tdkpin/original/tools/recover_win16_imports.py

#!/usr/bin/env python3
"""Populate the import ledger from Wine 11.15 Win16 specs and target evidence."""

from __future__ import annotations

import csv
import hashlib
import re
import urllib.request
from pathlib import Path

ROOT = Path(__file__).resolve().parent.parent
LEDGER = ROOT / "IMPORT_RECONSTRUCTION.tsv"
WINE_TAG = "wine-11.15"
WINE_BASE = f"https://gitlab.winehq.org/wine/wine/-/raw/{WINE_TAG}"
SPEC_SOURCES = {
    "USER": (
        "dlls/user.exe16/user.exe16.spec",
        "d873fa88079c4f922cbba2fb3a75f40e2ff23b4d56434b0f4326169397539e32",
    ),
    "GDI": (
        "dlls/gdi.exe16/gdi.exe16.spec",
        "3bd84b26a9d4b7aa91a9114d0a46dc13e4d6fcd71290120afdd9b7cfaffe76ab",
    ),
    "KERNEL": (
        "dlls/krnl386.exe16/krnl386.exe16.spec",
        "8d799b701d1cf16d4e882b5e354dc6e357056d4adcd83e629571dc4b33b02075",
    ),
    "MMSYSTEM": (
        "dlls/mmsystem.dll16/mmsystem.dll16.spec",
        "31e54179db61fa95a1d90a9a230d2ae5afb6013ba37ae54e78990d1795694a73",
    ),
    "KEYBOARD": (
        "dlls/keyboard.drv16/keyboard.drv16.spec",
        "ca9a9e697e9afe0baf5cff7099ebb442169f36fa05293e78a0bb50fbefaa096d",
    ),
}
SPEC_FUNCTION = re.compile(
    r"^\s*(?P<ordinal>\d+)\s+"
    r"(?P<kind>pascal|cdecl|varargs|stub|register)"
    r"(?P<flags>(?:\s+-\S+)*)\s+"
    r"(?P<name>[A-Za-z0-9_@]+)\((?P<arguments>[^)]*)\)"
)
SPEC_EQUATE = re.compile(
    r"^\s*(?P<ordinal>\d+)\s+equate\s+(?P<name>[A-Za-z0-9_@]+)\s+(?P<value>\S+)"
)


def fetch_specs() -> dict[str, tuple[str, list[str]]]:
    specs = {}
    for library, (path, expected_digest) in SPEC_SOURCES.items():
        url = f"{WINE_BASE}/{path}"
        with urllib.request.urlopen(url, timeout=30) as response:
            payload = response.read()
        digest = hashlib.sha256(payload).hexdigest()
        if digest != expected_digest:
            raise RuntimeError(
                f"Wine source drift for {path}: expected {expected_digest}, got {digest}"
            )
        specs[library] = (url, payload.decode("utf-8").splitlines())
    return specs


def parse_specs(
    specs: dict[str, tuple[str, list[str]]],
) -> dict[tuple[str, str], dict[str, str]]:
    declarations = {}
    for library, (url, lines) in specs.items():
        for line_number, line in enumerate(lines, 1):
            match = SPEC_FUNCTION.match(line)
            if match:
                name = match.group("name")
                flags = match.group("flags").strip()
                kind = match.group("kind")
                prefix = f"{kind}16"
                if flags:
                    prefix += f" {flags}"
                declaration = f"{prefix} {name}({match.group('arguments').strip()})"
                declarations[(library, name.upper())] = {
                    "prototype": declaration,
                    "evidence": f"Wine {WINE_TAG} {url}#L{line_number}",
                    "notes": (
                        f"Wine Win16 ordinal {match.group('ordinal')}; "
                        "every TDKPIN call site still requires argument-level review."
                    ),
                    "status": "partial",
                    "stage": "prototype-sourced",
                }
                continue
            match = SPEC_EQUATE.match(line)
            if match:
                name = match.group("name")
                declarations[(library, name.upper())] = {
                    "prototype": f"equate16 {name} = {match.group('value')}",
                    "evidence": f"Wine {WINE_TAG} {url}#L{line_number}",
                    "notes": (
                        f"Wine Win16 ordinal {match.group('ordinal')}; this is runtime data, "
                        "not a callable function."
                    ),
                    "status": "verified",
                    "stage": "runtime-data",
                }
    return declarations


def recover_special_declarations(
    declarations: dict[tuple[str, str], dict[str, str]],
) -> None:
    keyboard_url = f"{WINE_BASE}/{SPEC_SOURCES['KEYBOARD'][0]}"
    declarations[("KEYBOARD", "ORDINAL_5")] = {
        "prototype": "pascal16 -ret16 AnsiToOem(str ptr)",
        "evidence": f"Wine {WINE_TAG} {keyboard_url}#L5",
        "notes": "Imported by ordinal; Wine identifies ordinal 5 as AnsiToOem.",
        "status": "partial",
        "stage": "prototype-sourced",
    }
    declarations[("KEYBOARD", "ORDINAL_6")] = {
        "prototype": "pascal16 -ret16 OemToAnsi(str ptr)",
        "evidence": f"Wine {WINE_TAG} {keyboard_url}#L6",
        "notes": "Imported by ordinal; Wine identifies ordinal 6 as OemToAnsi.",
        "status": "partial",
        "stage": "prototype-sourced",
    }
    declarations[("MMTIMER", "SYSTEMTIMERMAKE")] = {
        "prototype": (
            "pascal16 -ret16 SystemTimerMake("
            "HWND16 recipient, word delay_ms, word resolution_ms, word one_shot)"
        ),
        "evidence": (
            "MMTIMER.DLL 1000:0039 disassembly; export ordinal 1; "
            "f1d9ac980c7bfba5dc53eaa9e7cb2c3cd9b82f879ee8962ad40bf863d641bb49"
        ),
        "notes": (
            "Binary-reviewed wrapper around timeSetEvent16: it installs MMTIMER 1000:0002, "
            "posts message 0x0580 to recipient, and returns the multimedia timer id."
        ),
        "status": "verified",
        "stage": "binary-reviewed",
    }


def main() -> int:
    with LEDGER.open(newline="", encoding="utf-8") as handle:
        reader = csv.DictReader(handle, delimiter="\t")
        columns = list(reader.fieldnames or ())
        rows = list(reader)
    if "stage" not in columns:
        columns.insert(columns.index("status") + 1, "stage")

    declarations = parse_specs(fetch_specs())
    recover_special_declarations(declarations)
    import_keys = {(row["library"], row["import_name"].upper()) for row in rows}
    missing = sorted(import_keys - set(declarations))
    if missing:
        raise RuntimeError(f"no Win16 declaration for: {missing}")

    for row in rows:
        recovered = declarations[(row["library"], row["import_name"].upper())]
        preserve_review = row.get("stage") not in {
            "",
            "untyped-import",
            "prototype-sourced",
            "runtime-data",
            "binary-reviewed",
        }
        preserved = (
            {key: row.get(key, "") for key in ("status", "stage", "evidence", "notes")}
            if preserve_review
            else None
        )
        row.update(recovered)
        if preserved:
            row.update(preserved)

    temporary = LEDGER.with_suffix(LEDGER.suffix + ".tmp")
    with temporary.open("w", newline="", encoding="utf-8") as handle:
        writer = csv.DictWriter(
            handle, fieldnames=columns, delimiter="\t", lineterminator="\n"
        )
        writer.writeheader()
        writer.writerows(rows)
    temporary.replace(LEDGER)

    statuses = {}
    for row in rows:
        statuses[row["status"]] = statuses.get(row["status"], 0) + 1
    print(f"updated_import_slots={len(rows)}")
    print(
        f"unique_declarations={len({(row['library'], row['import_name']) for row in rows})}"
    )
    print("statuses=" + ",".join(f"{key}:{statuses[key]}" for key in sorted(statuses)))
    return 0


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