feat(VM): 打通「清单项 -> cd 槽」的精确映射规则, 新增 tools/tdc_slotmap.py
装配循环(@4362)分支树读通:
kind===2(@4820) -> 不进 cd: 清空 r3 转存 r4, r4 在收尾 @4905 另作他用
v===null/undefined(@4799) -> 'null'
typeof v==='number'(@4738) -> (v||0)
kind===1(@4724) -> 原样拼(v 是 JSON 碎片串, 串内逗号会切出额外槽)
其余(@4703) -> 加引号
收尾 @4753: typeof w==='number' 才再 +1 槽
同一构建端到端校验: 37 项 -> 38 槽, 仅第 33 项(模块1)占 2 槽
('641547489,1791732001' -> 641547489 | 1791732001 精确命中)。
工具支持 --items(运行时取样) + --plain(解密明文) 对拍。
This commit is contained in:
yml2213
@@ -0,0 +1,174 @@
|
||||
# -*- coding: utf-8 -*-
|
||||
"""
|
||||
tdc_slotmap.py —— 把「清单项」精确映射到「cd 槽号」。
|
||||
|
||||
为什么需要单独一层（承 har/1ms_dianxuan/PROGRESS.md #34/#35/#36）:
|
||||
#34 槽位序 = 清单模块的 require 顺序（追加式数组，已核）
|
||||
#36 **槽数 ≠ 项数**: 装配循环里 `kind=1` 的分支是 `out += v`（不加引号直接拼），
|
||||
所以某些项的返回值**本身带逗号**时会一次性产出多个槽。
|
||||
实测同一构建: 37 项 -> 38 槽，多出的那一处是第 33 项（模块 1）
|
||||
get() 返回 "641547489,1791732001"。
|
||||
|
||||
因此映射不能是"第 i 项 = 第 i 槽"，必须按每个采集器的**返回值形状**推算:
|
||||
|
||||
kind = ret[2]（可能缺省）
|
||||
kind == 1: ret[0] 原样拼 -> 占 1 + ret[0].count(",") 槽（数字则 1 槽）
|
||||
其它: ret[0] 占 1 槽
|
||||
ret[1] 是 number: 再 +1 槽（实测 kind=2 的项第二元素是 null，不触发）
|
||||
|
||||
本模块提供:
|
||||
* `slot_map(prog, items)` —— 给静态导出的项 + 运行时取样到的返回值形状，算出映射
|
||||
* `from_runtime_items(prog)` —— 需要浏览器环境（iframe 里读清单各项 get()），见 CLI
|
||||
* CLI 校验模式: 与一份已解密的 collect 明文对拍，确认推算出的槽数 == 实际槽数
|
||||
|
||||
用法:
|
||||
# 用运行时取样 + 本次 getData 解密出的明文, 做一次端到端对拍
|
||||
.venv/bin/python tools/tdc_slotmap.py <tdc.js> \\
|
||||
--items /tmp/items.json --plain /tmp/plain.txt
|
||||
"""
|
||||
from __future__ import annotations
|
||||
|
||||
import argparse
|
||||
import json
|
||||
import sys
|
||||
from pathlib import Path
|
||||
from typing import Dict, List, Optional, Sequence, Tuple
|
||||
|
||||
ROOT = Path(__file__).resolve().parents[1]
|
||||
if str(ROOT) not in sys.path:
|
||||
sys.path.insert(0, str(ROOT))
|
||||
if str(ROOT / "tools") not in sys.path:
|
||||
sys.path.insert(0, str(ROOT / "tools"))
|
||||
|
||||
import tdc_layout # noqa: E402
|
||||
import tdc_program as tp # noqa: E402
|
||||
|
||||
|
||||
def slots_of_item(ret) -> int:
|
||||
"""一个采集器的 get() 返回值会占几个 cd 槽。
|
||||
|
||||
ret 是运行时取到的值（或其形状描述）。规则见 PROGRESS.md #36。
|
||||
"""
|
||||
if ret is None:
|
||||
return 1
|
||||
if not isinstance(ret, (list, tuple)) or not ret:
|
||||
return 1
|
||||
first = ret[0]
|
||||
kind = ret[2] if len(ret) >= 3 else None
|
||||
n = 1
|
||||
if kind == 1 and isinstance(first, str):
|
||||
# 不加引号直接拼 -> 串里的逗号会切出额外槽
|
||||
n += first.count(",")
|
||||
# 第二个元素是 number 时再占一槽（实测 kind=2 的第二元素是 null，不触发）
|
||||
if len(ret) >= 2 and isinstance(ret[1], (int, float)) and not isinstance(ret[1], bool):
|
||||
n += 1
|
||||
return n
|
||||
|
||||
|
||||
def slot_map(seq: Sequence[int], rets: Sequence) -> List[Tuple[int, int, int]]:
|
||||
"""返回 [(清单项下标, 模块 id, 该首槽在 cd 中的槽号)]。
|
||||
|
||||
某一项占多槽时，只给它的**首槽**；其余槽是你按 ret 内容自然拼出来的。
|
||||
"""
|
||||
out: List[Tuple[int, int, int]] = []
|
||||
slot = 0
|
||||
for i, m in enumerate(seq):
|
||||
out.append((i, m, slot))
|
||||
slot += slots_of_item(rets[i] if i < len(rets) else None)
|
||||
return out
|
||||
|
||||
|
||||
def total_slots(seq: Sequence[int], rets: Sequence) -> int:
|
||||
return sum(slots_of_item(rets[i] if i < len(rets) else None) for i in range(len(seq)))
|
||||
|
||||
|
||||
def load_items(path: str) -> List:
|
||||
"""读 cdp_eval 的取样输出（{items:[{head,types,len,...}]}）。
|
||||
|
||||
运行时只能带回 head（截断过），所以这里按 types/head 重建一个**够用的**形状。
|
||||
"""
|
||||
d = json.load(open(path, encoding="utf-8"))
|
||||
res = d.get("result", d)
|
||||
items = res.get("items") or []
|
||||
rets = []
|
||||
for it in items:
|
||||
types = it.get("types")
|
||||
head = it.get("head")
|
||||
if not isinstance(types, list):
|
||||
rets.append(None)
|
||||
continue
|
||||
rebuilt = []
|
||||
for t, h in zip(types, head or []):
|
||||
if t == "string":
|
||||
rebuilt.append(h if isinstance(h, str) else str(h))
|
||||
elif t == "number":
|
||||
rebuilt.append(h if isinstance(h, (int, float)) else 0)
|
||||
elif t == "null":
|
||||
rebuilt.append(None)
|
||||
elif t == "array":
|
||||
rebuilt.append([])
|
||||
else:
|
||||
rebuilt.append(None)
|
||||
rets.append(rebuilt)
|
||||
return rets
|
||||
|
||||
|
||||
def main() -> int:
|
||||
ap = argparse.ArgumentParser(description="清单项 -> cd 槽号的精确映射")
|
||||
ap.add_argument("path")
|
||||
ap.add_argument("--items", default=None, help="运行时取样的 items.json")
|
||||
ap.add_argument("--plain", default=None, help="已解密的 collect 明文（用于对拍校验）")
|
||||
ap.add_argument("--json", default=None, help="把映射写成 JSON")
|
||||
args = ap.parse_args()
|
||||
|
||||
prog = tp.load(args.path)
|
||||
reg = dict(prog.module_registry(0))
|
||||
_mid, _entry, seq = tdc_layout.find_manifest_module(prog)
|
||||
rets: List = load_items(args.items) if args.items else []
|
||||
|
||||
print(f"文件 : {args.path}")
|
||||
print(f"清单模块 : registry[{_mid}]")
|
||||
print(f"清单项数 : {len(seq)}")
|
||||
if args.items:
|
||||
print(f"运行时取样 : {len(rets)} 项")
|
||||
print(f"推算 cd 槽数 : {total_slots(seq, rets)}")
|
||||
|
||||
if rets:
|
||||
print("\n项 模块 工厂 取到的值(截断) 占槽 首槽")
|
||||
for i, m in enumerate(seq):
|
||||
e = reg.get(m)
|
||||
r = rets[i] if i < len(rets) else None
|
||||
s = str(r)[:36]
|
||||
n = slots_of_item(r)
|
||||
start = slot_map(seq, rets)[i][2]
|
||||
print(f"{i:2d} {m:3d} @{e:<6} {s:<38} {n:^3} {start:2d}")
|
||||
|
||||
if args.plain:
|
||||
cd = json.loads(open(args.plain, encoding="latin1").read()
|
||||
.rstrip("\x00").strip()).get("cd") or []
|
||||
print(f"\n明文实际 cd 槽数: {len(cd)}")
|
||||
if rets:
|
||||
ok = len(cd) == total_slots(seq, rets)
|
||||
print(f">>> 推算 == 实际 ? {ok}")
|
||||
if not ok:
|
||||
print(" (差 %d 槽: 检查是否有项返回值的逗号计数没算准)" % (len(cd) - total_slots(seq, rets)))
|
||||
else:
|
||||
print("\n逐项 -> 槽对位:")
|
||||
sm = slot_map(seq, rets)
|
||||
for (i, m, st) in sm:
|
||||
n = slots_of_item(rets[i])
|
||||
vals = [repr(cd[st + k])[:24] for k in range(min(n, len(cd) - st))]
|
||||
print(f" 项{i:2d}(模块{m:2d}) -> 槽 {st}..{st + n - 1} {' | '.join(vals)}")
|
||||
|
||||
if args.json and rets:
|
||||
Path(args.json).write_text(json.dumps(
|
||||
{"manifestModule": _mid, "seq": list(seq),
|
||||
"map": [{"item": i, "module": m, "firstSlot": st, "slots": slots_of_item(rets[i])}
|
||||
for (i, m, st) in slot_map(seq, rets)]},
|
||||
ensure_ascii=False, indent=1), encoding="utf-8")
|
||||
print(f"映射 -> {args.json}")
|
||||
return 0
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
raise SystemExit(main())
|
||||
Reference in New Issue
Block a user