElemDerived` representable
+	help: insert some indirection (e.g., a `Box`, `Rc`, or `&`) to make `ElemDerived` representable
+	   |
+	LL |     A(Box<ElemDerived>)
+	   |       ^^^^           ^
10	
11	error: aborting due to previous error
12	

Probably needs a re-bless.
@bors r-

[estebank]

@bors r=matthewjasper rebased

[bors]

📌 Commit 03552ec has been approved by matthewjasper