Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.
C++
CI smokes: serve smoke expects the server-package preflight, server smoke accepts OK (skipped=N), nested-telemetry fixture deep enough for Python 3.12+
Niko1221fcca2ccb84b1
diff --git a/tools/minimal_cli_smoke.py b/tools/minimal_cli_smoke.pyindex 27661f483..b48e531ba 100644--- a/tools/minimal_cli_smoke.py+++ b/tools/minimal_cli_smoke.py@@ -81,11 +81,15 @@ def smoke() -> dict:with tempfile.TemporaryDirectory(prefix="strata-minimal-cli-") as directory:missing = str(Path(directory) / "missing-config.json")- serve = _record(_run(str(ROOT / "strata.py"), "serve", "--profile", missing,- "--offline", "--output", "jsonl"), 5)- if (serve.get("event") != "failed" or serve.get("exit_code") != 5 or- "cannot read run config" not in (serve.get("error") or {}).get("message", "")):- raise RuntimeError("missing run config did not fail before server startup")+ # serve needs the API server's packages; without them strata.py stops in its+ # preflight (exit 5, a plain message, nothing installed) before reading the config.+ serve = _run(str(ROOT / "strata.py"), "serve", "--profile", missing,+ "--offline", "--output", "jsonl")+ if (serve.returncode != 5 or serve.stdout.strip() or+ "missing packages the API server needs" not in serve.stderr or+ "never install anything" not in serve.stderr):+ raise RuntimeError(f"serve without server packages did not stop in the preflight: "+ f"exit {serve.returncode}, stderr {serve.stderr.strip()!r}")code = _record(_run(str(ROOT / "strata.py"), "code", "run", "--prompt-file",str(Path(directory) / "missing-prompt.txt"), "--policy", "review","--output", "jsonl"), 5)@@ -98,7 +102,7 @@ def smoke() -> dict:"platform": platform.platform(), "python": platform.python_version(),"optional_packages_visible": {"rich": False, "textual": False},"commands": ["--help", "status --json", "monitor --once --output json",- "serve --offline --output jsonl (missing config)",+ "serve --offline --output jsonl (server packages absent: preflight exit 5)","code run --output jsonl (missing harness)"],"telemetry_gets": len(paths), "non_gets": 0,"actual_engine_started": False, "passed": True}diff --git a/tools/minimal_server_smoke.py b/tools/minimal_server_smoke.pyindex 81780639c..b4bebcd37 100644--- a/tools/minimal_server_smoke.py+++ b/tools/minimal_server_smoke.py@@ -40,14 +40,16 @@ def smoke() -> dict:result = subprocess.run([sys.executable, "-m", "unittest", TEST_MODULE], cwd=ROOT,capture_output=True, text=True, timeout=240, check=False)summary = result.stdout + result.stderr- count = re.search(r"Ran (\d+) tests in [0-9.]+s\s+OK\s*$", summary)- passed = result.returncode == 0 and count is not None and int(count[1]) >= MIN_TESTS+ # Windows-only tests skip on POSIX, so unittest ends "OK (skipped=N)", not a bare "OK".+ count = re.search(r"Ran (\d+) tests in [0-9.]+s\s+OK(?: \(skipped=(\d+)\))?\s*$", summary)+ executed = int(count[1]) - int(count[2] or 0) if count else 0+ passed = result.returncode == 0 and count is not None and executed >= MIN_TESTSreturn {"schema_version": 1, "scope": "source_checkout_fake_engine_runtime_only","platform": platform.platform(), "python": platform.python_version(),"requirements_sha256": hashlib.sha256((ROOT / "requirements.txt").read_bytes()).hexdigest(),"optional_packages_visible": {name: visible[name] for name in ("rich", "textual")},"server_packages_visible": {name: visible[name] for name in ("numpy", "jinja2", "requests")},- "test_module": TEST_MODULE, "tests_run": int(count[1]) if count else None,+ "test_module": TEST_MODULE, "tests_run": executed if count else None,"duration_s": round(time.monotonic() - started, 2),"test_exit_code": result.returncode, "actual_gpu_engine": False,"passed": bool(passed), "test_summary": summary[-1000:]}diff --git a/tools/portable_monitor_smoke.py b/tools/portable_monitor_smoke.pyindex 5f6f42ad8..fc72bd2e7 100644--- a/tools/portable_monitor_smoke.py+++ b/tools/portable_monitor_smoke.py@@ -117,7 +117,7 @@ def _server() -> tuple[ThreadingHTTPServer, list[str], threading.Event]:"topology": {"mode": "single", "selected_devices": []},"rates": {"output_tokens_server_counter": 0},"devices": [], "quality": {"telemetry": "available"},- }).encode() if not malformed.is_set() else b"[" * 1100 + b"0" + b"]" * 1100+ }).encode() if not malformed.is_set() else b"[" * 100000 + b"0" + b"]" * 100000self.send_response(200)self.send_header("Content-Type", "application/json")self.send_header("Content-Length", str(len(body)))