lsm / HyperNeo / 32425691780

80%

push

github

test(daemon): run validate-test-matrix guard against a fixture mini-repo (#2628)

* test(daemon): run validate-test-matrix guard against a fixture mini-repo

Each of the 39 guard scenarios now mutates its own clone of a temp
fixture repo (verbatim copies of the workflows/scripts/configs plus a
synthesized test-file skeleton and stub vitest/vite modules) instead of
the real repository, and all guard runs execute concurrently. Local file
time drops from 533s to ~29s; the real-repo green check is kept as one
scenario, and a pristine-fixture green scenario guards against fixture
drift.

* test(daemon): cap guard-worker concurrency at the effective CPU allowance

Review feedback (PR #2628): os.cpus().length ignores cgroup CPU quotas,
so a throttled container over-subscribed workers and starved guards into
the execFile timeout. The worker limit now respects cgroup v1/v2 quotas
(capped at 8) and the per-run/test timeouts have headroom (150s/180s).

44344 of 59754 branches covered (74.21%)

Branch coverage included in aggregate %.

60266 of 72031 relevant lines covered (83.67%)

3975.59 hits per line