test(infra): fix the 3 stale/flaky infra tests to the current contracts

- test_engine_image_and_rustfs_policy: root Dockerfile was refactored to the
  digest-pinned ousast-task-base (carries semantic+s3 deps) + a joern-build stage,
  installing the package --no-deps. Updated the stale `COPY benchmarks` (now
  --chown=1000) and `.[semantic,s3]` (now --no-deps + joern-cli copy + pinned base)
  assertions to the new contract.
- test_image_stages: plane/Dockerfile.engine-task refactored the same way; the
  split marker is now `FROM ${ENGINE_REF} AS replay`, base is ousast-task-base
  (not ax-task-runner:v0.3.1), and the replay stage no longer installs git (base
  provides it). Re-pinned the engine-stage SHA and updated the markers/assertions.
- test_partial_clone_lazy_fetch: passed alone but failed mid-suite — a leaked
  GIT_DIR/GIT_OBJECT_DIRECTORY/alternates from an earlier test changed the
  missing-blob count. Added an autouse fixture that neutralises inherited git
  control vars so every test in the file is self-contained (reproduced + verified).

Co-Authored-By: Claude Opus 4.8 <[email protected]>

tests/test_engine_queue.py

@@ -199,8 +199,12 @@ def test_engine_k8s_render_contract():
 
 def test_engine_image_and_rustfs_policy():
     dockerfile = (ROOT / "Dockerfile").read_text()
-    assert "COPY benchmarks ./benchmarks" in dockerfile
+    assert "COPY --chown=1000 benchmarks ./benchmarks" in dockerfile
-    assert ".[semantic,s3]" in dockerfile
+    # The image is built on the digest-pinned task base (which carries the semantic+s3 deps), so the package
+    # installs --no-deps and Joern (the semantic engine) is copied from the joern-build stage.
+    assert "ousast-task-base:${TASK_BASE_TAG}@${TASK_BASE_DIGEST}" in dockerfile
+    assert "/venv/bin/pip install --no-deps --no-cache-dir ." in dockerfile
+    assert "COPY --from=joern-build /opt/joern-cli /opt/joern-cli" in dockerfile
     for name in ["engine_trace_worker.py", "engine_trace_parse.py", "engine_worker_service.py"]:
         assert (ROOT / "benchmarks/learn" / name).stat().st_size > 0
     doc = (ROOT / "docs/rustfs.md").read_text().split("## Scoped account for engine workers")[1]

tests/test_unseen_extract.py

@@ -16,6 +16,25 @@ e = importlib.import_module("benchmarks.unseen.extract")
 s = importlib.import_module("benchmarks.unseen.source")
 
 
+@pytest.fixture(autouse=True)
+def _hermetic_git_env(monkeypatch):
+    """Neutralise inherited Git control vars so each test is self-contained.
+
+    A GIT_DIR / GIT_OBJECT_DIRECTORY / alternates value leaked from an earlier test in the suite redirects
+    object lookup: the partial-clone lazy-fetch test then sees a different missing-blob count (it passes alone
+    but failed mid-suite). These tests build their own repos under tmp_path and must not read ambient git env.
+    """
+    for var in (
+        "GIT_DIR",
+        "GIT_WORK_TREE",
+        "GIT_INDEX_FILE",
+        "GIT_OBJECT_DIRECTORY",
+        "GIT_ALTERNATE_OBJECT_DIRECTORIES",
+        "GIT_COMMON_DIR",
+    ):
+        monkeypatch.delenv(var, raising=False)
+
+
 @pytest.mark.parametrize("slow", [False, True])
 def test_process_deadline_restores_alarm(slow):
     original = signal.getsignal(signal.SIGALRM)

tests/test_unseen_image.py

@@ -4,13 +4,19 @@ import hashlib
 from pathlib import Path
 
 ROOT = Path(__file__).resolve().parents[1]
-ENGINE_SHA = "06ec67d411cd7abbfcd0735b1c2a70e4f3eb65ddcfeef89975709c664f59ad42"
+# The engine portion is pinned so that adding the replay stage cannot silently change the engine image.
+# Recompute after an intentional engine-stage change: it is sha256 of `engine.rstrip() + "\n"` where `engine`
+# is everything before the `FROM ${ENGINE_REF} AS replay` line.
+ENGINE_SHA = "77c34f3381cef6c6d1f7fd761cd3924a3656cc5e6eb20c51c819de77afb4b2fc"
 
 
 def test_image_stages():
     text = (ROOT / "plane/Dockerfile.engine-task").read_text()
-    engine, replay = text.split("FROM engine AS replay", 1)
+    # The replay stage builds FROM the engine stage (via the ENGINE_REF arg, default `engine`).
-    original = engine.replace("v0.3.1 AS engine", "v0.3.1").rstrip() + "\n"
+    engine, replay = text.split("FROM ${ENGINE_REF} AS replay", 1)
-    assert hashlib.sha256(original.encode()).hexdigest() == ENGINE_SHA
+    assert hashlib.sha256((engine.rstrip() + "\n").encode()).hexdigest() == ENGINE_SHA
-    assert "FROM ghcr.io/norandom/ax-task-runner:v0.3.1 AS engine" in engine
+    # Task base is digest-pinned (no floating fallback) and built in two stages: joern-build -> engine.
-    assert "git" in replay and "replay_entry.py" in replay and "ousast pre-push --help" in replay
+    assert "FROM ghcr.io/norandom/ousast-task-base:${TASK_BASE_TAG}@${TASK_BASE_DIGEST} AS joern-build" in engine
+    assert "FROM ghcr.io/norandom/ousast-task-base:${TASK_BASE_TAG}@${TASK_BASE_DIGEST} AS engine" in engine
+    # The replay stage is a pure addition on top of engine: the unseen replay harness + a smoke check.
+    assert "replay_entry.py" in replay and "ousast pre-push --help" in replay