coverage: Make CoverageSuccessors a struct

[Zalathar]

A small and self-contained improvement, extracted from some larger things I'm working on.

There are two minor differences implied by this change:

• A SwitchInt terminator with exactly one target is now eligible to be chained.
  • The previous behaviour didn't have any strong rationale, and this gets rid of a special case.
• A terminator with no coverage-relevant successors is now always marked as “not out-summable”.
  • This is more correct, because (for example) a Return terminator has no successors (so their sum of execution counts is 0), but the return itself can have a non-zero execution count.
  • The existing code that checks this flag doesn't care either way, but this is likely to matter in the future.

There are no observable changes to the output of any of the coverage tests.

[rustbot]

r? @cjgillot

rustbot has assigned @cjgillot.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

[Zalathar]

Incorporated into #132389.