Decompose oversized Rust files by responsibility

tools/lazaret_verify.py

@@ -14,6 +14,7 @@ import tempfile
 import tomllib
 
 ROOT = Path(__file__).resolve().parent.parent
+VERIFICATION_TEMP_ROOT = ROOT / "target" / "verification"
 MATERIALIZED_FUSER_MANIFEST = (
     ROOT / ".lazaret/vendor/fuser-0.17.0/Cargo.toml"
 ).resolve()
@@ -165,14 +166,13 @@ PROCESS_LINT_ALLOWANCES = {
     "crates/lazaret-linux/src/process.rs": {
         "append_to",
         "send_signal",
+    },
+    "crates/lazaret-linux/src/process/broker.rs": {
         "acquire",
         "acquire_with_stdio",
         "acquire_with_stdio_after_ack",
         "acquire_captured",
         "private_role_command",
-        "require_reapable_sigchld",
-        "cleanup_sigchld_probe",
-        "fail_sigchld_probe",
         "reap",
         "reap_status",
         "fail_running_captured",
@@ -180,16 +180,23 @@ PROCESS_LINT_ALLOWANCES = {
         "verify_direct_child_pidfd",
         "fail_before_pidfd",
         "fail_after_pidfd_closed",
-        "recover_process",
-        "current_process_identity",
-        "append_parent_identity",
-        "exec_target",
         "run_transient",
         "run_transient_with_stdout",
         "run_transient_captured",
         "spawn_bootstrap",
         "captured_process",
+    },
+    "crates/lazaret-linux/src/process/recovery.rs": {
+        "recover_process",
+        "current_process_identity",
         "match_recovery_process",
+        "require_reapable_sigchld",
+        "cleanup_sigchld_probe",
+        "fail_sigchld_probe",
+    },
+    "crates/lazaret-linux/src/process/launch.rs": {
+        "append_parent_identity",
+        "exec_target",
         "suspend_current_process",
     },
     "crates/lazaret-host-coordinator/tests/runtime-composition.rs": {"run_benchmark_guest"},
@@ -218,6 +225,12 @@ PROCESS_LINT_ALLOWANCES = {
         "retained_fixed_child_bind_mount_is_refused_when_mount_setup_is_available"
     },
 }
+RUST_FILE_LINE_LIMIT = 1000
+RUST_FILE_SIZE_EXCEPTIONS = {
+    "crates/lazaret-host-coordinator/src/lifecycle.rs": (
+        "host coordinator lifecycle transition state machine and invariant-preserving adapter"
+    ),
+}
 FORBIDDEN_PROCESS_SYSCALLS = tuple(
     re.compile(pattern)
     for pattern in (
@@ -250,6 +263,11 @@ class VerificationError(RuntimeError):
     pass
 
 
+def temporary_directory(prefix: str) -> tempfile.TemporaryDirectory[str]:
+    VERIFICATION_TEMP_ROOT.mkdir(parents=True, exist_ok=True)
+    return tempfile.TemporaryDirectory(prefix=prefix, dir=VERIFICATION_TEMP_ROOT)
+
+
 def require(condition: bool, message: str) -> None:
     if not condition:
         raise VerificationError(message)
@@ -402,6 +420,43 @@ def workspace_rust_sources(metadata: dict) -> list[Path]:
     return sorted(sources)
 
 
+def audit_rust_file_sizes(metadata: dict) -> None:
+    sources = workspace_rust_sources(metadata)
+    relative_sources = {str(path.relative_to(ROOT)): path for path in sources}
+    line_counts = {
+        relative: len(path.read_bytes().splitlines())
+        for relative, path in relative_sources.items()
+    }
+    for relative, responsibility in RUST_FILE_SIZE_EXCEPTIONS.items():
+        require(
+            relative == Path(relative).as_posix()
+            and not Path(relative).is_absolute()
+            and relative.endswith(".rs"),
+            f"malformed Rust file size exception path: {relative}",
+        )
+        require(
+            "\n" not in responsibility and len(responsibility.split()) >= 3,
+            f"malformed Rust file size exception responsibility: {relative}",
+        )
+        require(relative in relative_sources, f"stale Rust file size exception: {relative}")
+        require(
+            line_counts[relative] > RUST_FILE_LINE_LIMIT,
+            f"stale Rust file size exception below threshold: {relative}",
+        )
+
+    oversized = {
+        relative: lines
+        for relative, lines in line_counts.items()
+        if lines > RUST_FILE_LINE_LIMIT and relative not in RUST_FILE_SIZE_EXCEPTIONS
+    }
+    require(
+        not oversized,
+        "Rust files exceed the reviewed responsibility threshold: "
+        + ", ".join(f"{path} ({lines})" for path, lines in sorted(oversized.items())),
+    )
+    print("PASS tracked Rust file responsibility size boundary")
+
+
 def audit_dependencies(metadata: dict) -> None:
     policy = load_toml(ROOT / "deny.toml")
     allowed_licenses = set(policy["licenses"]["allow"])
@@ -466,7 +521,7 @@ def license_permitted(expression: str, allowed: set[str]) -> bool:
 def run_models() -> None:
     tlc = shutil.which("tlc")
     require(tlc is not None, "TLC must be installed on PATH")
-    with tempfile.TemporaryDirectory(prefix="lazaret-tlc-") as directory:
+    with temporary_directory("tlc-") as directory:
         workdir = Path(directory)
         for module, config in MODEL_CONFIGURATIONS:
             require(module.is_file(), f"missing model: {module.relative_to(ROOT)}")
@@ -561,7 +616,7 @@ def audit_process_syscalls(metadata: dict) -> None:
     require(sys.platform == "linux", "--process-syscalls requires Linux")
     binaries = build_workspace_binaries(metadata)
     require(binaries, "workspace has no binaries to trace")
-    with tempfile.TemporaryDirectory(prefix="lazaret-process-syscalls-") as directory:
+    with temporary_directory("process-syscalls-") as directory:
         for binary, executable in sorted(binaries.items()):
             trace_base = Path(directory) / binary
             command = [
@@ -620,7 +675,7 @@ def run_rust_tests() -> None:
 
 def run_ignored_workspace_corpus(metadata: dict) -> None:
     binaries = build_workspace_binaries(metadata)
-    with tempfile.TemporaryDirectory(prefix="lazaret-ignored-workspace-corpus-") as directory:
+    with temporary_directory("ignored-workspace-corpus-") as directory:
         root = Path(directory)
         run_physical(
             [
@@ -661,6 +716,7 @@ def verify(process_syscalls: bool) -> None:
     run_ignored_workspace_corpus(metadata)
     run([sys.executable, "-m", "unittest", "discover", "-s", "tests/tooling"])
     audit_workspace_lints(metadata)
+    audit_rust_file_sizes(metadata)
     audit_dependencies(metadata)
     run_models()
     if process_syscalls: