ffproto/gen_from_cs.py at main · OpenFusionProject/ffproto · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
#!/usr/bin/env python3
"""Generate packets.json IR from C# packet struct sources.

Walks a directory of C# files, finds every

    [StructLayout(LayoutKind.Sequential, CharSet = CharSet.Unicode,
                  Pack = N, Size = M)]
    public struct NAME { ... }

and emits a single JSON document describing each struct's fields in a
language-agnostic form suitable for codegen in C/C++/C#/Python/Rust/etc.

JSON shape:

    {
      "version": 1,
      "defaults": {"pack": 4, "charset": "utf-16le"},
      "structs": [
        {"name": "...", "size": N, "pack": 4?, "fields": [
            {"name": "f", "type": "i32"},
            {"name": "s", "type": "wstr",   "length": 10},
            {"name": "a", "type": "i32",    "array": 5},
            {"name": "n", "type": "struct", "ref": "sItemBase"}
        ]}
      ]
    }

Each emitted struct's computed size (assuming sequential layout with the
declared Pack) is verified against the declared Size attribute; mismatches
are reported on stderr and exit code is non-zero.
"""

from __future__ import annotations

import argparse
import json
import re
import sys
from dataclasses import dataclass, field
from pathlib import Path
from typing import Iterable


# ----- C# token -> IR primitive --------------------------------------------

MARSHAL_TO_IR = {
    "I1": "i8",
    "U1": "u8",
    "I2": "i16",
    "I4": "i32",
    "U4": "u32",
    "I8": "i64",
    "U8": "u64",
    "R4": "f32",
}

# C# field-declaration type tokens (used for ByValArray element type and
# bare Struct fields). Maps the C# alias to the IR primitive name; anything
# not in this map is treated as a nested struct reference.
CS_TYPE_TO_IR = {
    "sbyte": "i8",
    "byte": "u8",
    "short": "i16",
    "ushort": "u16",
    "int": "i32",
    "uint": "u32",
    "long": "i64",
    "ulong": "u64",
    "float": "f32",
    "double": "f64",
    "bool": "u8",
    "char": "u16",
}

PRIM_SIZE = {
    "i8": 1, "u8": 1,
    "i16": 2, "u16": 2,
    "i32": 4, "u32": 4, "f32": 4,
    "i64": 8, "u64": 8, "f64": 8,
}


# ----- Regexes -------------------------------------------------------------

# Captures Pack and Size out of the [StructLayout(...)] attribute. The
# attribute body can span multiple lines, so we use DOTALL.
RE_STRUCT_LAYOUT = re.compile(
    r"\[StructLayout\((?P<body>[^\]]*?)\)\]\s*"
    r"public\s+struct\s+(?P<name>\w+)\s*\{(?P<body2>.*?)\n\}",
    re.DOTALL,
)

RE_PACK = re.compile(r"\bPack\s*=\s*(\d+)")
RE_SIZE = re.compile(r"\bSize\s*=\s*(\d+)")

# A single field: optional [MarshalAs(...)] attribute followed by
#   public <csharp-type> name;
# csharp-type may include [] for arrays. We tolerate whitespace/newlines.
RE_FIELD = re.compile(
    r"(?:\[MarshalAs\(\s*UnmanagedType\.(?P<kind>\w+)"
    r"(?:\s*,\s*SizeConst\s*=\s*(?P<size_const>\d+))?\s*\)\]\s*)?"
    r"public\s+(?P<cs_type>[\w\[\]]+)\s+(?P<name>\w+)\s*;",
    re.DOTALL,
)


# ----- Data types ----------------------------------------------------------

@dataclass
class Field:
    name: str
    type: str
    length: int | None = None   # for wstr (chars)
    array: int | None = None    # for fixed arrays
    ref: str | None = None      # for struct refs

    def to_json(self) -> dict:
        out: dict = {"name": self.name, "type": self.type}
        if self.ref is not None:
            out["ref"] = self.ref
        if self.length is not None:
            out["length"] = self.length
        if self.array is not None:
            out["array"] = self.array
        return out


@dataclass
class Struct:
    name: str
    size: int
    pack: int
    fields: list[Field] = field(default_factory=list)
    source: str = ""


# ----- Parsing -------------------------------------------------------------

def strip_array_suffix(cs_type: str) -> str:
    return cs_type[:-2] if cs_type.endswith("[]") else cs_type


def parse_field(match: re.Match) -> Field:
    kind = match.group("kind")
    size_const = match.group("size_const")
    cs_type = match.group("cs_type")
    name = match.group("name")
    base_cs = strip_array_suffix(cs_type)

    if kind in MARSHAL_TO_IR:
        return Field(name=name, type=MARSHAL_TO_IR[kind])

    if kind == "ByValTStr":
        if size_const is None:
            raise ValueError(f"ByValTStr without SizeConst on field {name!r}")
        return Field(name=name, type="wstr", length=int(size_const))

    if kind == "ByValArray":
        if size_const is None:
            raise ValueError(f"ByValArray without SizeConst on field {name!r}")
        elem_ir = CS_TYPE_TO_IR.get(base_cs)
        if elem_ir is not None:
            return Field(name=name, type=elem_ir, array=int(size_const))
        # Otherwise it's an array of structs.
        return Field(name=name, type="struct", ref=base_cs,
                     array=int(size_const))

    if kind == "Struct":
        return Field(name=name, type="struct", ref=base_cs)

    if kind is None:
        # No MarshalAs attribute. Infer from the C# type.
        elem_ir = CS_TYPE_TO_IR.get(base_cs)
        if elem_ir is not None:
            return Field(name=name, type=elem_ir)
        return Field(name=name, type="struct", ref=base_cs)

    raise ValueError(f"unsupported MarshalAs kind {kind!r} on field {name!r}")


def parse_file(path: Path) -> Iterable[Struct]:
    text = path.read_text(encoding="utf-8", errors="replace")
    for m in RE_STRUCT_LAYOUT.finditer(text):
        body_attr = m.group("body")
        body = m.group("body2")
        pack_m = RE_PACK.search(body_attr)
        size_m = RE_SIZE.search(body_attr)
        s = Struct(
            name=m.group("name"),
            size=int(size_m.group(1)) if size_m else 0,
            pack=int(pack_m.group(1)) if pack_m else 4,
            source=str(path),
        )
        for fm in RE_FIELD.finditer(body):
            s.fields.append(parse_field(fm))
        # If no Size attribute and we found public fields, this isn't a
        # wire struct we know how to describe — skip.
        if size_m is None and s.fields:
            continue
        # Empty placeholder structs (no public fields, just the C#
        # decompiler's private _0024PRIVATE_0024 byte) carry an explicit
        # padding byte in the IR so consumers don't have to special-case
        # zero-sized structs. The byte is real: it's what the client
        # actually puts on the wire (Marshal.SizeOf == 1).
        if not s.fields:
            s.fields.append(Field(name="_unused", type="u8"))
            s.size = 1
        yield s


# ----- Size verification ---------------------------------------------------

def primitive_size(ir_type: str) -> int | None:
    return PRIM_SIZE.get(ir_type)


def compute_struct_layout(s: Struct, by_name: dict[str, "Struct"],
                          cache: dict[str, tuple[int, int]]
                          ) -> tuple[int, int] | None:
    """Compute (size, alignment) for a struct.

    Alignment is the max alignment of any field, capped by the struct's
    own Pack. Returns None if a nested ref is unresolved.
    """
    if s.name in cache:
        return cache[s.name]
    pack = s.pack
    offset = 0
    max_align = 1
    for fld in s.fields:
        layout = field_size_align(fld, by_name, cache)
        if layout is None:
            return None
        fsize, falign = layout
        align = min(falign, pack)
        if offset % align:
            offset += align - (offset % align)
        offset += fsize
        max_align = max(max_align, align)
    struct_align = min(max_align, pack)
    if offset % struct_align:
        offset += struct_align - (offset % struct_align)
    cache[s.name] = (offset, struct_align)
    return offset, struct_align


def field_size_align(fld: Field, by_name: dict[str, "Struct"],
                     cache: dict[str, tuple[int, int]]
                     ) -> tuple[int, int] | None:
    if fld.type == "wstr":
        assert fld.length is not None
        return fld.length * 2, 2
    if fld.type == "struct":
        assert fld.ref is not None
        nested = by_name.get(fld.ref)
        if nested is None:
            return None
        layout = compute_struct_layout(nested, by_name, cache)
        if layout is None:
            return None
        nested_size, nested_align = layout
        count = fld.array if fld.array is not None else 1
        return nested_size * count, nested_align
    prim = primitive_size(fld.type)
    if prim is None:
        return None
    count = fld.array if fld.array is not None else 1
    return prim * count, prim


# ----- Main ----------------------------------------------------------------

def main() -> int:
    ap = argparse.ArgumentParser(description=__doc__)
    ap.add_argument("input_dir", type=Path,
                    help="Directory of C# source files to scan.")
    ap.add_argument("-o", "--output", type=Path, default=Path("packets.json"),
                    help="Output JSON path (default: packets.json)")
    ap.add_argument("--indent", type=int, default=2,
                    help="JSON indent (default: 2)")
    ap.add_argument("--strict", action="store_true",
                    help="Refuse to write output if any size mismatch or "
                         "unresolved struct reference is found.")
    args = ap.parse_args()

    if not args.input_dir.is_dir():
        print(f"error: {args.input_dir} is not a directory", file=sys.stderr)
        return 2

    structs: list[Struct] = []
    for path in sorted(args.input_dir.rglob("*.cs")):
        try:
            structs.extend(parse_file(path))
        except Exception as exc:
            print(f"warn: failed to parse {path}: {exc}", file=sys.stderr)

    if not structs:
        print("error: no [StructLayout]+Size structs found", file=sys.stderr)
        return 2

    by_name = {s.name: s for s in structs}
    cache: dict[str, tuple[int, int]] = {}
    mismatches: list[tuple[str, int, int]] = []
    unresolved: list[str] = []

    for s in structs:
        if s.size == 0 and not s.fields:
            # Empty packet: no body to verify.
            continue
        layout = compute_struct_layout(s, by_name, cache)
        if layout is None:
            unresolved.append(s.name)
            continue
        computed, _ = layout
        if computed != s.size:
            mismatches.append((s.name, s.size, computed))
            # There are some mismatches due to stale annotations
            # (like on sNanoTuneNeedItemInfo2CL in 1013).
            # Warn, but tolerate without `--strict` and update the IR size to match the computed size.
            if computed > s.size:
                s.size = computed

    packs = {s.pack for s in structs}
    default_pack = 4 if 4 in packs else next(iter(packs))

    if args.strict and (mismatches or unresolved):
        if unresolved:
            print(f"error: {len(unresolved)} struct(s) had unresolved nested"
                  f" refs: {', '.join(unresolved[:5])}"
                  f"{'...' if len(unresolved) > 5 else ''}", file=sys.stderr)
        if mismatches:
            print(f"error: {len(mismatches)} size mismatch(es):",
                  file=sys.stderr)
            for name, declared, computed in mismatches[:20]:
                print(f"  {name}: declared={declared} computed={computed}",
                      file=sys.stderr)
            if len(mismatches) > 20:
                print(f"  ... and {len(mismatches) - 20} more",
                      file=sys.stderr)
        print(f"error: refusing to write {args.output} due to --strict",
              file=sys.stderr)
        return 1

    out = {
        "defaults": {"pack": default_pack, "charset": "utf-16le"},
        "structs": [
            {
                "name": s.name,
                "size": s.size,
                **({"pack": s.pack} if s.pack != default_pack else {}),
                "fields": [f.to_json() for f in s.fields],
            }
            for s in structs
        ],
    }

    args.output.write_text(
        json.dumps(out, indent=args.indent) + "\n", encoding="utf-8")

    print(f"wrote {len(structs)} structs to {args.output}", file=sys.stderr)
    if unresolved:
        print(f"note: {len(unresolved)} struct(s) had unresolved nested refs"
              f" (size not verified): {', '.join(unresolved[:5])}"
              f"{'...' if len(unresolved) > 5 else ''}", file=sys.stderr)
    if mismatches:
        print(f"warn: {len(mismatches)} size mismatch(es):", file=sys.stderr)
        for name, declared, computed in mismatches:
            print(f"  {name}: declared={declared} computed={computed}",
                  file=sys.stderr)
    return 0


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