Don't call FieldPlacement::count when count is too large

[pnkfelix]

Sidestep ICE in FieldPlacement::count by not calling it when count will not fit in host's usize.

(I briefly played with trying to fix this by changing FieldPlacement::count to return a u64. However, based on how FieldPlacement is used, it seems like this would be a largely pointless pursuit... I'm open to counter-arguments, however.)

Fix #57038

[rust-highfive]

r? @michaelwoerister

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

[Mark-Simulacrum]

Cc @eddyb

[pnkfelix]

Note that this PR doesn't actually make it so that our test suite runs cleanly for a host=i686 adn target=x86_64.

Namely, I observe this failure on huge-enum.rs now atop this PR.

failures:

---- [ui] ui/huge-enum.rs stdout ----
diff of stderr:

-       error: the type `std::option::Option<[u32; N]>` is too big for the current architecture
+       error: the type `[u32; 35184372088831]` is too big for the current architecture
2
3       error: aborting due to previous error
4

So ... is this an improvement over the status quo (and we just should revise the test suite a touch, maybe)? Or should we consider a broader change so that we get closer behavior in these various contexts...?

[michaelwoerister]

Let's merge it for now. It's definitely an improvement.

@bors r+ rollup

[bors]

📌 Commit 0eacf2c has been approved by michaelwoerister

[bors]

⌛ Testing commit 0eacf2c with merge 5edef531f2fdbfbdad7ea53fc4254d9cf747e7bd...

[bors]

💔 Test failed - status-appveyor

[emilyalbini]

@bors retry
AppVeyor... what's wrong with you today?