#!/usr/bin/env python3
"""Warn when the working tree has CR bytes but git index is LF-only.

Tarballs copied from Windows often have CRLF on disk while `git pull` on Ubuntu
deploys LF from the index. Run after unpack:

  python3 scripts/check_worktree_eol.py
  bash scripts/normalize_worktree_lf.sh   # optional fix
"""

from __future__ import annotations

import subprocess
import sys
from pathlib import Path

ROOT = Path(__file__).resolve().parents[1]

# Same critical paths as check_text_eol.py
PATTERNS = [
    "*.sh",
    "server/**/*.py",
    "scripts/**/*.py",
    "deploy/**/*.sh",
    "docker/**/*.py",
    "web/agent/**/*.py",
    "frontend/src/**/*.vue",
    "frontend/src/**/*.ts",
    ".cursor/rules/**/*.mdc",
]


def git_ls_files(pattern: str) -> list[str]:
    r = subprocess.run(
        ["git", "ls-files", pattern],
        cwd=ROOT,
        capture_output=True,
        text=True,
        check=False,
    )
    return [p for p in r.stdout.splitlines() if p.strip()]


def index_has_cr(path: str) -> bool:
    r = subprocess.run(
        ["git", "show", f":{path}"],
        cwd=ROOT,
        capture_output=True,
        check=False,
    )
    return r.returncode == 0 and b"\r" in r.stdout


def worktree_has_cr(path: str) -> bool:
    full = ROOT / path
    if not full.is_file():
        return False
    return b"\r" in full.read_bytes()


def main() -> int:
    seen: set[str] = set()
    mismatch: list[str] = []
    worktree_only: list[str] = []

    for pattern in PATTERNS:
        for path in git_ls_files(pattern):
            if path in seen:
                continue
            seen.add(path)
            wt = worktree_has_cr(path)
            if not wt:
                continue
            if index_has_cr(path):
                worktree_only.append(path)
            else:
                mismatch.append(path)

    if not mismatch and not worktree_only:
        print(f"check_worktree_eol: OK ({len(seen)} paths, working tree matches LF index)")
        return 0

    if mismatch:
        print(
            f"CRLF in working tree but LF in git index ({len(mismatch)} file(s)) — "
            "bash/scripts may fail on Ubuntu:",
            file=sys.stderr,
        )
        for p in sorted(mismatch)[:50]:
            print(f"  {p}", file=sys.stderr)
        if len(mismatch) > 50:
            print(f"  ... and {len(mismatch) - 50} more", file=sys.stderr)

    if worktree_only:
        print(f"CR in both worktree and index ({len(worktree_only)} file(s)) — run check_text_eol.py", file=sys.stderr)

    print(
        "\nFix tracked files:\n"
        "  git config core.autocrlf false\n"
        "  bash scripts/normalize_worktree_lf.sh",
        file=sys.stderr,
    )
    return 1


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