dissect.cstruct 4.8.dev8__tar.gz → 5.0.dev0__tar.gz

{dissect_cstruct-4.8.dev8 → dissect_cstruct-5.0.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dissect.cstruct
-Version: 4.8.dev8
+Version: 5.0.dev0
 Summary: A Dissect module implementing a parser for C-like structures: structure parsing in Python made easy
 Author-email: Dissect Team <dissect@fox-it.com>
 License-Expression: Apache-2.0

{dissect_cstruct-4.8.dev8 → dissect_cstruct-5.0.dev0}/dissect/cstruct/cstruct.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import copy
 import ctypes as _ctypes
 import struct
 import sys
@@ -9,7 +10,7 @@ from typing import TYPE_CHECKING, Any, BinaryIO, TypeVar, cast
 
 from dissect.cstruct.exceptions import ResolveError
 from dissect.cstruct.expression import Expression
-from dissect.cstruct.parser import CStyleParser, TokenParser
+from dissect.cstruct.parser import CStyleParser
 from dissect.cstruct.types import (
     LEB128,
     BaseArray,
@@ -25,6 +26,7 @@ from dissect.cstruct.types import (
     Void,
     Wchar,
 )
+from dissect.cstruct.types.base import MetaType
 
 if TYPE_CHECKING:
     from collections.abc import Iterable
@@ -54,7 +56,6 @@ class cstruct:
         self.endian = endian
 
         self.consts = {}
-        self.lookups = {}
         self.includes = []
         # fmt: off
         self.typedefs = {
@@ -206,6 +207,24 @@ class cstruct:
 
         raise AttributeError(f"Invalid attribute: {attr}")
 
+    def __copy__(self) -> cstruct:
+        cs = cstruct(endian=self.endian, pointer=self.pointer.__name__)
+        cs._anonymous_count = self._anonymous_count
+        cs.consts = self.consts.copy()
+        cs.includes = self.includes.copy()
+
+        # Update typedefs to point to the new cstruct instance
+        for name, type in self.typedefs.items():
+            if name in cs.typedefs:
+                continue
+
+            if isinstance(type, MetaType):
+                new_type = copy.copy(type)
+                new_type.cs = cs
+                cs.typedefs[name] = new_type
+
+        return cs
+
     def _next_anonymous(self) -> str:
         name = f"__anonymous_{self._anonymous_count}__"
         self._anonymous_count += 1
@@ -253,11 +272,7 @@ class cstruct:
         Definitions can be parsed using different parsers. Currently, there's
         only one supported parser - DEF_CSTYLE. Parsers can add types and
         modify this cstruct instance. Arguments can be passed to parsers
-        using kwargs.
-
-        The CSTYLE parser was recently replaced with token based parser,
-        instead of a strictly regex based one. The old parser is still available
-        by using DEF_LEGACY.
+        using ``kwargs``.
 
         Args:
             definition: The definition to parse.
@@ -267,9 +282,9 @@ class cstruct:
         deftype = deftype or cstruct.DEF_CSTYLE
 
         if deftype == cstruct.DEF_CSTYLE:
-            TokenParser(self, **kwargs).parse(definition)
-        elif deftype == cstruct.DEF_LEGACY:
             CStyleParser(self, **kwargs).parse(definition)
+        else:
+            raise ValueError(f"Unknown definition type: {deftype}")
 
         return self
 
@@ -328,6 +343,14 @@ class cstruct:
 
         raise ResolveError(f"Recursion limit exceeded while resolving type {name}")
 
+    def copy(self) -> cstruct:
+        """Create a copy of this cstruct instance.
+
+        Returns:
+            A new cstruct instance with the same types and settings as this one.
+        """
+        return copy.copy(self)
+
     def _make_type(
         self,
         name: str,

{dissect_cstruct-4.8.dev8 → dissect_cstruct-5.0.dev0}/dissect/cstruct/exceptions.py

@@ -5,6 +5,10 @@ class Error(Exception):
     pass
 
 
+class LexerError(Error):
+    pass
+
+
 class ParserError(Error):
     pass
 
@@ -23,7 +27,3 @@ class ArraySizeError(Error):
 
 class ExpressionParserError(Error):
     pass
-
-
-class ExpressionTokenizerError(Error):
-    pass

dissect_cstruct-5.0.dev0/dissect/cstruct/expression.py

@@ -0,0 +1,222 @@
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+from dissect.cstruct.exceptions import ExpressionParserError
+from dissect.cstruct.lexer import IDENTIFIER_TYPES, Lexer, TokenCursor, TokenType
+from dissect.cstruct.utils import offsetof, sizeof
+
+if TYPE_CHECKING:
+    from collections.abc import Callable
+
+    from dissect.cstruct import cstruct
+    from dissect.cstruct.lexer import Token
+
+BINARY_OPERATORS: dict[TokenType, Callable[[int, int], int]] = {
+    TokenType.PIPE: lambda a, b: a | b,
+    TokenType.CARET: lambda a, b: a ^ b,
+    TokenType.AMPERSAND: lambda a, b: a & b,
+    TokenType.LSHIFT: lambda a, b: a << b,
+    TokenType.RSHIFT: lambda a, b: a >> b,
+    TokenType.PLUS: lambda a, b: a + b,
+    TokenType.MINUS: lambda a, b: a - b,
+    TokenType.STAR: lambda a, b: a * b,
+    TokenType.SLASH: lambda a, b: a // b,
+    TokenType.PERCENT: lambda a, b: a % b,
+}
+
+UNARY_OPERATORS: dict[TokenType, Callable[[int], int]] = {
+    TokenType.UNARY_MINUS: lambda a: -a,
+    TokenType.TILDE: lambda a: ~a,
+}
+
+OPERATORS = set(BINARY_OPERATORS.keys()) | set(UNARY_OPERATORS.keys())
+
+FUNCTION_TOKENS = {
+    TokenType.SIZEOF: 1,
+    TokenType.OFFSETOF: 2,
+}
+
+PRECEDENCE_LEVELS = {
+    TokenType.PIPE: 0,
+    TokenType.CARET: 1,
+    TokenType.AMPERSAND: 2,
+    TokenType.LSHIFT: 3,
+    TokenType.RSHIFT: 3,
+    TokenType.PLUS: 4,
+    TokenType.MINUS: 4,
+    TokenType.STAR: 5,
+    TokenType.SLASH: 5,
+    TokenType.PERCENT: 5,
+    TokenType.UNARY_MINUS: 6,
+    TokenType.TILDE: 6,
+    # Functions
+    TokenType.SIZEOF: 7,
+    TokenType.OFFSETOF: 7,
+}
+
+
+def precedence(o1: TokenType, o2: TokenType) -> bool:
+    # Unary (prefix) operators are right-associative, so use strict greater-than
+    if o2 in UNARY_OPERATORS:
+        return PRECEDENCE_LEVELS[o1] > PRECEDENCE_LEVELS[o2]
+    return PRECEDENCE_LEVELS[o1] >= PRECEDENCE_LEVELS[o2]
+
+
+class Expression(TokenCursor):
+    """Expression parser for calculations in definitions."""
+
+    def __init__(self, expression: str):
+        self.expression = expression
+
+        tokens = Lexer(expression).tokenize()
+        super().__init__(tokens)
+        self._stack: list[TokenType] = []
+        self._queue: list[int | str] = []
+
+    def __repr__(self) -> str:
+        return self.expression
+
+    def _reset(self) -> None:
+        """Reset the expression state for a new input."""
+        self._reset_cursor()
+        self._stack = []
+        self._queue = []
+
+    def _error(self, msg: str, *, token: Token | None = None) -> ExpressionParserError:
+        return ExpressionParserError(f"line {(token if token is not None else self._current()).line}: {msg}")
+
+    def _evaluate_expression(self, cs: cstruct) -> None:
+        operator = self._stack.pop(-1)
+        result = 0
+
+        if operator in UNARY_OPERATORS:
+            if len(self._queue) < 1:
+                raise ExpressionParserError("Invalid expression: not enough operands")
+
+            result = UNARY_OPERATORS[operator](self._queue.pop(-1))
+        elif operator in BINARY_OPERATORS:
+            if len(self._queue) < 2:
+                raise ExpressionParserError("Invalid expression: not enough operands")
+
+            right = self._queue.pop(-1)
+            left = self._queue.pop(-1)
+            result = BINARY_OPERATORS[operator](left, right)
+        elif operator in FUNCTION_TOKENS:
+            num_args = FUNCTION_TOKENS[operator]
+            if len(self._queue) < num_args:
+                raise ExpressionParserError("Invalid expression: not enough operands")
+
+            args = [self._queue.pop(-1) for _ in range(num_args)][::-1]
+            if operator == TokenType.SIZEOF:
+                type_ = cs.resolve(args[0])
+                result = sizeof(type_)
+            elif operator == TokenType.OFFSETOF:
+                type_ = cs.resolve(args[0])
+                result = offsetof(type_, args[1])
+
+        self._queue.append(result)
+
+    def evaluate(self, cs: cstruct, context: dict[str, int] | None = None) -> int:
+        """Evaluates an expression using a Shunting-Yard implementation."""
+        self._reset()
+        context = context or {}
+
+        while (token := self._current()).type != TokenType.EOF:
+            if token.type == TokenType.NUMBER:
+                self._queue.append(int(self._take().value, 0))
+
+            elif token.type in OPERATORS:
+                while (
+                    len(self._stack) != 0
+                    and self._stack[-1] != TokenType.LPAREN
+                    and precedence(self._stack[-1], token.type)
+                ):
+                    self._evaluate_expression(cs)
+
+                self._stack.append(self._take().type)
+
+            elif token.type in FUNCTION_TOKENS:
+                func = self._take().type
+                self._stack.append(func)
+
+                self._expect(TokenType.LPAREN)
+
+                num_args = FUNCTION_TOKENS[func]
+                while num_args > 1:
+                    self._queue.append(self._collect_until(TokenType.COMMA))
+                    self._expect(TokenType.COMMA)
+                    num_args -= 1
+
+                self._queue.append(self._collect_until(TokenType.RPAREN))
+                self._expect(TokenType.RPAREN)
+
+                # Evaluate immediately
+                self._evaluate_expression(cs)
+
+            elif token.type in IDENTIFIER_TYPES:
+                ident = self._take().value
+
+                obj = None
+                for i, part in enumerate(ident.split(".")):
+                    if i == 0:
+                        if part in context:
+                            obj = context[part]
+                        elif part in cs.consts:
+                            obj = cs.consts[part]
+                        elif part in cs.typedefs:
+                            obj = cs.resolve(part)
+                        else:
+                            raise self._error(f"Unknown identifier: '{ident}'", token=token)
+                    else:
+                        if isinstance(obj, dict) and part in obj:
+                            obj = obj[part]
+                        elif hasattr(obj, part):
+                            obj = getattr(obj, part)
+                        else:
+                            raise self._error(f"Unknown identifier: '{ident}'", token=token)
+
+                try:
+                    self._queue.append(int(obj))
+                except (ValueError, TypeError):
+                    raise self._error(f"Identifier '{ident}' does not resolve to an integer", token=token)
+
+            elif token.type == TokenType.LPAREN:
+                if self._previous().type == TokenType.NUMBER:
+                    raise self._error(
+                        f"Parser expected sizeof or an arethmethic operator instead got: '{self._previous().value}'",
+                        token=self._previous(),
+                    )
+
+                self._stack.append(self._take().type)
+
+            elif token.type == TokenType.RPAREN:
+                if self._previous().type == TokenType.LPAREN:
+                    raise self._error(
+                        "Parser expected an expression, instead received empty parenthesis.",
+                        token=self._previous(),
+                    )
+
+                if len(self._stack) == 0:
+                    raise self._error("Mismatched parentheses")
+
+                while self._stack[-1] != TokenType.LPAREN:
+                    self._evaluate_expression(cs)
+                    if len(self._stack) == 0:
+                        raise self._error("Mismatched parentheses")
+
+                self._stack.pop(-1)  # Pop the '('
+                self._take()
+
+            else:
+                raise self._error(f"Unmatched token: '{token.value}'", token=token)
+
+        while len(self._stack) != 0:
+            if TokenType.LPAREN in self._stack:
+                raise self._error("Mismatched parentheses")
+            self._evaluate_expression(cs)
+
+        if len(self._queue) != 1:
+            raise self._error("Invalid expression: too many operands")
+
+        return self._queue[0]