Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 7 additions & 2 deletions .github/workflows/package-smoke.yml
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,7 @@ on:
- "packages/loopx-core/**"
- "loopx/**"
- "tests/test_core_distribution.py"
- "tests/control_plane/test_installed_core_context.py"
- "packages/loopx-repo-health/**"
- "packages/loopx-obelisk/**"
- "packages/loopx-codex-provider-routing/**"
Expand All @@ -20,6 +21,7 @@ on:
- "packages/loopx-core/**"
- "loopx/**"
- "tests/test_core_distribution.py"
- "tests/control_plane/test_installed_core_context.py"
- "packages/loopx-repo-health/**"
- "packages/loopx-obelisk/**"
- "packages/loopx-codex-provider-routing/**"
Expand Down Expand Up @@ -56,10 +58,13 @@ jobs:
uv venv "$RUNNER_TEMP/core-env" --python .venv/bin/python
uv pip install --python "$RUNNER_TEMP/core-env/bin/python" "$RUNNER_TEMP"/core-dist/loopx_core-*.whl
echo "LOOPX_CORE_TEST_PYTHON=$RUNNER_TEMP/core-env/bin/python" >> "$GITHUB_ENV"
- name: Qualify installed execution, recovery, replay and refusal
- name: Qualify installed execution, context, recovery, replay and refusal
env:
LOOPX_USAGE_PING: "0"
run: uv run --extra test python -m pytest -q tests/test_core_distribution.py
run: >-
uv run --extra test python -m pytest -q
tests/test_core_distribution.py
tests/control_plane/test_installed_core_context.py

codex-provider-routing:
name: loopx-codex-provider-routing
Expand Down
29 changes: 29 additions & 0 deletions docs/architecture/rfcs/monorepo-distribution-split-v0.md
Original file line number Diff line number Diff line change
Expand Up @@ -3608,3 +3608,32 @@ at an immutable revision; the procedure, consent and check-scope clauses remain
unchanged. It is an on-demand supporting reference, not a runtime fetch or new
authority. Preserve the earlier failed installation gate and disk-limited base
replays separately from the repaired-head validation.

### 2026-11-13 — Installed context regression enters the package lane

The standalone core lane now invokes `tests/test_core_distribution.py` and
`tests/control_plane/test_installed_core_context.py` together, and both pull
request and main-push filters observe the context suite. The previous lane
omitted that suite despite its local qualification. An independent workflow
check rejects the unchanged base for all three omissions and accepts the
repaired command and filters. The clean Python 3.11.15 wheel's 1,885 tracked
runtime files match the source bytes; the installed artifact remains separate
from the source test environment.

Additional macOS arm64/Python 3.13.5/Node 26 observations use the same installed
core 2.3.0 and isolated SQLite/File complete-Turn fixtures, with synthetic Hosts
and usage telemetry disabled. The two selected Turn processes still load 525
LoopX Python modules, including 111 capability modules, and issue 220/221 Effect
RPCs. All 1,024 observed Python connect/address-resolution events target
loopback; the two Node listeners and their DNS lookups also use loopback. Four
Node processes, including the Host supervisors, load 247 distinct files. The
observer's own local-connection and named-import subprocess controls pass;
the earlier incomplete observer remains separate from this checked observation.

These observations identify local RPC traffic in this fixture; they do not
establish OS-wide network completeness, ordinary telemetry behavior, model
context adoption, paired performance or full costs. This delivery closes the
installed protocol's CI coverage gap without selecting D2/D5 or a minimal
preset. Real Host initial/revised/compacted/lost-ACK adoption, complete V1–V5/V10,
R1/M3, Standard/workspace publication and original-goal regression remain open.
Preserve the earlier failures, frozen budgets and unchanged governance owners.
23 changes: 23 additions & 0 deletions docs/architecture/rfcs/monorepo-distribution-split-v0.zh-CN.md
Original file line number Diff line number Diff line change
Expand Up @@ -2813,3 +2813,26 @@ owner 和验收约束保持开放。
核验的公开源文件,并固定不可变 revision;原流程、同意边界和检查范围条款保持
原义。它是按需支持资料,不是 runtime fetch 或新 authority。早期安装门失败和
受磁盘限制的 base 重放,与修复 head 的验证分别保留。

### 2026-11-13 — 安装上下文回归进入 package lane

独立 core lane 现在一起执行 `tests/test_core_distribution.py` 和
`tests/control_plane/test_installed_core_context.py`,pull request 与 main push
的路径过滤器也都观察上下文测试。此前该测试虽通过本地资格验证,lane 却未执行。
独立 workflow 检查在未改动的 base 上因这三处遗漏拒绝,在修复后的命令和过滤器上
通过。干净 Python 3.11.15 wheel 的 1,885 个 tracked runtime 文件与源码字节一致;
安装产物继续与源码测试环境隔离。

补充观察在 macOS arm64/Python 3.13.5/Node 26 上,使用同一独立 core 2.3.0、
隔离 SQLite/File 完整 Turn 夹具、合成 Host,并关闭使用遥测。两个选定 Turn
进程仍加载 525 个 LoopX Python 模块,其中 111 个 capability 模块,发出
220/221 次 Effect RPC。观察到的 1,024 次 Python 连接/地址解析事件均指向
loopback;两个 Node listener 及其 DNS lookup 也使用 loopback。包含 Host
supervisor 在内的四个 Node 进程加载 247 个不同文件。观察器自身的本地连接与
named-import 子进程对照通过;早期不完整观察器与本次核验后的观察分别保留。

这些观察识别了此夹具中的本地 RPC 流量,不证明 OS 全域网络完整性、普通遥测
行为、模型上下文采用、配对性能或完整成本。本交付补齐安装协议的 CI 覆盖缺口,
不选择 D2/D5 或最小 preset。真实 Host 的 initial/revised/compacted/lost-ACK
采用、完整 V1–V5/V10、R1/M3、Standard/workspace 发布及原始目标回归继续开放。
保留早期失败、冻结预算和未变的治理 owner。
Loading