Make comptime functions ineligible for coverage

[Zalathar]

• Enforce that we don't generate code for comptime fns #159777
• Do not generate coverage records for compile-time-only functions #161808

---

As reported in #161808, the assertion in #159777 causes the compiler to ICE when it tries to generate an unused-function record for comptime functions, because comptime functions aren't allowed to have a symbol name. This can occur when trying to instrument core, for example.

Compile-time-only functions don't generate code, so instrumenting them for coverage is useless anyway. This PR therefore makes them ineligible for coverage, which avoids the problem.

---

This PR is essentially a rewrite of #161808. I wanted to request several modifications to that PR, and the overall change is small enough that it was easier for me to just recreate it from scratch.

The main differences are:

• Adds a coverage test instead of a ui test
• Comments focus more on the fact that instrumenting comptime functions isn't useful, with avoiding the ICE being a secondary motivation

cc @fmease @oli-obk as I'm not familiar with comptime

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

[rustbot]

r? @mati865

rustbot has assigned @mati865.
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

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler
• compiler expanded to 75 candidates
• Random selection from 19 candidates

[fs-rachel]

Looks reasonable to me, I'd like to ask for two things to be brought over from my version:

1. Please can you add Co-Authored-By: Rachel Barker <rachel.barker@ferrous-systems.com> to the commit message?

2. Please can you add the test first in its own commit, so that it's easy to check that the test fails before the main change is applied and succeeds afterward?

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[Zalathar]

1. Please can you add Co-Authored-By: Rachel Barker <rachel.barker@ferrous-systems.com> to the commit message?

Sounds reasonable, done.

2. Please can you add the test first in its own commit, so that it's easy to check that the test fails before the main change is applied and succeeds afterward?

Normally I wouldn't add a failing test in its own commit, specifically because I don't want to have a failing intermediate commit.

Since you asked, what I've done instead is temporarily add the test as a tests/crashes test that “successfully” crashes, and then migrate it over to tests/coverage in the commit that fixes the crash.

(If it were a tests/ui test, I could just add a test that expects to crash, and then change the test expectation from crash to pass. But coverage tests aren't allowed to crash the compiler.)

[fs-rachel]

1. Please can you add Co-Authored-By: Rachel Barker <rachel.barker@ferrous-systems.com> to the commit message?

Sounds reasonable, done.

Thanks :)

2. Please can you add the test first in its own commit, so that it's easy to check that the test fails before the main change is applied and succeeds afterward?

Normally I wouldn't add a failing test in its own commit, specifically because I don't want to have a failing intermediate commit.

Since you asked, what I've done instead is temporarily add the test as a tests/crashes test that “successfully” crashes, and then migrate it over to tests/coverage in the commit that fixes the crash.

(If it were a tests/ui test, I could just add a test that expects to crash, and then change the test expectation from crash to pass. But coverage tests aren't allowed to crash the compiler.)

Fair enough, I'm happy with that

[oli-obk]

I'm changing how comptime fns are handled to make all this a bit simpler, but for now, this is the correct check and mirrors the other checks in the compiler

r? @oli-obk

@bors r+ rollup

View changes since this review

[rust-bors]

📌 Commit 0027ee1 has been approved by oli-obk

It is now in the queue for this repository.

[JonathanBrouwer]

@bors try jobs=test-armhf-gnu

[rust-bors]

☀️ Try build successful (CI)
Build commit: fff678a (fff678a10a9c4a1f5f465322647dd2f3af976448)
Base parent: 656a9da (656a9da186dacaf3bf8f7f7296a825d256cb4ae3)