gitreins-tier1-go-test-stage-step-timeout-120s-default

Answer

Diagnosed and verified. Full write-up: ~/solution-gitreins-step-timeout.md.

Root cause

Two timeout sources feed two different step populations in engine/pipeline.py:

• Explicit pipeline.stages script steps use only the step's own key with a hard-coded default: python timeout=step_def.get("timeout", 120) # _run_script_step guards.test_timeout is never consulted.
• guards.test_timeout is read only in tier1_plan() (line ~1277), which injects test_step["timeout"] = test_timeout into the synthesized default Tier 1 steps. load_pipeline_config calls tier1_plan only when there is no pipeline: section, so an explicit pipeline skips that path entirely.

Bonus discrepancy: guard path defaults to 180s, pipeline fallback to 120s.

Fix

Add an explicit timeout: to each script step in .gitreins/config.yaml:

guards:
  test_timeout: 300
  test_command: "go test ./..."
pipeline:
  stages:
    - id: tier1
      parallel: true
      on: [pre-commit, pre-eval]
      steps:
        - id: secrets
          type: script
          run: "echo scan"
          timeout: 300
        - id: lint
          type: script
          run: "echo lint"
          timeout: 300
        - id: tests
          type: script
          run: "go test ./..."
          timeout: 300      # was missing -> hard 120s

Optional upstream hardening: make _run_script_step fall back to self.config["guards"]["test_timeout"] (default 180) instead of 120.

Verification (installed gitreins 0.15.0, stubbed run_bounded)

A) explicit step, no timeout key, guards.test_timeout=300 -> 120   # bug
B) explicit step, timeout: 300,           guards.test_timeout=300 -> 300
C) synthesized tier1 tests step timeout -> 300

End-to-end with a real config file: before fix the loaded tests step had no timeout key; after fix it loads "timeout": 300 and the effective runtime timeout is 300.