port(2V0-a): EterPythonLib 窗口系统的头进镜像

2V0 的入口实测：audit/slices/2V0.json 的 212 个头里 164 个已经在镜像中，缺 48 个
（1 个 EterLib、5 个 EterPythonLib、42 个 UserInterface）；10 个单元里已移植 2 个。
本轮补 EterPythonLib 这 5 个，外加必须先有的 EterPythonLib/StdAfx.h —— port_header_gate
按 40250 的做法把每个头放在所属库的 StdAfx.h 后面编译，没有它这五个既看不到
<Python.h> 也看不到 STL 和 CSingleton。

6 个文件全是 port_copy.py 的机械拷贝，0 处手改。

顺带修正清单的一处误报：EterLib/Dimm.h 不是 40250 的源文件，是 Windows SDK <dimm.h>
的 MIDL 产物，port/common/shim/win32/Dimm.h 已经顶掉了；镜像一份反而遮住 shim，
EterLib/IME.h 的 #include "Dimm.h" 会去找 rpc.h。拷进来验证后删掉。

只有头，没有 .cpp，所以没有 port-map 条目 —— 只有声明，没有可调用的东西。
port_header_gate 新增 6 个 TU 全绿；ctest 27/27。

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

audit/history.jsonl

@@ -500,3 +500,4 @@
 {"time": "2026-09-23T22:35:00Z", "event": "known_issue", "unit": "Android GBK 编解码挂起；优先保证 macOS 可用", "ported": [], "adapted": [], "deleted": [], "divergent": [], "needs_live": [], "decision": {"by": "user", "statement": "先不管这个问题了，先保证 macOS 能正常使用，Android 后期再想办法"}, "finding": {"symptom": "build-android.sh 默认 API 24 下 extension/src/net/text_codec.cpp 编译失败：use of undeclared identifier 'iconv_open' / 'iconv' / 'iconv_close'", "root_cause": "text_codec.cpp:10 用 __has_include(<iconv.h>) 探测。NDK 里该头在任何 API 等级都存在，但声明被 __BIONIC_AVAILABILITY_GUARD(28) 挡住，探测方式与 bionic 的可用性模型不匹配", "measured": "NDK 28.2.13676358 clang++ -fsyntax-only：--target=aarch64-linux-android24 报 3 处 undeclared；android28 通过", "why_api_bump_is_wrong": "NDK iconv.h 注释写明 bionic 只支持 utf8/ascii/usascii/utf16be/utf16le/utf32be/utf32le/wchart，没有 GB2312/GBK。升到 28 后 iconv_open(\"GB2312\",\"UTF-8\") 运行期返回 -1，encode_one()/decode_wire() 对所有输入都返回 false，连纯 ASCII 也被拒，比现有 #else 的 ASCII 回退分支（text_codec.cpp:104-127）更糟。附带成本：APK min_sdk 需 >=28；config/android/pyconfig.h 是 API 24 下 configure 的（HAVE_LANGINFO_H 因 nl_langinfo 需 API 26 被关掉），要重跑 gen_pyconfig.sh android 并重验内建模块表", "future_fix": "1) 探测改为 #if __has_include(<iconv.h>) && !defined(__ANDROID__)，不要写 __ANDROID_API__ >= 28；2) Android 自带 GBK<->Unicode 表（约 22k 码位）。40250 原客户端走 Windows MultiByteToWideChar(CP949/CP936)，跨平台无对应物；macOS 能用只是系统 libiconv 恰好带 GBK"}, "docs": ["docs/PORT-PLAN.md 新增「已知问题：Android 的 GBK 编解码（挂起，2026-09-23）」小节"], "tests": "仅文档改动；诊断由 NDK clang++ -fsyntax-only 双 API 等级实测得出", "not_done": "text_codec.cpp 未改动；Android 扩展构建目前仍是红的"}
 {"time": "2026-09-23T23:30:00Z", "event": "test_fix", "unit": "net 测试去 flaky：net_classic_session_test + net_loopback_test", "ported": [], "adapted": [], "deleted": [], "divergent": [], "needs_live": [], "trigger": "macOS 全量 ctest 连续两次各挂 1 个 net 测试，且每次挂的不是同一个（第一次 net.loopback_flow，第二次 net.classic_session）", "root_causes": ["accept_loopback(): poll() 的 `<= 0` 把 EINTR 当成对端从未连接；且单片 500ms 超时在有负载时不够。50 轮压测里 4 次失败全部源于此", "net_classic_session_test.cpp:570 的 pump 循环条件抄自上面的死端口用例，但此处 connect 会成功，stage 先经过 LoggingIn，循环提前一步退出", "同文件 5 处手写 pump 循环只给 10-20ms 等真实 loopback socket（非阻塞 connect 完成 / connect 被拒 / 对端 FIN）", "net_loopback_test.cpp 10 处 run(..., N, []{return false;}) 是固定空转而非等待条件，等不够就断言失败"], "fixes": ["accept_loopback() 改为 EINTR 重试 + 5s 截止时间（listener 本就是非阻塞，超时后 accept 不会挂住）", "新增 pump_until(session, want, max_ms=2000)，替换全部 5 处手写循环；等的是断言真正关心的那个状态", "替换 replacement PHASE_LOGIN 处的单次 pump（前一条 CHECK 已经断言过「不在同一帧」，这里只需要后续某次 pump 收到）", "net_loopback_test 的 10 处固定空转改为条件等待；服务端批量收包用该批最后一个包的标志作条件（TCP 有序，最后一个到了说明前面都到了），客户端事件用 decltype 声明的本地累加器在谓词里 drain"], "speedup": {"net_classic_session_test": "61.2s -> 3.5s：两处 this-*-host-must-not-resolve.invalid 换成带空标签的名字。本机解析器不按 RFC 6761 本地拒绝 .invalid，每次发往上游等满 30s 超时；空标签让 getaddrinfo 在语法层返回同一个 EAI_NONAME，实测 <10ms", "net_loopback_test": "固定空转改条件退出后单跑 0.94s", "ctest_total": "78.7s -> 8.7s"}, "tests": "net_classic_session_test 100/100（修复前 50 轮挂 4 轮）；net_loopback_test 100/100（修复前单独跑第 7 轮即挂）；全量 ctest 27/27；port_map.py check 0 error；参考 EterPack.cpp 的 8 个 pack key 值逐个 git grep --cached -w，均未入库", "not_done": "只改测试，未改产品代码。这些都是测试自身的计时/错误处理缺陷，不是 mtnet 的行为问题"}
 {"time": "2026-09-23T11:00:00Z", "event": "port_round", "unit": "PORT-PLAN 批次 2P step 4 (macOS): the 40250 script layer starts inside the Godot process", "ported": [], "adapted": ["UserInterface/UserInterface.cpp:RunMainScript -> extension/src/platform/ScriptLib/PythonBoot.cpp (Start/RunMainScript/RunLine/Stop; 40250 keeps CPythonLauncher on Main()'s stack, Godot owns this process so the launcher has to outlive the call that starts it)", "EterBase/Debug.cpp:LogBox, EterBase/Debug.cpp:LogBoxf -> extension/src/platform/EterBase/Debug.cpp (stderr plus LastLogBoxMessage(); ScriptLib Traceback() reports through LogBoxf after PyErr_Fetch has cleared the exception, so this is the only place a script traceback survives)"], "deleted": [], "divergent": ["RunMainScript RegisterCommandLine passes the command line through unchanged: the -cs/-ncs auto-login decoding needs EterBase/Utils.cpp SplitLine and EterLib/Util.cpp base64_decode, both declared but not ported (批次 2V1). Nothing in this host passes -cs today.", "RunMainScript drops '--pause-before-create-window' (system(\"pause\")) and the XTrap init check: no console and no XTrap on macOS/Android.", "Of the 37 module initializers RunMainScript calls, only initpack() is called — the other 36 are 2V slices."], "needs_live": [], "staging": {"native": "extension/src/platform/ScriptLib/PythonBoot.{h,cpp} (godot-free, placed under ScriptLib because that is the directory port_logic/port_platform drop when `mtpython` is absent; compiled with the UserInterface library's definitions so its `#ifdef _DISTRIBUTE` gives __DEBUG__ = 0). extension/src/platform/EterBase/LogBox.h. extension/src/python_host_node.{h,cpp} -> GDScript class Metin2PythonHost (start/is_running/run_main_script/run_line/stop; each returns \"\" or the reason, and a build without the interpreter answers with a distinguishable message instead of failing to register).", "gdscript": "project/python_host_check.gd in two halves — the interpreter + stdlib half needs no 40250 pack and runs on any platform, the system.py half needs the pack and is skipped without it. project/python_host_test.gd (--script) is the desktop entry; MT_TEST_MODE=python in project/client_main.gd is the on-device entry, since export templates do not accept --script.", "boot": "Metin2PythonHost.start(\"\") stages the stdlib through Metin2Python.stdlib_path(), which until now had no caller outside python_stdlib_test.gd."}, "status_note": "macOS only. Android/iOS in-process is untouched: the APK cross-build is blocked on the GBK iconv issue (see the PORT-PLAN known-issue section, deferred by the user 2026-09-23), and a device has no 40250 Client/pack, so on a phone only the interpreter + stdlib half of the check can run. prototype.RunApp() is 2V0: system.py still stops at its first import. The real boot path (AppFlow) does not start the interpreter.", "tests": ["godot --headless --path project --script python_host_test.gd -> PASS: interpreter starts in-process, is_running(), a second start() is a no-op, `import os, string, types, codecs`, u'\\xe4'.encode('cp1252'), sys.path non-empty, system.py stops at ImportError: No module named app (identical to ctest port.python_launcher), stop() tears it down", "MT_TEST_MODE=python godot --headless --path project -> same result through client_main.gd", "ctest --test-dir build -> 27/27 passed (14.51s)"], "not_done": "Android/iOS 进程内验证；正式启动路径起解释器（2V0）；RunMainScript 的 -cs 命令行解码（2V1）", "commit": "2P-step4-macos"}
+{"time": "2026-09-23T11:40:00Z", "event": "port_round", "unit": "PORT-PLAN 批次 2V0-a: the EterPythonLib window-system headers enter the mirror", "ported": ["EterPythonLib/StdAfx.h", "EterPythonLib/PythonGraphic.h", "EterPythonLib/PythonWindow.h", "EterPythonLib/PythonWindowManager.h", "EterPythonLib/PythonSlotWindow.h", "EterPythonLib/PythonGridSlotWindow.h"], "adapted": [], "deleted": [], "divergent": [], "needs_live": [], "staging": {"scope": "headers only; no .cpp unit and therefore no port-map entry — declarations, nothing to call. The EterPythonLib .cpp units are 2V0-b.", "mechanical": "port_copy.py copy, 0 of 6 with manual edits: every one compiles as a mechanical CP949->UTF-8 / CRLF->LF / include-case copy.", "stdafx": "EterPythonLib/StdAfx.h had to come first: port_header_gate compiles each header behind its library's StdAfx.h, as 40250 does, and without it these five see neither <Python.h>, the STL, nor CSingleton.", "slice_list_correction": "audit/slices/2V0.json lists EterLib/Dimm.h as a missing header. It is not: 40250 vendors the MIDL output of the Windows SDK <dimm.h> as EterLib/Dimm.h, and extension/src/port/common/shim/win32/Dimm.h already shims it (a mirror copy shadows the shim for EterLib/IME.h's quoted include and breaks the build on 'rpc.h'). Copied, found, and removed again."}, "status_note": "2V0 entry state measured: of the slice's 212 headers 164 were already in the mirror, 48 were not (1 EterLib = the SDK shim above, 5 EterPythonLib = this round, 42 UserInterface = 2V0-d, all behind UserInterface/PythonApplication.h). Of the 10 units, 2 are ported (ScriptLib/PythonLauncher.cpp, UserInterface/PythonPackModule.cpp) and 8 are not.", "tests": ["cmake --build build --target port_header_gate -> the 6 new gate translation units compile (each header alone, twice, behind EterPythonLib/StdAfx.h)", "ctest --test-dir build -> 27/27 passed"], "not_done": "EterPythonLib/*.cpp（2V0-b）；wndMgr/grp/grpImage/grpText 模块（2V0-c）；42 个 UserInterface 头 + PythonApplicationModule.cpp（2V0-d）", "commit": "2V0-a"}