test: Integrate mq_core tests into GitHub Actions

.github/workflows/all_test.yml

@@ -65,10 +65,15 @@ jobs:
           # Scene selection:
           # - ci_top_attention_doc_page_build validates doc build through the prebuilt Docker image.
           # - ci_top_attention_bin_kvtest keeps the Rust kv_test entry under the testbed scene contract.
+          # - ci_top_attention_mq_core keeps MQ correctness coverage inside the same CI testbed contract.
           suite["scenes"] = {
               key: value
               for key, value in suite["scenes"].items()
-              if key in ("ci_top_attention_doc_page_build", "ci_top_attention_bin_kvtest")
+              if key in (
+                  "ci_top_attention_doc_page_build",
+                  "ci_top_attention_bin_kvtest",
+                  "ci_top_attention_mq_core",
+              )
           }
 
           # Profile selection:
@@ -91,6 +96,7 @@ jobs:
           # - Keep the original per-scene scales from ci_test_list.yaml.
           # - ci_top_attention_doc_page_build stays on n1_kvowner_dram_3gib.
           # - ci_top_attention_bin_kvtest stays on n1_kvowner_dram_20gib.
+          # - ci_top_attention_mq_core stays on n1_kvowner_dram_20gib.
 
           out_path.write_text(
               yaml.safe_dump(suite, sort_keys=False, allow_unicode=False),

AGENTS.md

@@ -9,6 +9,9 @@ Keep this document concise.
 - Git operations are limited to basic `stage`, `unstage`, `commit`, and `push`. Do not use other Git operations.
 - Prefer contraction over compatibility by default. Do not add compatibility layers, deprecated paths, or aliases unless the task explicitly requires them.
 - Prefer one canonical name for one concept. Avoid synonym parameters, duplicated entrypoints, and parallel config surfaces.
+- For test entrypoints, match the real execution model directly. If a test is a standalone script/process test, invoke it as a script/process; do not wrap it in `pytest` just for uniformity.
+- Do not forward pytest-style flags (`-k`, `-q`, node selectors, etc.) through direct-process test wrappers unless the wrapper explicitly implements and documents that selector surface.
+- For new integration or process-lifecycle tests, prefer direct process startup with explicit arguments and explicit exit-code checks over adding new pytest-only wrappers.
 - Control branching deliberately. Prefer a small, explicit, enumerated set of supported branches in the style of a Rust enum over open-ended proliferation of near-duplicate cases.
 - When extending a surface, prefer folding the new case into an existing finite branch set. If a new branch is unavoidable, make it explicit, bounded, and easy to list exhaustively.
 - Names for testbed-scoped concepts should say `testbed` explicitly. Avoid generic names for testbed-only modes, ports, roots, workdirs, and other testbed-scoped settings.

AGENTS_CN.md

@@ -9,6 +9,9 @@
 - Git 操作仅限基础的 `stage`、`unstage`、`commit` 和 `push`。不要使用其他 Git 操作
 - 默认优先收束而不是兼容。除非任务明确要求，否则不要添加兼容层、废弃路径或别名
 - 一个概念优先只保留一个正式名字。避免同义参数、重复入口和并行配置面
+- 对测试入口，要直接匹配真实执行模型。如果测试本质上是独立脚本 / 独立进程测试，就按脚本 / 进程直接启动；不要为了表面统一再额外包一层 `pytest`
+- 对直接启动进程的测试包装器，不要透传 `-k`、`-q`、node selector 等 pytest 风格参数，除非该包装器显式实现并文档化了这组筛选接口
+- 新增集成测试或进程生命周期测试时，优先采用“直接启动进程 + 显式参数 + 显式检查退出码”的模式，而不是继续新增 pytest 专用包装层
 - 有意识地控制分支。优先采用类似 Rust enum 的小而显式、可穷举罗列的有限分支集合，而不是开放式扩散出一批近似重复分支
 - 扩展一个 surface 时，优先把新情况折叠进已有的有限分支集合；如果确实必须新增分支，就让它保持显式、边界清楚、易于完整罗列
 - testbed 作用域内的概念，命名里应显式带上 `testbed`。对仅属于 testbed 的 mode、port、root、workdir 等设置，避免使用过于泛化的名字

fluxon_doc_cn/design/teststack_1_当前架构与CI测试流程.md

@@ -9,10 +9,11 @@
 **稳定结论：**
 
 - `teststack` 由三层组成：
-  - **suite 编译层**：将 `scene × scale × profile` 组合成可执行 case；
-  - **testbed 编排层**：拉起共享 testbed，保持 controller 在线，并基于 `fluxon_ops` / ops 接口调度部署 workload；
-  - **case 执行层**：处理每个 case 的 prepare、execute、collect、finalize。
+  - **上层：suite 编译层**：将 `scene × scale × profile` 组合成可执行 case；
+  - **中层：统一 case plan / dispatch 层**：把编译结果收敛成统一的 `prepare / execute / collect / finalize` 外壳，并按 runtime backend 分发；
+  - **下层：runtime backend 执行层**：分别承接 `CI` backend 和 `TEST_STACK` backend 的具体 prepare、execute、collect、finalize 实现。
 - `test_runner.py` 是统一执行器，覆盖 `CI` case、`TEST_STACK` benchmark case，以及 UI / GitOps 集成入口。
+- `test_runner.py` 当前主要承载上层和中层；`test_runner_runtime_backend.py` 承载下层 runtime backend 实现。
 - `start_test_bed.py` 只负责共享 testbed 的启动与 controller 侧 apply 编排，不承担通用测试执行职责。
 - `ci_2_virt_node.py` 是 GitHub Actions / 本地双逻辑节点 CI 的封装入口，负责串联打包、dispatch、拉起 testbed、运行 runner、构建文档等步骤。
 
@@ -56,6 +57,46 @@ flowchart TD
     B --> I[test_runner UI + GitOps]
 ```
 
+### 4.0 `test_runner` 内部的上中下分层
+
+这里要把“teststack 三层”与“`test_runner` 内部三层”区分开看。
+
+`teststack` 整体上仍然是：
+
+- suite 编译层
+- testbed 编排层
+- case 执行层
+
+但在 `test_runner` 自身内部，当前稳定实现已经进一步分成三层：
+
+| 层级 | 作用 | 当前主要落点 |
+| --- | --- | --- |
+| 上层 | 解析 suite、selector、`scene/scale/profile`，并 materialize `resolved_case` | `test_runner.py` |
+| 中层 | 将不同 case family 收敛成统一 `_CasePlan` 外壳，并负责统一 dispatch | `test_runner.py` |
+| 下层 | 按 runtime backend 执行具体 runtime 逻辑 | `test_runner_runtime_backend.py` |
+
+这里的关键点是：
+
+- **上层统一的是 schema 和 case 编译模型**；
+- **中层统一的是 `prepare / execute / collect / finalize` 的外壳**；
+- **下层不再按 `scene/scale/profile` 切分，而是按 runtime backend 切分**。
+
+这意味着：
+
+- `scene / scale / profile` 仍是一套统一输入模型；
+- `CI` 与 `TEST_STACK` 的差异，主要落在下层 runtime backend，而不是上层 schema。
+
+当前对应关系可以简化理解为：
+
+```text
+scene / scale / profile
+    -> resolved_case
+    -> _CasePlan
+    -> runtime backend dispatch
+       -> CI backend
+       -> TEST_STACK backend
+```
+
 ### 4.1 suite 编译层
 
 本层输入为 `ci_test_list.yaml`，主要定义三类核心对象与一类产物注册表：
@@ -135,6 +176,25 @@ flowchart TD
 - `summary.yaml` 是单次 run_dir 的终态摘要；
 - `resolved_case.yaml` / `resolved_case_full.yaml` 是单次 run 的编译产物。
 
+### 4.4 `test_runner.py` 与 `test_runner_runtime_backend.py` 的边界
+
+当前 repo 内已经开始把 `test_runner` 主体按“统一编译/分发”和“runtime backend 执行”拆开。
+
+稳定边界如下：
+
+| 文件 | 主要职责 | 不负责什么 |
+| --- | --- | --- |
+| `fluxon_test_stack/test_runner.py` | 上层 suite/schema/case 编译；中层 `_CasePlan` 编译与统一 dispatch；runner 入口、workdir 历史、通用 util | 不再直接承载大段 `CI` / `TEST_STACK` backend 细节 |
+| `fluxon_test_stack/test_runner_runtime_backend.py` | 下层 backend 运行逻辑：`_prepare_ci_case`、`_execute_ci_case`、`_prepare_test_stack_case`、`_execute_test_stack_case`、对应 finalize / result wait | 不解析 suite，不决定 `scene × scale × profile` 的组合空间 |
+
+这层拆分的目的不是制造第二套 case 模型，而是把：
+
+- **统一 case schema**
+- **统一 `_CasePlan` 外壳**
+- **不同 runtime backend 实现**
+
+三者分开，避免把所有逻辑继续堆在一个 `test_runner.py` 里。
+
 ## 5. teststack 的公共契约
 
 ### 5.1 两类场景
@@ -407,6 +467,12 @@ sequenceDiagram
 
 `test_runner.py` 会先把每个 case 编译成 `_CasePlan`。这里有一个通用骨架：所有 case 都分成 `prepare_phases / execute_phases / collect_phases` 三段。不同场景的差异不在“三段结构本身”，而在于每段里放哪些 runtime phase、每个 phase 覆盖哪些 instance，以及 run_dir 怎样 staging。
 
+这里要明确：
+
+- `_CasePlan` 属于中层统一外壳；
+- `CI` 和 `TEST_STACK` 都要先落到 `_CasePlan`；
+- 真正的 backend 差异，延后到下层 runtime backend 才展开。
+
 通用语义如下：
 
 - prepare phase 先准备场景依赖的 runtime、配置、脚本和共享目录；

fluxon_py/tests/test_config.yaml

@@ -1,2 +1,5 @@
-deployconf_path: ../../deployment/deployconf.yaml
 kv_svc_type: fluxon
+etcd_address: 127.0.0.1:2379
+cluster_name: fluxon-example-cluster
+shared_memory_path: /tmp/fluxon-example-cluster/shm
+shared_file_path: /tmp/fluxon-example-cluster/share

fluxon_py/tests/test_lib.py

@@ -32,12 +32,11 @@
 from setup_and_pack.utils.repo_config_utils import (
     _verify_host_port,
     _verify_url,
-    load_deployconf_etcd_address,
-    load_deployconf_fluxon_cluster_name,
-    load_deployconf_fluxon_shared_file_path,
-    load_deployconf_fluxon_shared_memory_path,
     load_test_config_mapping,
-    load_test_deployconf_path,
+    load_test_etcd_address_from_test_config,
+    load_test_fluxon_cluster_name_from_test_config,
+    load_test_fluxon_shared_file_path_from_test_config,
+    load_test_fluxon_shared_memory_path_from_test_config,
     load_test_kv_svc_type_from_test_config,
 )
 
@@ -50,10 +49,9 @@ def load_test_kv_svc_type(*, config_path: Optional[Path] = None) -> str:
 
 
 def load_test_kv_svc_ip(*, config_path: Optional[Path] = None) -> str:
-    """Load test backend host from the shared deployconf."""
-    deployconf_path = load_test_deployconf_path(config_path=config_path)
-    etcd_addr = load_deployconf_etcd_address(config_path=deployconf_path)
-    s, _port = _verify_host_port(etcd_addr, field="deployconf.global_envs.ETCD_FULL_ADDRESS")
+    """Load test backend host from test_config.yaml."""
+    etcd_addr = load_test_etcd_address_from_test_config(config_path=config_path)
+    s, _port = _verify_host_port(etcd_addr, field="test_config.yaml.etcd_address")
     if "://" in s or not s:
         raise ValueError("test backend host should be a host or IP without scheme, e.g. 127.0.0.1")
     return s
@@ -81,21 +79,18 @@ def load_test_mooncake_master_server_address(*, config_path: Optional[Path] = No
 
 
 def load_test_fluxon_cluster_name(*, config_path: Optional[Path] = None) -> str:
-    """Load required fluxon cluster name from the shared deployconf."""
-    deployconf_path = load_test_deployconf_path(config_path=config_path)
-    return load_deployconf_fluxon_cluster_name(config_path=deployconf_path)
+    """Load required Fluxon cluster name from test_config.yaml."""
+    return load_test_fluxon_cluster_name_from_test_config(config_path=config_path)
 
 
 def load_test_fluxon_share_mem_path(*, config_path: Optional[Path] = None) -> str:
-    """Load required fluxon shared memory path from the shared deployconf."""
-    deployconf_path = load_test_deployconf_path(config_path=config_path)
-    return load_deployconf_fluxon_shared_memory_path(config_path=deployconf_path)
+    """Load required Fluxon shared-memory path from test_config.yaml."""
+    return load_test_fluxon_shared_memory_path_from_test_config(config_path=config_path)
 
 
 def load_test_fluxon_share_file_path(*, config_path: Optional[Path] = None) -> str:
-    """Load required fluxon shared file path from the shared deployconf."""
-    deployconf_path = load_test_deployconf_path(config_path=config_path)
-    return load_deployconf_fluxon_shared_file_path(config_path=deployconf_path)
+    """Load required Fluxon shared-file path from test_config.yaml."""
+    return load_test_fluxon_shared_file_path_from_test_config(config_path=config_path)
 
 
 def load_test_chan_config(*, config_path: Optional[Path] = None) -> Dict[str, int]:
@@ -105,10 +100,9 @@ def load_test_chan_config(*, config_path: Optional[Path] = None) -> Dict[str, in
     """
     return {"capacity": 10, "ttl_seconds": 90, "weight": 1}
 
-# Resolve ETCD host/port and test configuration via config utils (no direct field access)
-_TEST_DEPLOYCONF_PATH = load_test_deployconf_path()
-_ETCD_ADDRESS = load_deployconf_etcd_address(config_path=_TEST_DEPLOYCONF_PATH)
-ETCD_HOST, _ETCD_PORT = _verify_host_port(_ETCD_ADDRESS, field="deployconf.global_envs.ETCD_FULL_ADDRESS")
+# Resolve ETCD host/port and test configuration via test_config.yaml (single explicit authority)
+_ETCD_ADDRESS = load_test_etcd_address_from_test_config()
+ETCD_HOST, _ETCD_PORT = _verify_host_port(_ETCD_ADDRESS, field="test_config.yaml.etcd_address")
 ETCD_PORT = int(_ETCD_PORT)
 KV_SVC_TYPE = load_test_kv_svc_type()
 KV_SVC_IP = load_test_kv_svc_ip()

fluxon_release/test_rsc/source/prepare.yaml

@@ -19,6 +19,8 @@ python_runtime:
           source: wheel
         - pinned: readerwriterlock==1.0.9
           source: wheel
+        - pinned: pytest==8.3.5
+          source: wheel
     zerorpc:
       requirements:
         - pinned: zerorpc==0.6.3

fluxon_test_stack/benchmark_full_matrix.yaml

@@ -193,8 +193,7 @@ artifact_sets:
       region: us-east-1
       key_prefix: profiles/fluxon_fastws
     release_artifacts:
-      python_wheel: fluxon-0.2.1-py3-none-any.whl
-      pyo3_wheel: fluxon_pyo3-0.2.1-cp38-abi3-manylinux_2_28_x86_64.whl
+      wheel: fluxon-0.2.1-py3-none-any.whl
     test_rsc_source: &test_rsc_source_fastws
       kind: FLUXON_OPS_FS_S3
       bucket: fluxon-release
@@ -215,8 +214,7 @@ artifact_sets:
       region: us-east-1
       key_prefix: profiles/fluxon_tquic
     release_artifacts:
-      python_wheel: fluxon-0.2.1-py3-none-any.whl
-      pyo3_wheel: fluxon_pyo3-0.2.1-cp38-abi3-manylinux_2_28_x86_64.whl
+      wheel: fluxon-0.2.1-py3-none-any.whl
     test_rsc_source: &test_rsc_source_tquic
       kind: FLUXON_OPS_FS_S3
       bucket: fluxon-release
@@ -237,8 +235,7 @@ artifact_sets:
       region: us-east-1
       key_prefix: profiles/fluxon_sockudo_ws
     release_artifacts:
-      python_wheel: fluxon-0.2.1-py3-none-any.whl
-      pyo3_wheel: fluxon_pyo3-0.2.1-cp38-abi3-manylinux_2_28_x86_64.whl
+      wheel: fluxon-0.2.1-py3-none-any.whl
     test_rsc_source: &test_rsc_source_sockudo_ws
       kind: FLUXON_OPS_FS_S3
       bucket: fluxon-release
@@ -259,8 +256,7 @@ artifact_sets:
       region: us-east-1
       key_prefix: profiles/fluxon_tcp
     release_artifacts:
-      python_wheel: fluxon-0.2.1-py3-none-any.whl
-      pyo3_wheel: fluxon_pyo3-0.2.1-cp38-abi3-manylinux_2_28_x86_64.whl
+      wheel: fluxon-0.2.1-py3-none-any.whl
     test_rsc_source: &test_rsc_source_tcp
       kind: FLUXON_OPS_FS_S3
       bucket: fluxon-release

fluxon_test_stack/ci_test_list.yaml

@@ -29,6 +29,14 @@ scenes:
       scales: [n1_kvowner_dram_20gib]
       profiles: [fluxon_tcp]
 
+  ci_top_attention_mq_core:
+    ci:
+      subject: mq
+      runtime_contract: cluster_kv_owner
+    select:
+      scales: [n1_kvowner_dram_20gib]
+      profiles: [fluxon_tcp]
+
   kv_read_heavy_zipf:
     test_stack:
       mode: KVSTORE
@@ -221,8 +229,7 @@ artifact_sets:
       region: us-east-1
       key_prefix: profiles/fluxon_fastws
     release_artifacts:
-      python_wheel: fluxon-0.2.1-py3-none-any.whl
-      pyo3_wheel: fluxon_pyo3-0.2.1-cp38-abi3-manylinux_2_28_x86_64.whl
+      wheel: fluxon-0.2.1-py3-none-any.whl
     test_rsc_source: &test_rsc_source_fastws
       kind: FLUXON_OPS_FS_S3
       bucket: fluxon-release
@@ -242,8 +249,7 @@ artifact_sets:
       region: us-east-1
       key_prefix: profiles/fluxon_tquic
     release_artifacts:
-      python_wheel: fluxon-0.2.1-py3-none-any.whl
-      pyo3_wheel: fluxon_pyo3-0.2.1-cp38-abi3-manylinux_2_28_x86_64.whl
+      wheel: fluxon-0.2.1-py3-none-any.whl
     test_rsc_source: &test_rsc_source_tquic
       kind: FLUXON_OPS_FS_S3
       bucket: fluxon-release
@@ -263,8 +269,7 @@ artifact_sets:
       region: us-east-1
       key_prefix: profiles/fluxon_sockudo_ws
     release_artifacts:
-      python_wheel: fluxon-0.2.1-py3-none-any.whl
-      pyo3_wheel: fluxon_pyo3-0.2.1-cp38-abi3-manylinux_2_28_x86_64.whl
+      wheel: fluxon-0.2.1-py3-none-any.whl
     test_rsc_source: &test_rsc_source_sockudo_ws
       kind: FLUXON_OPS_FS_S3
       bucket: fluxon-release
@@ -284,8 +289,7 @@ artifact_sets:
       region: us-east-1
       key_prefix: profiles/fluxon_tcp
     release_artifacts:
-      python_wheel: fluxon-0.2.1-py3-none-any.whl
-      pyo3_wheel: fluxon_pyo3-0.2.1-cp38-abi3-manylinux_2_28_x86_64.whl
+      wheel: fluxon-0.2.1-py3-none-any.whl
     test_rsc_source: &test_rsc_source_tcp
       kind: FLUXON_OPS_FS_S3
       bucket: fluxon-release
@@ -315,6 +319,7 @@ profiles:
             doc_site_base_url: example.com
           ci_top_attention_bin_kvtest:
             kv_test_rounds: all
+          ci_top_attention_mq_core: {}
         runtime_contracts:
           cluster_kv_owner: &cluster_kv_owner_runtime
             base_runtime:
@@ -460,6 +465,7 @@ profiles:
             doc_site_base_url: example.com
           ci_top_attention_bin_kvtest:
             kv_test_rounds: all
+          ci_top_attention_mq_core: {}
       test_stack:
         <<: *common_test_stack_runtime
   fluxon_sockudo_ws:
@@ -472,6 +478,7 @@ profiles:
             doc_site_base_url: example.com
           ci_top_attention_bin_kvtest:
             kv_test_rounds: all
+          ci_top_attention_mq_core: {}
       test_stack:
         <<: *common_test_stack_runtime
   fluxon_tcp:
@@ -484,6 +491,7 @@ profiles:
             doc_site_base_url: example.com
           ci_top_attention_bin_kvtest:
             kv_test_rounds: all
+          ci_top_attention_mq_core: {}
       test_stack:
         <<: *common_test_stack_runtime
   redis_sharded: