"""Convert Git porcelain-v2 -z bytes to JSON without splitting path whitespace.

No Git commands, file writes, or network calls are performed by this converter.
"""
import base64
import json
import sys

MAX_BYTES = 1024 * 1024


def path_value(raw):
    if not raw:
        raise ValueError("empty_path")
    try:
        display = raw.decode("utf-8", errors="strict")
    except UnicodeDecodeError:
        display = None
    return {"bytes_base64": base64.b64encode(raw).decode("ascii"), "utf8": display}


def parse_status(raw):
    if len(raw) > MAX_BYTES:
        raise ValueError("input_too_large")
    if raw and not raw.endswith(b"\0"):
        raise ValueError("missing_final_nul")
    chunks = raw.split(b"\0")[:-1] if raw else []
    records, headers, index = [], 0, 0
    layouts = {
        b"1": (8, ["xy", "submodule", "head_mode", "index_mode", "worktree_mode", "head_oid", "index_oid"]),
        b"2": (9, ["xy", "submodule", "head_mode", "index_mode", "worktree_mode", "head_oid", "index_oid", "score"]),
        b"u": (10, ["xy", "submodule", "stage1_mode", "stage2_mode", "stage3_mode", "worktree_mode", "stage1_oid", "stage2_oid", "stage3_oid"]),
    }
    while index < len(chunks):
        record = chunks[index]
        index += 1
        if record.startswith(b"# "):
            headers += 1
            continue
        if record.startswith((b"? ", b"! ")):
            records.append({"kind": "untracked" if record[:1] == b"?" else "ignored", "path": path_value(record[2:])})
            continue
        tag = record[:1]
        if tag not in layouts or record[1:2] != b" ":
            raise ValueError("unknown_record_type")
        count, fields = layouts[tag]
        parts = record.split(b" ", count)
        if len(parts) != count + 1 or any(not value for value in parts[1:count]):
            raise ValueError("truncated_record")
        try:
            metadata = dict(zip(fields, [value.decode("ascii") for value in parts[1:count]]))
        except UnicodeDecodeError as error:
            raise ValueError("non_ascii_metadata") from error
        item = {"kind": {b"1": "ordinary", b"2": "rename_or_copy", b"u": "unmerged"}[tag], **metadata, "path": path_value(parts[count])}
        if tag == b"2":
            if index >= len(chunks):
                raise ValueError("missing_original_path")
            item["original_path"] = path_value(chunks[index])
            index += 1
        records.append(item)
    return {"ok": True, "format": "git-porcelain-v2-z", "records": records, "ignored_headers": headers}


def main():
    raw = sys.stdin.buffer.read(MAX_BYTES + 1)
    try:
        result = parse_status(raw)
    except ValueError as error:
        print(json.dumps({"ok": False, "error": str(error)}))
        return 1
    print(json.dumps(result, ensure_ascii=True))
    return 0


if __name__ == "__main__":
    raise SystemExit(main())