AGENTIC QE V3.8.10 REPLACES FABRICATED COVERAGE WITH REAL PER-FILE METRICS AND TREND TRACKING
Agentic QE v3.8.10 fixes bogus coverage scoring and switches quality gates to real per-file metrics with trend tracking. The release [v3.8.10](https://github.c...
Agentic QE v3.8.10 fixes bogus coverage scoring and switches quality gates to real per-file metrics with trend tracking.
The release v3.8.10 removes fabricated coverage fallbacks that could show up to 95% on files with zero tests. It now uses actual coverage data or marks it as unavailable.
The test runner passes --coverage to Vitest/Jest and reads results from disk, with per-file coverage stored via a key-value API. A coverage:previous rotation enables simple trend detection release.
Expect more honest gates and dashboards. Teams without coverage tooling no longer get unfair penalties, but may need to align key names and paths after upgrading.
Restores trust in coverage metrics by removing fabricated values that inflated scores.
Prevents false penalties and mis-prioritized work when coverage data is missing or partial.
-
terminal
In CI, run agentic-qe test_execute_parallel with Vitest/Jest and verify per-file coverage keys exist and zero-test files are not inflated.
-
terminal
Temporarily remove coverage reports to confirm it shows "unavailable" and gates behave correctly; enable coverage:previous and check trend outputs.
Legacy codebase integration strategies...
- 01.
Map existing coverage artifact paths and keys to agentic-qe’s key-value schema to avoid mismatches after the upgrade.
- 02.
Re-baseline quality gates since fabricated scores may have masked gaps; expect possible initial score drops.
Fresh architecture paradigms...
- 01.
Start with per-file coverage storage and coverage:previous rotation enabled to establish clean baselines from day one.
- 02.
Use parallel test execution with coverage flags to collect metrics without extending CI time.