#!/usr/bin/env python3
"""Generate deterministic real EROFS fixtures for B19a."""

from __future__ import annotations

import argparse
import hashlib
import json
import os
from pathlib import Path
import shutil
import struct
import subprocess


SUPER = 1024
MAGIC = 0xE0F5E1E2
CRC32C_POLY = 0x82F63B78


def sha256(path: Path) -> str:
    digest = hashlib.sha256()
    with path.open("rb") as stream:
        for chunk in iter(lambda: stream.read(1024 * 1024), b""):
            digest.update(chunk)
    return digest.hexdigest()


def crc32c(data: bytes | bytearray, seed: int = 0xFFFFFFFF) -> int:
    value = seed
    for byte in data:
        value ^= byte
        for _ in range(8):
            value = (value >> 1) ^ (CRC32C_POLY if value & 1 else 0)
    return value & 0xFFFFFFFF


def deterministic_value(label: str, length: int) -> bytes:
    seed = (label + "|").encode("ascii")
    return (seed * (length // len(seed) + 1))[:length]


def load_spec(path: Path) -> dict:
    spec = json.loads(path.read_text(encoding="ascii"))
    if (
        spec.get("schema") != 1
        or spec.get("batch") != "B19a"
        or spec.get("candidate") != "P15-022"
    ):
        raise SystemExit("invalid B19a fixture spec identity")
    return spec


class Image:
    def __init__(self, data: bytes | bytearray):
        self.data = bytearray(data)
        if len(self.data) < SUPER + 144 or self.u32(SUPER) != MAGIC:
            raise ValueError("invalid EROFS image")
        self.block_bits = self.data[SUPER + 12]
        if self.block_bits < 9 or self.block_bits > 16:
            raise ValueError("invalid EROFS block size")
        self.block_size = 1 << self.block_bits
        self.blocks = self.u32(SUPER + 36)
        self.limit = self.blocks << self.block_bits
        self.meta_blkaddr = self.u32(SUPER + 40)
        self.xattr_blkaddr = self.u32(SUPER + 44)
        self.root_nid = self.u16(SUPER + 14)
        if self.limit > len(self.data):
            raise ValueError("EROFS image is truncated")

    @classmethod
    def load(cls, path: Path) -> "Image":
        return cls(path.read_bytes())

    def u16(self, offset: int) -> int:
        if offset < 0 or offset + 2 > len(self.data):
            raise ValueError("u16 read is out of bounds")
        return struct.unpack_from("<H", self.data, offset)[0]

    def u32(self, offset: int) -> int:
        if offset < 0 or offset + 4 > len(self.data):
            raise ValueError("u32 read is out of bounds")
        return struct.unpack_from("<I", self.data, offset)[0]

    def inode(self, nid: int) -> dict[str, int]:
        offset = (self.meta_blkaddr << self.block_bits) + (nid << 5)
        if offset > self.limit or 64 > self.limit - offset:
            raise ValueError("inode is out of bounds")
        inode_format = self.u16(offset)
        inode_size = 64 if inode_format & 1 else 32
        xattr_count = self.u16(offset + 2)
        xattr_size = 0 if xattr_count == 0 else 12 + 4 * (xattr_count - 1)
        size = struct.unpack_from(
            "<Q" if inode_size == 64 else "<I", self.data, offset + 8
        )[0]
        return {
            "nid": nid,
            "offset": offset,
            "inode_size": inode_size,
            "xattr_size": xattr_size,
            "layout": (inode_format >> 1) & 7,
            "size": size,
            "start_block": self.u32(offset + 16),
        }

    def inode_data(self, inode: dict[str, int], logical: int, length: int) -> bytes:
        if logical > inode["size"] or length > inode["size"] - logical:
            raise ValueError("inode data range is out of bounds")
        if inode["layout"] == 0:
            physical = (inode["start_block"] << self.block_bits) + logical
        elif inode["layout"] == 2:
            tail_start = ((inode["size"] + self.block_size - 1) // self.block_size - 1) * self.block_size
            if logical < tail_start:
                if logical + length > tail_start:
                    raise ValueError("inode data read crosses inline tail")
                physical = (inode["start_block"] << self.block_bits) + logical
            else:
                physical = (
                    inode["offset"]
                    + inode["inode_size"]
                    + inode["xattr_size"]
                    + logical
                    - tail_start
                )
        else:
            raise ValueError("unsupported directory layout")
        if physical > self.limit or length > self.limit - physical:
            raise ValueError("inode data is outside the image")
        return bytes(self.data[physical : physical + length])

    @staticmethod
    def parse_dirblock(data: bytes) -> list[tuple[bytes, int]]:
        if len(data) < 12:
            raise ValueError("directory block is short")
        first_name = struct.unpack_from("<H", data, 8)[0]
        if first_name < 12 or first_name >= len(data) or first_name % 12:
            raise ValueError("directory first name offset is invalid")
        count = first_name // 12
        entries = []
        previous_offset = 0
        previous_name: bytes | None = None
        for index in range(count):
            entry = index * 12
            nid = struct.unpack_from("<Q", data, entry)[0]
            start = struct.unpack_from("<H", data, entry + 8)[0]
            end = (
                struct.unpack_from("<H", data, entry + 20)[0]
                if index + 1 < count
                else len(data)
            )
            if (
                start < first_name
                or start >= len(data)
                or (index == 0 and start != first_name)
                or (index != 0 and start <= previous_offset)
                or end <= start
                or end > len(data)
            ):
                raise ValueError("directory name bounds are invalid")
            span = data[start:end]
            if index + 1 < count:
                if b"\0" in span:
                    raise ValueError("non-trailing directory name contains NUL")
                name = span
            else:
                name = span.split(b"\0", 1)[0]
            if not name or len(name) > 255 or b"/" in name:
                raise ValueError("directory name is invalid")
            if previous_name is not None and previous_name >= name:
                raise ValueError("directory names are not strictly ordered")
            entries.append((name, nid))
            previous_offset = start
            previous_name = name
        return entries

    def directory_entries(self, inode: dict[str, int]) -> list[tuple[bytes, int]]:
        result = []
        previous_name: bytes | None = None
        logical = 0
        while logical < inode["size"]:
            length = min(self.block_size, inode["size"] - logical)
            block = self.parse_dirblock(self.inode_data(inode, logical, length))
            if previous_name is not None and previous_name >= block[0][0]:
                raise ValueError("directory block boundary is not ordered")
            result.extend(block)
            previous_name = block[-1][0]
            logical += length
        return result

    def resolve(self, path: str) -> dict[str, int]:
        inode = self.inode(self.root_nid)
        for component in path.strip("/").encode("ascii").split(b"/"):
            if not component:
                continue
            matches = [nid for name, nid in self.directory_entries(inode) if name == component]
            if len(matches) != 1:
                raise ValueError(f"fixture path is absent or ambiguous: {path}")
            inode = self.inode(matches[0])
        return inode

    def body_offset(self, inode: dict[str, int]) -> int:
        return inode["offset"] + inode["inode_size"]

    def body_shape(self, inode: dict[str, int]) -> tuple[int, list[int], list[int]]:
        body = self.body_offset(inode)
        size = inode["xattr_size"]
        if size <= 12 or body > self.limit or size > self.limit - body:
            raise ValueError("target xattr body is invalid")
        shared_count = self.data[body + 4]
        header_size = 12 + shared_count * 4
        if header_size > size:
            raise ValueError("target shared count is invalid")
        shared_ids = [self.u32(body + 12 + index * 4) for index in range(shared_count)]
        inline_entries = []
        cursor = body + header_size
        end = body + size
        while cursor < end:
            if cursor + 4 > end:
                raise ValueError("target inline xattr header is truncated")
            name_length = self.data[cursor]
            value_length = self.u16(cursor + 2)
            total = (4 + name_length + value_length + 3) & ~3
            if total > end - cursor:
                raise ValueError("target inline xattr entry is truncated")
            inline_entries.append(cursor)
            cursor += total
        return header_size, shared_ids, inline_entries

    def update_checksum(self) -> None:
        if self.u32(SUPER + 8) & 1:
            struct.pack_into("<I", self.data, SUPER + 4, 0)
            struct.pack_into(
                "<I", self.data, SUPER + 4, crc32c(self.data[SUPER : self.block_size])
            )

    def write(self, path: Path) -> None:
        path.write_bytes(self.data)


def set_epoch(path: Path) -> None:
    for entry in sorted(path.rglob("*"), reverse=True):
        os.utime(entry, (0, 0), follow_symlinks=False)
    os.utime(path, (0, 0), follow_symlinks=False)


def create_source(path: Path, spec: dict) -> None:
    path.mkdir(parents=True)
    path.chmod(0o755)
    shared_value = deterministic_value(
        "shared-value-p15-022", spec["fixture"]["shared_value_bytes"]
    )
    inline_value = deterministic_value(
        "inline-value-p15-022", spec["fixture"]["inline_value_bytes"]
    )
    files = []
    for index in range(spec["fixture"]["peer_count"]):
        prefix = f"peer-{index:03d}-"
        suffix = "x" * (spec["fixture"]["peer_name_bytes"] - len(prefix) - 4)
        files.append(path / f"{prefix}{suffix}.bin")
    target = path / "target.bin"
    files.append(target)
    for index, entry in enumerate(files):
        entry.write_bytes(f"P15-022 file {index:03d}\n".encode("ascii"))
        entry.chmod(0o644)
        os.setxattr(entry, b"user.shared", shared_value)
        if entry == target:
            os.setxattr(entry, b"user.inline", inline_value)
        else:
            os.setxattr(
                entry,
                b"user.peer",
                deterministic_value(f"peer-{index:03d}", 31),
            )

    budget = path / "budget"
    budget.mkdir()
    budget.chmod(0o755)
    for index in range(spec["fixture"]["budget_file_count"]):
        entry = budget / f"budget-{index:03d}.bin"
        entry.write_bytes(b"B19a budget\n")
        entry.chmod(0o644)
        value = bytearray(
            deterministic_value(
                f"budget-value-{index:03d}",
                spec["fixture"]["budget_value_bytes"],
            )
        )
        struct.pack_into("<I", value, 0, index)
        os.setxattr(entry, b"user.budget", bytes(value))
    set_epoch(path)


def build_image(source: Path, output: Path, spec: dict) -> list[str]:
    mkfs = shutil.which("mkfs.erofs")
    if mkfs is None:
        raise SystemExit("mkfs.erofs is required")
    command = [
        mkfs,
        "-d0",
        "-T0",
        "--all-time",
        "--all-root",
        "--workers=1",
        "--sort=path",
        f"-U{spec['fixture']['uuid']}",
        "-x2",
        "-Eforce-inode-extended",
        str(output),
        str(source),
    ]
    completed = subprocess.run(
        command, check=False, stdout=subprocess.PIPE, stderr=subprocess.STDOUT
    )
    if completed.returncode != 0:
        raise SystemExit(
            "mkfs.erofs failed: " + completed.stdout.decode("utf-8", "replace")
        )
    return command


def mutate(valid: Path, output: Path, mutation: str, spec: dict) -> None:
    image = Image.load(valid)
    target = image.resolve(spec["fixture"]["target"])
    body = image.body_offset(target)
    _, shared_ids, inline_entries = image.body_shape(target)
    if mutation == "corrupt-shared-count":
        image.data[body + 4] = 255
    elif mutation == "corrupt-shared-id":
        if not shared_ids:
            raise SystemExit("valid fixture has no shared xattr ID")
        struct.pack_into("<I", image.data, body + 12, 0xFFFFFFFF)
    elif mutation == "corrupt-inline-name":
        if not inline_entries or image.data[inline_entries[0]] == 0:
            raise SystemExit("valid fixture has no inline xattr entry")
        image.data[inline_entries[0] + 4] = 0
    else:
        raise SystemExit(f"unknown B19a mutation: {mutation}")
    image.update_checksum()
    image.write(output)


def generate(output: Path, work: Path, spec: dict) -> dict:
    if output.exists() or work.exists():
        raise SystemExit("B19a output and work paths must be absent")
    output.mkdir(parents=True)
    source = work / "source"
    create_source(source, spec)
    valid = output / "valid.erofs"
    command = build_image(source, valid, spec)
    image = Image.load(valid)
    root = image.inode(image.root_nid)
    root_entries = image.directory_entries(root)
    target = image.resolve(spec["fixture"]["target"])
    _, shared_ids, inline_entries = image.body_shape(target)
    if root["size"] <= image.block_size:
        raise SystemExit("B19a root directory did not span multiple blocks")
    if not shared_ids or not inline_entries:
        raise SystemExit("B19a target lacks both shared and inline xattrs")
    target_positions = [
        index for index, (name, _) in enumerate(root_entries) if name == b"target.bin"
    ]
    if target_positions != [len(root_entries) - 1]:
        raise SystemExit("B19a target is not uniquely sorted after all peers")
    for mutation in (
        "corrupt-shared-count",
        "corrupt-shared-id",
        "corrupt-inline-name",
    ):
        mutate(valid, output / f"{mutation}.erofs", mutation, spec)
    fsck = shutil.which("fsck.erofs")
    if fsck is None:
        raise SystemExit("fsck.erofs is required")
    completed = subprocess.run(
        [fsck, str(valid)], check=False, stdout=subprocess.PIPE, stderr=subprocess.STDOUT
    )
    if completed.returncode != 0:
        raise SystemExit(
            "fsck.erofs rejected B19a valid fixture: "
            + completed.stdout.decode("utf-8", "replace")
        )
    hashes = {path.name: sha256(path) for path in sorted(output.glob("*.erofs"))}
    frozen = spec.get("fixture_sha256", {})
    if frozen and hashes != frozen:
        raise SystemExit(f"B19a fixture hashes differ: {hashes}")
    report = {
        "block_size": image.block_size,
        "command": [*command[:-2], "OUTPUT/valid.erofs", "WORK/source"],
        "fixture_sha256": hashes,
        "inline_entries": len(inline_entries),
        "root_directory_blocks": (root["size"] + image.block_size - 1) // image.block_size,
        "root_entry_count": len(root_entries),
        "shared_count": len(shared_ids),
        "status": "PASS",
        "target_entry_index": target_positions[0],
    }
    (output / "manifest.json").write_text(
        json.dumps(report, indent=2, sort_keys=True) + "\n", encoding="ascii"
    )
    return report


def main() -> int:
    parser = argparse.ArgumentParser()
    parser.add_argument("--spec", type=Path, required=True)
    parser.add_argument("--output", type=Path, required=True)
    parser.add_argument("--work", type=Path, required=True)
    args = parser.parse_args()
    spec = load_spec(args.spec)
    try:
        report = generate(args.output, args.work, spec)
    finally:
        shutil.rmtree(args.work, ignore_errors=True)
    print(json.dumps(report, sort_keys=True))
    return 0


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