Make function pointers implement traits for up to 12 parameters

[Manishearth]

(12 was chosen to be consistent with what we do for tuples)

Fixes #28559

[rust-highfive]

r? @aturon

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

[nrc]

@bors: r+

[bors]

📌 Commit 5f66c70 has been approved by nrc

[bors]

⌛ Testing commit 5f66c70 with merge 6217b00...

[bors]

☀️ Test successful - auto-linux-32-nopt-t, auto-linux-32-opt, auto-linux-64-nopt-t, auto-linux-64-opt, auto-linux-64-x-android-t, auto-mac-32-opt, auto-mac-64-nopt-t, auto-mac-64-opt, auto-win-gnu-32-nopt-t, auto-win-gnu-32-opt, auto-win-gnu-64-nopt-t, auto-win-gnu-64-opt, auto-win-msvc-32-opt, auto-win-msvc-64-opt

[petrochenkov]

@Manishearth
Implementations for function pointers were "trimmed" a bit in #28268 to reduce some pretty nasty code bloat. 8 arguments would be enough for backward compatibility with stable.
Well, yes, it required a crater run~~, but who on earth uses functions with 12 arguments?!~~

[Manishearth]

Functions with a large number of arguments being used in other structs are basically inevitable in bindings.

[alexcrichton]

It's always good to do a bit of investigation into what caused regressions if they do happen, and as @petrochenkov mentioned we did indeed cut these back to reduce the massive metadata size of libcore. The number chosen for tuples was completely arbitrary, so there's not real reason to match it for functions, and I would prefer to have only the bare minimum necessary to get things working until we find a better solution.

This is ok because it already landed, but always nice to not be too hasty!

[Manishearth]

Will do, sorry about the trouble.

[gkoz]

Is there an issue documenting the bloat problem?

[alexcrichton]

triage: beta-nominated

Nominating for a backport as this fixes a reported regression with chipmunk-sys and is pretty harmless.

[alexcrichton]

The libs team decided to accept this for a backport to beta