Zack Buhman
ddf46ce8fd
The overall intent is to make writing a decompiler require less code duplication. "bits.hpp" and "stmt_enum.hpp" are replaced with "stmt_ins.hpp" which is generated directly from dsp-notes.csv.
129 lines
2.8 KiB
Python
129 lines
2.8 KiB
Python
import csv
|
|
import sys
|
|
from dataclasses import dataclass
|
|
from typing import Union
|
|
|
|
class _X:
|
|
def __repr__(self):
|
|
return 'x'
|
|
|
|
X = _X()
|
|
|
|
bit = Union[int, _X]
|
|
|
|
@dataclass
|
|
class Bits:
|
|
bits: tuple[bit, bit, bit, bit, bit, bit, bit, bit,
|
|
bit, bit, bit, bit, bit, bit, bit, bit,
|
|
bit, bit, bit, bit, bit, bit, bit, bit,
|
|
bit, bit, bit, bit, bit, bit, bit, bit]
|
|
|
|
def __repr__(self):
|
|
return ''.join(repr(i) for i in self.bits)
|
|
|
|
def __getitem__(self, key):
|
|
return list(reversed(self.bits))[key]
|
|
|
|
@dataclass
|
|
class Const:
|
|
name: str
|
|
bits: Bits
|
|
|
|
@dataclass
|
|
class Enum:
|
|
name: str
|
|
consts: list[Const]
|
|
|
|
@dataclass
|
|
class Struct:
|
|
name: str
|
|
bits: Bits
|
|
fields: list[tuple[int, str]]
|
|
|
|
@dataclass
|
|
class Namespace:
|
|
name: str
|
|
|
|
def read(path):
|
|
with open(path, newline='') as f:
|
|
reader = csv.reader(f, delimiter=',', quotechar='"')
|
|
return list(reader)
|
|
|
|
label_col = 2
|
|
bits_col = 3
|
|
|
|
def parse_struct(rows, ix):
|
|
return None, ix + 1
|
|
|
|
def validate_consts(consts):
|
|
by_index = set([
|
|
frozenset([ix for ix, bit in enumerate(const.bits) if bit is not X])
|
|
for const in consts
|
|
])
|
|
assert len(by_index) == 1
|
|
|
|
def get_bits(row):
|
|
return [
|
|
int(i) if i in {"0", "1"} else X
|
|
for i in row[bits_col:bits_col+32]
|
|
]
|
|
|
|
def get_nonbits(row):
|
|
return [
|
|
(31-ix, i)
|
|
for ix, i in enumerate(row[bits_col:bits_col+32])
|
|
if i and i not in {"0", "1"}
|
|
]
|
|
|
|
def get_name(row, key):
|
|
label0 = row[label_col]
|
|
assert label0.startswith(f'{key} '), label0
|
|
_, name = label0.split(' ', maxsplit=1)
|
|
return name
|
|
|
|
def parse_enum(rows, ix):
|
|
enum_name = get_name(rows[ix], "enum")
|
|
ix += 1
|
|
consts = []
|
|
|
|
while ix < len(rows):
|
|
label = rows[ix][label_col]
|
|
if not label or ' ' in label:
|
|
validate_consts(consts)
|
|
return Enum(enum_name, consts), ix
|
|
else:
|
|
const_name = label
|
|
bits = get_bits(rows[ix])
|
|
assert len(bits) == 32
|
|
consts.append(Const(const_name, Bits(bits)))
|
|
ix += 1
|
|
|
|
def parse_struct(rows, ix):
|
|
struct_name = get_name(rows[ix], "struct")
|
|
bits = get_bits(rows[ix])
|
|
fields = get_nonbits(rows[ix])
|
|
ix += 1
|
|
|
|
struct = Struct(struct_name, Bits(bits), fields)
|
|
return struct, ix
|
|
|
|
def parse(rows):
|
|
ix = 0
|
|
while ix < len(rows):
|
|
if rows[ix][0].strip():
|
|
yield Namespace(rows[ix][0].lower())
|
|
|
|
label = rows[ix][label_col]
|
|
if label.startswith('struct'):
|
|
ret, ix = parse_struct(rows, ix)
|
|
yield ret
|
|
elif label.startswith('enum' ):
|
|
ret, ix = parse_enum(rows, ix)
|
|
yield ret
|
|
else:
|
|
ix += 1
|
|
|
|
#events = list(parse(read(sys.argv[1])))
|
|
#from pprint import pprint
|
|
#pprint(events)
|