coverage: Rename the three main coverage-info structs

[Zalathar]

This PR renames the three main structures used to carry per-function coverage info through different stages of compilation:

- CoverageInfoHi       => CoverageEarlyInfo   // collected early, at the THIR/MIR boundary during MIR building
- FunctionCoverageInfo => CoverageMirInfo     // collected during the main MIR instrumentation pass
- CoverageIdsInfo      => CoverageCodegenInfo // collected during codegen, after MIR opts

The new names should hopefully help to keep the three structs distinct, while also avoiding the historical inconsistency of the previous names.

The renaming also resolves an old TODO of mine about coverage_ids_info no longer making sense as a name.

There should be no change to compiler behaviour.

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

This PR changes MIR

cc @oli-obk, @RalfJung, @JakobDegen, @vakaras

[rustbot]

r? @mejrs

rustbot has assigned @mejrs.
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, mir
• compiler, mir expanded to 75 candidates
• Random selection from 18 candidates

[mejrs]

I'm not a fan of putting "early" in names of things. In my experience it tends to be somewhat vague and it ages poorly. How about CoverageHirInfo?

[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]

I'm not a fan of putting "early" in names of things. In my experience it tends to be somewhat vague and it ages poorly. How about CoverageHirInfo?

Hmm, I see the general concern, but in this instance I think it's doing good work, and there's less risk of it going stale than if we tried to use a less-vague name.

The main distinction I'm trying to make is that of the two side-data structs attached to a MIR body, this is the one that has to be prepared at a time before the main InstrumentCoverage pass, because it preserves information that is only available during MIR building.

The struct itself contains whatever currently needs to be collected early and stored as side-data. It's less “important” than the side-data recorded during instrumentation, and ceases to have value after instrumentation. Any my hope is that follow-up work to #161517 will allow it to be removed entirely.

---

The specific suggestion of CoverageHirInfo doesn't work, partly because it clashes with ExtractedHirInfo which is extracted later (during instrumentation), and partly because the “HIR” part would be more misleading than accurate.

I could imagine it being named something verbose like CoverageEarlyMirBuildInfo, but going down that road seems to just make things clunkier without making them clearer. The actual role of the struct can't be fully captured in its name, so I'd prefer a distinctive name that presents the key distinction (early collection) and otherwise gets out of the way.

[mejrs]

Sounds good.

@bors r+ rollup

View changes since this review

[rust-bors]

📌 Commit 0bce041 has been approved by mejrs

It is now in the queue for this repository.