core/qbft: preserve local value on comparison failure

[KaloyanTanev]

Fixes the compare flow losing the already-fetched local value when the comparator reads the value and errors back-to-back: both buffered channel sends can complete before the select polls, which then picks the error case at random and drops the value. Subsequent rounds then block forever on the already-consumed input value source channel, so the node never prepares again, as seen in TestChainSplit flakes reaching max round. Extracts the select loop into awaitCompare and drains the pending value on the error and timeout paths.

category: bug
ticket: none

[Copilot]

Pull request overview

Fixes a QBFT compare-loop race where a locally-fetched value could be dropped if the comparator sends both a value and an error on buffered channels before the select loop observes them, which could later deadlock subsequent rounds waiting on an already-consumed input source.

Changes:

• Extracts the comparator select loop into a new helper (awaitCompare) and ensures any pending local value is drained/retained on error and timeout paths.
• Updates the timeout path to return the latest available local value (if already produced) instead of potentially returning a stale/empty value.
• Adds a regression test that repeatedly exercises the “both channels ready” select-randomness scenario.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
core/qbft/qbft.go	Refactors compare waiting logic into awaitCompare and drains a pending compare value on error/timeout to avoid losing the consumed local value.
core/qbft/qbft_internal_test.go	Adds a regression test covering the racy “value+error both ready” case to prevent deadlocks in subsequent rounds.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[pinebit]

Love generics

[KaloyanTanev]

Controversial topic in golang... haha

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 58.15%. Comparing base (3e85e42) to head (ef34194).
⚠️ Report is 5 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4628      +/-   ##
==========================================
+ Coverage   57.91%   58.15%   +0.23%     
==========================================
  Files         246      247       +1     
  Lines       33937    34043     +106     
==========================================
+ Hits        19656    19797     +141     
+ Misses      11821    11776      -45     
- Partials     2460     2470      +10     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud