diff --git a/asm/creole.py b/asm/creole.py index b1481dd..9fdd242 100644 --- a/asm/creole.py +++ b/asm/creole.py @@ -6,25 +6,52 @@ class MalformedArgument(Exception): pass def word_2c(w): - """ Negate a non-negative integer using two's compliment. """ + """ Negate a non-negative integer using 32 bit two's compliment. + + :param w: An integer in two's compliment. A non-negative Python + integer will work. + :return: The negation of the integer stored as a two's compliment + integer. + """ return (~w + 1) & 0xFFFFFFFF def ti(w): - """ Explicitly transform integer into two's compliment representation. """ + """ Explicitly transform integer into two's compliment representation. + + :param w: A python integer. + :return: The integer in two's compliment. + """ return w if w >= 0 else word_wc(-w) def from_2c(w): - """ Turn two's compliment word into Python integer. """ + """ Turn two's compliment word into Python integer. + + :param w: An integer in 32 bit twos compliment. + :return: The integer as a proper Python string. + """ if (w >> 31) & 1 == 0: return w return -word_2c(w) class Argument: + """ Class of arguments. Not used directly: It is used to store + intermediate information during the assembly process. """ + def __init__(self, argtype, val, sign=False): + """ Initialize an argument. + + :param argtype: Type of the argument (instance of ArgType). + :param val: Python integer value of the argument. + :param sign: If the argument should be treated as signed. + Otherwise, the integer will be interpreted in execution + as an unsigned integer. + """ self.at = argtype self.sign = sign self.val = val def __str__(self): return f'({self.at}, {self.sign}, {self.val})' def high_bits(self): + """ Returns the high bits that the argument would have + in the opcode. """ return int(self.sign) << 1 | (self.at == ArgType.REG) class ArgType(Enum): @@ -47,11 +74,21 @@ class ArgType(Enum): and returns a tuple with the first the first element being the type and the second element being the integer value of the argument. + + Valid parameters are: + + * `r` followed by a nonnegative integer (register) + * `l` followed by a nonnegative integer (label) + * any integer (immediate value) + + :param s: String representing the argument. + :return: The Argument object representing the argument. + :raises MalformedArgument: """ if s.isnumeric(): return Argument(ArgType.IMM, int(s)) elif s[0] == "-" and s[1:].isnumeric(): - return Argument(ArgType.IMM, word_2c(int(s[1:]))) + return Argument(ArgType.IMM, word_2c(int(s[1:])), True) elif s[0] == 'r' and s[1:].isnumeric(): return Argument(ArgType.REG, int(s[1:])) elif s[0] == 'l' and s[1:].isnumeric(): @@ -61,7 +98,11 @@ class ArgType(Enum): def typecheck(self, s): """ Parses the type of the string and returns it if it fits - the type of the enum value. """ + the type of the enum value. + + :param s: String argument representing an argument. + :return: The Argument class containing the object, or None + if the string does not fit the type of self. """ t = ArgType.gettype(s) if self == ArgType.VAL: if t.at == ArgType.REG or t.at == ArgType.IMM: @@ -76,60 +117,95 @@ class ArgType(Enum): class OpcodeException(Exception): pass class TypecheckLenException(Exception): + """ Exception thrown when arguments to an instruction are of the + incorrect length. """ def __init__(self, opcode, insargs, argtypelen): self.opcode = opcode self.insargs = insargs self.argtypelen = argtypelen def __str__(self): - return f'arguments {insargs} to opcode {self.opcode} not length {self.argtypelen}' + return f'''\ +arguments {insargs} to opcode {self.opcode} not of length {self.argtypelen}\ +''' class TypecheckException(Exception): + """ Exception thrown when an argument to an instruction are of the + incorrect type. """ def __init__(self, argtype, sarg, i, opcode): self.argtype = argtype self.sarg = sarg self.i = i self.opcode = opcode def __str__(self): - return f'opcode {self.opcode} has invalid value {self.sarg} (expected {self.argtype} in position {self.i}' + return f'''\ +opcode {self.opcode} has invalid value {self.sarg} +(expected {self.argtype} in position {self.i})\ +''' class Instruction(Enum): """ Class of microcode instructions. The first number is the opcode and the suceeding values are the types of each of the arguments. The first argument is the opcode and the second - argument is if the argument is a signed variant of another - opcode. """ - NOP = 0, False - PUSH = 1, False, ArgType.VAL - POP = 2, False, ArgType.REG - ADD = 3, False, ArgType.REG, ArgType.VAL, ArgType.VAL - MUL = 4, False, ArgType.REG, ArgType.VAL, ArgType.VAL - DIV = 5, False, ArgType.REG, ArgType.VAL, ArgType.VAL - SDIV = "DIV", True, ArgType.REG, ArgType.VAL, ArgType.VAL - SYS = 6, False, ArgType.VAL - CLB = 7, False, ArgType.LAB - JL = 8, False, ArgType.LAB, ArgType.VAL, ArgType.VAL - JLS = "JL", True, ArgType.LAB, ArgType.VAL, ArgType.VAL - JLE = 9, False, ArgType.LAB, ArgType.VAL, ArgType.VAL - JLES = "JLE", True, ArgType.LAB, ArgType.VAL, ArgType.VAL - JE = 10, False, ArgType.LAB, ArgType.VAL, ArgType.VAL - JNE = 11, False, ArgType.LAB, ArgType.VAL, ArgType.VAL + argument is what function is used to compile the instruction + (some instructions are actually versions of other instructions). """ + NOP = 0, "_default_render" + PUSH = 1, "_default_render", ArgType.VAL + POP = 2, "_default_render", ArgType.REG + ADD = 3, "_default_render", ArgType.REG, ArgType.VAL, ArgType.VAL + MUL = 4, "_default_render", ArgType.REG, ArgType.VAL, ArgType.VAL + DIV = 5, "_default_render", ArgType.REG, ArgType.VAL, ArgType.VAL + SDIV = "DIV", "_render_change_args", ArgType.REG, ArgType.VAL, ArgType.VAL + SYS = 6, "_default_render", ArgType.VAL + CLB = 7, "_default_render", ArgType.LAB + JL = 8, "_default_render", ArgType.LAB, ArgType.VAL, ArgType.VAL + JLS = "JL", "_render_change_args", ArgType.LAB, ArgType.VAL, ArgType.VAL + JLE = 9, "_default_render", ArgType.LAB, ArgType.VAL, ArgType.VAL + JLES = "JLE", "_render_change_args", ArgType.LAB, ArgType.VAL, ArgType.VAL + JE = 10, "_default_render", ArgType.LAB, ArgType.VAL, ArgType.VAL + JNE = 11, "_default_render", ArgType.LAB, ArgType.VAL, ArgType.VAL def __int__(self): - return self.opcode + """ Returns the opcode associated with the Instruction. + If it is a virtual instruction, it will resolve the string + name of the opcode and return its opcode. """ + if type(self.opcode) is int: + return self.opcode + return int(Instruction[self.opcode]) + + def __init__(self, opcode, renderfun, *args): + """ Initialize an Instruction. Do not call this function: it is + used to make enum values. To add a new instruction, modify + the Instruction enum. + This function sometimes takes string arguments because + certain values may not be loaded until later. + + :param opcode: Opcode of the instruction, or a string + containing the case-sensitive name of the instruction from + which this instruction derives from. + + :param renderfun: a string with the name of a function + in the class that returns the instruction opcode. + + :param *args: Type of each argument to the instruction. + The amount of arguments denotes the amount of instructions. + """ - def __init__(self, opcode, signed_instruction, *args): if type(opcode) is int and (opcode > 0x7F or opcode < 0): raise OpcodeException(opcode) self.opcode = opcode self.argtypes = args - if signed_instruction: - self.render = self._render_change_args - else: - self.render = self._default_render + self.render = getattr(self, renderfun) def typecheck(self, sargs): """ Pass arguments to the instruction and check if the - arguments are correct. """ + arguments are correct. + + :param sargs: List of arguments to the instruction + as strings. + :return: List of arguments (as Argument objects). + :raises TypeCheckException: + :raises TypecheckLenException: + """ rargs = [] if len(sargs) != len(self.argtypes): raise TypecheckLenException(self.opcode, sargs, @@ -143,9 +219,6 @@ class Instruction(Enum): rargs.append(t) return rargs - # The following will be called using OPCODE.render() instead of being - # called directly. - def _render_change_args(self, args): for i in range(0,len(args)): if args[i].at != ArgType.LAB: diff --git a/creole.c b/creole.c index fbde87e..30176c9 100644 --- a/creole.c +++ b/creole.c @@ -428,6 +428,11 @@ handle_compiletime_immediate(struct creole_env *env, } } +/* TODO: The compile step can be completely removed in favor of directly + * executing the bytecode, disassembling it with creole_parse_line() + * at every instance. This will also make the implementation simpler. + */ + enum creole_compiler_ret creole_compile(struct creole_env *env, struct creole_reader *r) {