#!/usr/bin/env python3
"""Compare the 40250 GC header registry with the classic wire adapter.

This is intentionally a static audit, not a replacement for wire tests.  It
keeps the reference-side registry count and dynamic/static classification
machine-checkable so a new packet alias or compatibility branch cannot be
silently treated as 40250 parity.
"""

from __future__ import annotations

import argparse
import json
import re
from pathlib import Path

import refroot  # script directory is on sys.path when run directly


REFERENCE_DYNAMIC = {
    "HEADER_GC_QUEST_INFO",
    "HEADER_GC_DUEL_START",
    "HEADER_GC_CHAT",
    "HEADER_GC_SYNC_POSITION",
    "HEADER_GC_SHOP",
    "HEADER_GC_SCRIPT",
    "HEADER_GC_MESSENGER",
    "HEADER_GC_GUILD",
    "HEADER_GC_DUNGEON",
    "HEADER_GC_NPC_POSITION",
    "HEADER_GC_LAND_LIST",
    "HEADER_GC_HYBRIDCRYPT_KEYS",
    "HEADER_GC_HYBRIDCRYPT_SDB",
}

# These are documented current-client adaptations.  They remain differences:
# allowing them here makes the report useful without pretending they are in the
# reference CMainPacketHeaderMap.
KNOWN_CURRENT_DYNAMIC_ADAPTATIONS = {
    "HDR_GC_WHISPER",
    "HDR_GC_SYMBOL_DATA",
}


def active_lines(text: str, *, gaiden: bool = False) -> list[str]:
    """Keep the 40250 branch used by the checked-in Win32 project.

    The snapshot contains both GAIDEN and non-GAIDEN enum aliases in Packet.h.
    The UserInterface project does not define GAIDEN, so evaluating that one
    branch is necessary before comparing numeric header identities.  Other
    optional blocks are intentionally retained: the registry source itself
    uses the improved-encryption symbols supplied by ServiceDefs.h.
    """
    out: list[str] = []
    stack: list[bool] = []
    enabled = True
    for line in text.splitlines():
        stripped = line.strip()
        if stripped in {"#if defined(GAIDEN)", "#ifdef GAIDEN"}:
            stack.append(enabled)
            enabled = enabled and gaiden
            continue
        if stripped == "#ifndef GAIDEN":
            stack.append(enabled)
            enabled = enabled and not gaiden
            continue
        if stripped == "#else" and stack:
            parent = stack[-1]
            enabled = parent and not enabled
            continue
        if stripped.startswith("#endif") and stack:
            enabled = stack.pop()
            continue
        if enabled and not stripped.startswith("//"):
            out.append(line)
    return out


def reference_registry(path: Path) -> list[dict[str, str]]:
    row = re.compile(
        r"\bSet\((HEADER_GC_[A-Z0-9_]+),\s*"
        r"CNetworkPacketHeaderMap::TPacketType\(sizeof\(([^)]+)\),\s*"
        r"(STATIC_SIZE_PACKET|DYNAMIC_SIZE_PACKET)\)"
    )
    out: list[dict[str, str]] = []
    for line in active_lines(path.read_text(errors="replace")):
        match = row.search(line)
        if match:
            out.append({"header": match.group(1), "struct": match.group(2), "kind": match.group(3)})
    return out


def function_body(text: str, name: str, return_type: str) -> str:
    match = re.search(
        rf"constexpr\s+{re.escape(return_type)}\s+{re.escape(name)}\s*\([^)]*\)\s*\{{(?P<body>.*?)\n\}}",
        text,
        re.S,
    )
    if not match:
        raise ValueError(f"cannot find {name} in wire header")
    return match.group("body")


def current_wire(path: Path) -> dict[str, list[str]]:
    text = path.read_text(errors="replace")
    size_body = function_body(text, "packet_size_gc", "int")
    dynamic_body = function_body(text, "is_dynamic_gc", "bool")
    case_re = re.compile(r"\bcase\s+(HDR_[A-Z0-9_]+)\s*:")
    return {
        "size_cases": case_re.findall(size_body),
        "dynamic_cases": case_re.findall(dynamic_body),
    }


def header_values(path: Path, prefix: str) -> dict[str, int]:
    values: dict[str, int] = {}
    pattern = re.compile(rf"\b({re.escape(prefix)}[A-Z0-9_]+)\s*=\s*(0x[0-9A-Fa-f]+|\d+)")
    for line in active_lines(path.read_text(errors="replace"), gaiden=False):
        match = pattern.search(line)
        if match:
            values.setdefault(match.group(1), int(match.group(2), 0))
    return values


def audit(reference_path: Path, wire_path: Path) -> dict:
    ref = reference_registry(reference_path)
    wire = current_wire(wire_path)
    reference_values = header_values(reference_path.parent / "Packet.h", "HEADER_GC_")
    current_values = header_values(wire_path, "HDR_")
    ref_dynamic = {row["header"].replace("HEADER_", "HDR_") for row in ref if row["kind"] == "DYNAMIC_SIZE_PACKET"}
    current_dynamic = set(wire["dynamic_cases"])
    extras = sorted(current_dynamic - ref_dynamic)
    missing = sorted(ref_dynamic - current_dynamic)
    unexpected_extras = sorted(set(extras) - KNOWN_CURRENT_DYNAMIC_ADAPTATIONS)
    reference_by_value: dict[int, list[dict[str, str]]] = {}
    for row in ref:
        value = reference_values.get(row["header"])
        if value is not None:
            reference_by_value.setdefault(value, []).append(row)

    current_registered_names = set(wire["size_cases"]) | current_dynamic
    current_by_value: dict[int, list[str]] = {}
    for name in current_registered_names:
        value = current_values.get(name)
        if value is not None:
            current_by_value.setdefault(value, []).append(name)

    reference_values_set = set(reference_by_value)
    current_values_set = set(current_by_value)
    numeric_missing = sorted(reference_values_set - current_values_set)
    numeric_extras = sorted(current_values_set - reference_values_set)

    kind_mismatches: list[dict[str, object]] = []
    for value, rows in sorted(reference_by_value.items()):
        expected = {row["kind"] for row in rows}
        actual = {
            "DYNAMIC_SIZE_PACKET" if name in current_dynamic else "STATIC_SIZE_PACKET"
            for name in current_by_value.get(value, [])
        }
        if actual and not actual.intersection(expected):
            kind_mismatches.append({"value": value, "reference": rows, "current": current_by_value[value]})

    result = {
        "reference": {
            "registered": len(ref),
            "static": sum(row["kind"] == "STATIC_SIZE_PACKET" for row in ref),
            "dynamic": sum(row["kind"] == "DYNAMIC_SIZE_PACKET" for row in ref),
            "dynamic_headers": sorted(ref_dynamic),
            "numeric_values": len(reference_values_set),
        },
        "current": {
            "packet_size_gc_cases": len(set(wire["size_cases"])),
            "is_dynamic_gc_cases": len(current_dynamic),
            "dynamic_headers": sorted(current_dynamic),
            "numeric_values": len(current_values_set),
        },
        "difference": {
            "missing_reference_dynamic": missing,
            "current_dynamic_extras": extras,
            "unapproved_current_dynamic_extras": unexpected_extras,
            "known_adaptations": sorted(set(extras) & KNOWN_CURRENT_DYNAMIC_ADAPTATIONS),
            "missing_reference_numeric_values": numeric_missing,
            "current_extra_numeric_values": numeric_extras,
            "kind_mismatches": kind_mismatches,
        },
        "status": "PASS"
        if not missing and not extras and not numeric_missing and not numeric_extras and not kind_mismatches
        else "PARTIAL",
    }
    return result


def main() -> int:
    parser = argparse.ArgumentParser()
    default_root = Path(__file__).resolve().parents[4]
    parser.add_argument("--repo-root", type=Path, default=default_root)
    parser.add_argument("--strict", action="store_true", help="return non-zero when the registry is not exact")
    args = parser.parse_args()

    root = args.repo_root.resolve()
    reference = refroot.reference_root(root) / "UserInterface/PythonNetworkStream.cpp"
    wire = root / "extension/src/net/classic/wire_classic.h"
    try:
        result = audit(reference, wire)
    except (OSError, ValueError) as exc:
        print(json.dumps({"status": "ERROR", "error": str(exc)}, indent=2, sort_keys=True))
        return 2
    print(json.dumps(result, indent=2, sort_keys=True))
    return 1 if args.strict and result["status"] != "PASS" else 0


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