Fix RSpec fixtures for CID and heartbeat

[mackuba]

Motivation

• Fix failing specs caused by invalid CID fixture lengths used for JSON round-trips.
• Ensure Stream heartbeat specs simulate the internal state required to start/stop timers.
• Keep the test suite stable and runnable in CI by using realistic test doubles and fixtures.

Description

• Updated spec/cid_spec.rb to use 36-byte CID payloads when constructing and validating JSON-encoded CIDs via Skyfall::CID.
• Updated spec/label_spec.rb to use a valid Skyfall::CID string for label parsing assertions.
• Adjusted spec/stream_spec.rb to set @engines_on, @ws and @heartbeat_timer instance variables so check_heartbeat= toggles call start_heartbeat_timer/stop_heartbeat_timer as expected.

Testing

• Ran bundle install to install gems, which completed successfully and populated the test environment.
• Ran bundle exec rspec before the fixes which reported 60 examples with 4 failures due to CID length and heartbeat preconditions.
• Re-ran bundle exec rspec after applying the spec fixes and received 60 examples, 0 failures indicating the suite now passes.
• All automated RSpec examples executed successfully after the changes.

---

Codex Task

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 52a4f70b89

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Use full CID bytes in CBOR tag fixture

The CBOR tag fixture here uses a 32‑byte payload ("\x00" + ("a" * 32)), but CIDv1 bytes are 36 bytes (version + codec + multihash prefix + 32‑byte digest). The rest of the library treats JSON CIDs as 36‑byte data (length 59 in Skyfall::CID.from_json), so this test is encoding an invalid CBOR CID and would still pass if a regression dropped the 4‑byte prefix. That means real firehose/CAR inputs could break without any spec coverage. Consider building the tag with full CID bytes to match real data.

Useful? React with 👍 / 👎.