# -*- coding: utf-8 -*-
r'''给 .ps1 文件补回 UTF-8 BOM，并检查哪些文件缺 BOM。

为什么必须补：
    PowerShell 读 .ps1 时，**没有 BOM 就按系统代码页（中文 Windows = GBK）解**。
    文件本身是 UTF-8，于是中文全变乱码，接着就是一堆
    "Unexpected token" 语法错误 —— 脚本直接跑不了。

    edit 工具会把 BOM 剥掉，所以每次用 edit 改完 .ps1 都要补一次。
    （这个坑今天踩了两次。）

用法：python _build/fix_ps1_bom.py
'''
from __future__ import annotations

import sys
from pathlib import Path

ROOT = Path(r"C:\Users\shu\Desktop\deepseekagent\LanRemote")
BOM = b"\xef\xbb\xbf"

SKIP_DIRS = (".git", "dist_exe", "pyinstaller_work", "_build\\jc", "out_java",
             "build", "node_modules")


def should_skip(path: Path) -> bool:
    text = str(path)
    return any(s in text for s in SKIP_DIRS)


def main() -> int:
    fixed = []
    already = []
    no_bom = []

    for path in sorted(ROOT.rglob("*.ps1")):
        if should_skip(path):
            continue
        raw = path.read_bytes()
        if raw.startswith(BOM):
            already.append(path)
            continue
        # 补 BOM
        path.write_bytes(BOM + raw)
        fixed.append(path)

    print("  补了 BOM 的 .ps1（%d 个）：" % len(fixed))
    for path in fixed:
        print("    " + str(path.relative_to(ROOT)))
    if not fixed:
        print("    （无）")

    print()
    print("  本来就有 BOM 的: %d 个" % len(already))

    # 复核：所有 .ps1 都该有 BOM
    print()
    print("  复核：")
    missing = []
    for path in sorted(ROOT.rglob("*.ps1")):
        if should_skip(path):
            continue
        if not path.read_bytes().startswith(BOM):
            missing.append(path)
    if missing:
        print("    还有 %d 个缺 BOM：" % len(missing))
        for path in missing:
            print("      " + str(path.relative_to(ROOT)))
        return 1
    print("    所有 .ps1 都有 BOM ✓")

    # 顺手看看 .bat 是不是 GBK（另一种编码规则）
    print()
    print("  .bat 编码检查（应该是 GBK/ANSI）：")
    for path in sorted(ROOT.rglob("*.bat")):
        if should_skip(path):
            continue
        raw = path.read_bytes()
        try:
            raw.decode("gbk")
            note = "GBK ✓"
        except UnicodeDecodeError:
            try:
                raw.decode("utf-8")
                note = "其实是 UTF-8（可能有问题，.bat 要用 GBK）"
            except UnicodeDecodeError:
                note = "两种都不是？"
        print("    %-42s %s" % (str(path.relative_to(ROOT)), note))
    return 0


if __name__ == "__main__":
    sys.exit(main())