[Testing] Speed up unit test suite: slow-first scheduling and long-pole parallelization

[janezpodhostnik]

Speeds up the full unit test suite (make unittest-main) from ~147s to ~84s median wall time on a 32-core machine, without weakening any assertion (statement coverage verified identical per package before/after).

• Makefile: schedule the slowest packages first when running the full suite (SLOW_TEST_PACKAGES, only applied for the default ./... package set; CI per-job package subsets are unaffected). go test schedules packages alphabetically, so the long, internally-sequential packages previously started at ~t+60s and formed a straggler tail that dominated the wall time (147s -> 97s).
• Parallelize long-pole packages whose tests ran fully serially (zero t.Parallel anywhere): scoring registry unit tests and the 5 subtests of TestScoreRegistry_SpamRecordWithoutDuplicateMessagesPenalty (5 sequential 5s require.Never windows now overlap), inspector validation ConfigToggle subtests, ledger compactor tests. All verified to have per-test fixtures and no shared state.
• connection gater integration test: run the two pubsub silence directions concurrently on two distinct topics (identical 5s never-receive windows, fully isolated), and overlap the pubsub and unicast protocol exchanges (package 65s -> 53s).
• Compress pacing intervals where assertions are per-heartbeat/per-tick, preserving semantics: ALSP heartbeat and peer manager update interval 1s -> 100ms (RepeatOffender test 21s -> 2s), parallelize the 6 verification happy path subtests (30s -> 8s).
• scoring libp2p integration tests run in parallel (they use random ports and noop metrics).

Deliberately kept serial: 4 scoring registry/decay tests whose score assertions tolerate only ~5-10% deviation, which the continuous spam-penalty decay exceeds under parallel contention (found via -race stress; parallelism there made them permanently fail, not just slower).

Validation: -race stress runs on all affected packages and repeated 10x full-suite campaigns (10/10 clean).

Depends on: #8640

---

^{Need help on this PR? Tag @codesmith-bot with what you need. Autofix is disabled.}

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 35e09b9d-5720-432c-a838-897dc53f7eab

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Dependency Review

✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.

Scanned Files

None

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!