[Testing] Restore exact EVM updated-register count assertion

[janezpodhostnik]

Follow-up to #8626, which relaxed assert.Len(state.UpdatedRegisterIDs(), 13) to
NotEmpty in the COA dryCall tests because the count varied between runs — losing the
ability to catch write-amplification regressions.

The variance is now fully characterized: the test env uses a random block ID, and
environment.uuidPartition (= sha256(blockID)[0] at txnIndex 0) selects the UUID
register. Partition 0 (probability 1/256) reuses the legacy uuid register → 12 updated
registers; all other partitions create a fresh uuid_N register → 13. This 1-in-256 case
is what made the original exact assertion flaky.

This PR restores an exact assertion via a small helper that computes the expected count
(12 or 13) from the block ID, keeping the fixture's full randomness. Both branches
verified: 15 random-block runs and 3 runs with partition 0 forced, all with -race.

Test-only change; no production code touched.

---

^{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: 849b6878-c62f-4e02-bece-39a8a0fe161f

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.}

[m-Peter]

LGTM!