ci: Apply a consistent Cargo profile

[abdnh]

Linked issue

Related: #5102 (comment)

Summary / motivation

This adds a new Cargo profile (ci) for use in all Rust build commands on CI. The goal is to reduce unnecessary recompilation of the same crates in dev/release profiles.

Steps to reproduce (before)

View the logs of the last CI run on main and notice that some crates are getting compiled with the release profile, e.g. compilation ends with "Finished release profile [optimized]".

How to test (after)

View the logs of the last CI run in this PR and confirm all Rust compilation commands end with "Finished ci profile [unoptimized]", indicating that only a single profile is being used.

Checklist (minimum)

• I ran ./ninja check or an equivalent relevant check locally.
• I added or updated tests when the change is non-trivial or behavior changed.

Scope

• This PR is focused on one change (no unrelated edits).

[abdnh]

We set RELEASE=2 in the release workflow so this has no effect there.

[iamllama]

Changes LGTM, but i'm seeing the opposite on ubuntu: ~68s for ./check on main vs ~96s for CI=true ./check on the pr branch 🤔

[fernandolins]

Yeah, I agree with @iamllama. The changes LGTM, but I cannot confirm that they demonstrably reduce time. Is there any way you can demonstrate an improvement with these changes?

[read-the-docs-community]

Documentation build overview

📚 Anki | 🛠️ Build #34026890 | 📁 Comparing db85a77 against latest (48c8e10)

  🔍 Preview build  

48 files changed · ± 36 modified · - 12 deleted

± Modified

• api-aqt-modules.html
• api-python-modules.html
• api-python.html
• architecture.html
• build.html
• contributing.html
• development.html
• editing.html
• genindex.html
• language_bridge.html
• and 26 more...

- Deleted

• autoapi/anki/exporting/index.html
• autoapi/anki/importing/index.html
• autoapi/aqt/exporting/index.html
• autoapi/aqt/importing/index.html
• autoapi/anki/importing/anki2/index.html
• autoapi/anki/importing/apkg/index.html
• autoapi/anki/importing/base/index.html
• autoapi/anki/importing/csvfile/index.html
• autoapi/anki/importing/mnemo/index.html
• autoapi/anki/importing/noteimp/index.html
• and 2 more...

[abdnh]

@iamllama

but i'm seeing the opposite on ubuntu: ~68s for ./check on main vs ~96s for CI=true ./check on the pr branch 🤔

Do you see a meaningful difference if you set opt-level = 1 for the ci profile? Appears to be slower on CI:

• opt-level = 0: https://github.com/ankitects/anki/actions/runs/30927538265
• opt-level = 1: https://github.com/ankitects/anki/actions/runs/30929285251

[abdnh]

I don't see consistent differences between this PR and main however (https://github.com/ankitects/anki/actions/runs/30923778268) and I'm not sure that's possible to confirm considering how unreliable GitHub runners are.

Applying a single compilation profile though appears to work for the purpose of reducing recompilation. We could set opt-level = 1 later if it turns out tests are getting slower.

[iamllama]

Do you see a meaningful difference if you set opt-level = 1 for the ci profile? Appears to be slower on CI:

It's definitely slower for me as well when run locally

[Luc-Mcgrady]

Does incremental compilation not get used in subsequent tests? I thought it would because of actions/cache/restore but I haven't seen it practice.

[abdnh]

Does incremental compilation not get used in subsequent tests?

It looks like it's common to disable it on CI: dtolnay/rust-toolchain#26

I thought it would because of actions/cache/restore but I haven't seen it practice.

Yeah I think something is not working here - the last run recompiled some stuff despite the commit only touching ci.yml. Maybe we should use the Swatinem/rust-cache action: #5134 (comment)

I'll revert the change for now until that's figured out.

[Luc-Mcgrady]

LGTM