def __init__(self, in_address):
self.tokenizer = Tokenizer(in_address)
self.curr_token = self.tokenizer.get_current_token()
self.out_address = in_address.replace(".jack", ".xml")
self.output = ""
self.indent = 0
self.compile_class()
def tokens_to_xml(path):
"""Write the tokens into a xml file with its type as tags. The outpath
is the dirpath of the a new directory in the module path to avoid name
clashes."""
paths = retrive_files(path)
out_dirpath = os.path.join(path, 'Xmlresult')
for path in paths:
outfile = os.path.basename(path).replace('.jack', 'T.xml')
outpath = os.path.join(out_dirpath, outfile)
tokenizer = Tokenizer(path)
analyzer = TokenAnalyzer(outpath)
while tokenizer.has_more_tokens():
tokenizer.advance()
t_type = tokenizer.token_type
tag = token_tags[t_type]
if t_type == T_KEYWORD:
analyzer.write_info(tokenizer.keyword, tag)
elif t_type == T_SYMBOL:
analyzer.write_info(tokenizer.symbol, tag)
elif t_type == T_ID:
analyzer.write_info(tokenizer.identifier, tag)
elif t_type == T_INTEGER:
analyzer.write_info(tokenizer.intval, tag)
elif t_type == T_STRING:
analyzer.write_info(tokenizer.stringval, tag)
analyzer.close()
def __init__(self, inpath, outpath):
self.tokenizer = Tokenizer(inpath)
self.symboltable = SymbolTable()
self.vmwriter = VMWriter(outpath)
self._class_name = None
if self.tokenizer.has_more_tokens():
self.compile_class()
self.vmwriter.close()
print("{0} completed.".format(outpath))
def __init__(self, in_address):
self.tokenizer = Tokenizer(in_address)
self.symbol_table = SymbolTable()
self.vm_writer = VMWriter(in_address.replace(".jack", ".vm"))
self.curr_token = self.tokenizer.get_current_token()
self.out_address = in_address.replace(".jack", ".xml")
self.output = ""
self.indent = 0
self.label_count = -1
self.class_name = ""
self.compile_class()
def __init__(self, inFile):
self.t = Tokenizer(inFile)
self.symTable = SymbolTable()
self.vmName = inFile.rstrip('.jack') + '.vm'
self.vm = VMWriter(self.vmName)
self.className = ''
self.types = ['int', 'char', 'boolean', 'void']
self.stmnt = ['do', 'let', 'if', 'while', 'return']
self.subroutType = ''
self.whileIndex = 0
self.ifIndex = 0
self.fieldNum = 0
def __init__(self, filepath):
self.jackfiles = []
self.fileOrDir(filepath)
for file in self.jackfiles:
print(file)
tokenizer = Tokenizer(file)
outputname = filepath.split('.')[0]+".0000.xml"
print(outputname)
print(tokenizer.file)
def SetClass(self, input_path, output_path):
self.tokenizer = Tokenizer(input_path)
self.output_file = open("{0}.xml".format(output_path), 'w')
self.code_file = open(output_path, 'w')
self.tokenizer.advance()
self.indent_level = 0
self.current_class_name = None
self.current_sub_name = None
def run_jack_compiler(file_path):
jack_files = []
if file_path.endswith(".jack"):
jack_files.append(file_path)
else:
file_list = listdir(file_path)
for file in file_list:
if file.endswith(".jack"):
jack_files.append(file_path + '/' + file)
for jack_file in jack_files:
# print(" Now processing: ", jack_file)
curr_tokenizer = Tokenizer(jack_file)
curr_compilation_engine = CompilationEngine(curr_tokenizer, jack_file)
curr_compilation_engine.compile_class()
def run_jack_tokenizer(file_path):
jack_files = []
if file_path.endswith(".jack"):
jack_files.append(file_path)
else:
file_list = listdir(file_path)
for file in file_list:
if file.endswith(".jack"):
jack_files.append(file_path + '/' + file)
for jack_file in jack_files:
print(" Now processing: ", jack_file)
curr_tokenizer = Tokenizer(jack_file)
print("writing into", jack_file[:-5]+"test.xml")
xml_stream = open(jack_file[:-5]+"test.xml", "w+")
xml_stream.write("<tokens>\n")
single_file_tokenize(curr_tokenizer, xml_stream)
xml_stream.write("</tokens>\n")
xml_stream.close()
class CompilationEngine:
_OPEN_PARENTHESIS = "\("
_CLOSE_PARENTHESIS = "\)"
_OPEN_BRACKET = "\["
_CLOSE_BRACKET = "\]"
_DOT = "\."
_OPS = "\+|-|\*|\/|&|\||<|>|="
def __init__(self, in_address):
self.tokenizer = Tokenizer(in_address)
self.curr_token = self.tokenizer.get_current_token()
self.out_address = in_address.replace(".jack", ".xml")
self.output = ""
self.indent = 0
self.compile_class()
def write_file(self):
with open(self.out_address, 'w') as f:
f.write(self.output)
def write(self, to_write):
"""
Writes to the output, with indentation.
:param to_write: The string to write
"""
self.output += (self.indent * " ") + to_write + "\n"
def compile_class(self):
"""
Compiles a complete class.
"""
def comp_class():
self.eat("class")
self.eat(NAME_REG)
self.eat("{")
self.compile_class_var_dec()
self.compile_subroutine()
self.eat("}")
self.wrap("class", comp_class)
def compile_class_var_dec(self):
"""
Compiles a static or field declaration.
:return:
"""
var_type_reg = "static|field"
if self.peek_token(var_type_reg):
self.wrap("classVarDec", self.__class_var_dec)
self.compile_class_var_dec()
def __class_var_dec(self):
"""
Compiles a single class var declaration.
"""
var_type_reg = "static|field"
# (static|field)
var_type = self.curr_token.get_token()
self.eat(var_type_reg)
# type
self.__compile_type(False)
# Compile varName combo until no more ","
self.__single_var()
self.eat(";")
def __single_var(self):
"""
Compiles a single set of variables separated by commas.
"""
# varName
self.eat(NAME_REG)
if self.peek_token(","):
self.eat(",")
self.__single_var()
def __compile_type(self, for_function):
"""
Compiles a type for a function or variable, determined by
a received boolean value.
:param for_function: True if is type of function, false otherwise.
:return:
"""
type_reg = r"int|char|boolean|[A-Za-z_]\w*"
if for_function:
type_reg += "|void"
self.eat(type_reg)
def compile_subroutine(self):
"""
Compiles a complete method, function or constructor.
:return:
"""
sub_regex = "(constructor|function|method)"
def subroutine_dec():
self.eat(sub_regex)
self.__compile_type(True)
# subroutine name
self.__compile_name()
self.eat(CompilationEngine._OPEN_PARENTHESIS)
self.compile_parameter_list()
self.eat(CompilationEngine._CLOSE_PARENTHESIS)
self.wrap("subroutineBody", subroutine_body)
def subroutine_body():
self.eat("{")
if self.peek_token("var"):
self.compile_var_dec()
self.compile_statements()
self.eat("}")
if self.peek_token(sub_regex):
self.wrap("subroutineDec", subroutine_dec)
# Handle next subroutine if there is one
self.compile_subroutine()
def __compile_name(self):
if self.peek_type() == IDENTIFIER:
self.eat(NAME_REG)
else:
print("ERROR: Identifier Expected")
exit(-1)
def compile_parameter_list(self):
"""
Compiles a possibly empty parameter list, not including the
enclosing ()
:return:
"""
self.wrap("parameterList", self.__params)
def __params(self):
type_reg = r"int|char|boolean|[A-Za-z_]\w*"
if self.peek_token(type_reg):
self.__compile_type(False)
self.eat(NAME_REG)
if self.peek_token(","):
self.eat(",")
self.__params()
def compile_var_dec(self):
"""
Compiles a var declaration.
:return:
"""
self.wrap("varDec", self.__comp_var_dec)
if self.peek_token("var"):
self.compile_var_dec()
def __comp_var_dec(self):
self.eat("var")
self.__compile_type(False)
self.__single_var()
self.eat(";")
def compile_statements(self):
"""
Compiles a sequence of statements, not including the enclosing {}
:return:
"""
def statement():
"""
Determines the type of statement and compiles it. Calls itself
afterwards to check for more statements.
:return:
"""
# statement_reg = "let|if|while|do|return"
# if self.peek_token(statement_reg):
if self.peek_token("let"):
self.compile_let()
statement()
if self.peek_token("if"):
self.compile_if()
statement()
if self.peek_token("while"):
self.compile_while()
statement()
if self.peek_token("do"):
self.compile_do()
statement()
if self.peek_token("return"):
self.compile_return()
statement()
self.wrap("statements", statement)
def compile_do(self):
"""
Compiles a do statement
:return:
"""
def do():
self.eat("do")
self.__subroutine_call()
self.eat(";")
self.wrap("doStatement", do)
def __comp_do(self):
self.eat("do")
self.__subroutine_call()
self.eat(";")
def compile_let(self):
"""
Compiles a let statement
:return:
"""
self.wrap("letStatement", self.__comp_let)
def __comp_let(self):
self.eat("let")
self.__compile_name()
# Determine [expression]
if self.peek_token(CompilationEngine._OPEN_BRACKET):
self.eat(CompilationEngine._OPEN_BRACKET)
self.compile_expression()
self.eat(CompilationEngine._CLOSE_BRACKET)
self.eat("=")
self.compile_expression()
self.eat(";")
def compile_while(self):
"""
Compiles a while statement.
:return:
"""
def comp_while():
self.eat("while")
self.eat(CompilationEngine._OPEN_PARENTHESIS)
self.compile_expression()
self.eat(CompilationEngine._CLOSE_PARENTHESIS)
self.eat("{")
self.compile_statements()
self.eat("}")
self.wrap("whileStatement", comp_while)
def compile_return(self):
"""
Compiles a return statement.
:return:
"""
def comp_return():
self.eat("return")
# if next is expression:
if self.is_term():
self.compile_expression()
self.eat(";")
self.wrap("returnStatement", comp_return)
def compile_if(self):
"""
Compiles an if statement, possibly with a trailing else clause.
:return:
"""
def comp_if():
self.eat("if")
self.eat(CompilationEngine._OPEN_PARENTHESIS)
self.compile_expression()
self.eat(CompilationEngine._CLOSE_PARENTHESIS)
self.eat("{")
# self.indent += 1
self.compile_statements()
self.eat("}")
# Handle else:
if self.peek_token("else"):
self.eat("else")
self.eat("{")
self.compile_statements()
self.eat("}")
self.wrap("ifStatement", comp_if)
def compile_expression(self):
"""
Compiles an expression.
:return:
"""
def comp_expression():
self.compile_term()
# Case: term op term
if self.peek_token(CompilationEngine._OPS):
self.eat(CompilationEngine._OPS)
self.compile_term()
self.wrap("expression", comp_expression)
def compile_term(self):
"""
Compiles a term.
:return:
"""
def term():
curr_type = self.peek_type()
is_const = curr_type == STRING_CONST or \
curr_type == INT_CONST or \
curr_type == KEYWORD
# Case: term is integerConstant or stringConstant or
# keywordConstant
if is_const:
self.write(self.tokenizer.get_current_token().get_xml_wrap())
self.__advance_token()
# Case: token is a varName or a subroutineName
elif curr_type == IDENTIFIER:
# self.write(self.tokenizer.get_current_token().get_xml_wrap())
# self.tokenizer.advance()
# Case: varName [ expression ]
if self.peek_next(CompilationEngine._OPEN_BRACKET):
self.__var_name_array()
# Case: subroutineCall:
elif self.peek_next(
CompilationEngine._OPEN_PARENTHESIS) or self.peek_next(
CompilationEngine._DOT):
self.__subroutine_call()
else:
self.eat(NAME_REG)
# Case: ( expression )
elif self.peek_token(CompilationEngine._OPEN_PARENTHESIS):
self.eat(CompilationEngine._OPEN_PARENTHESIS)
self.compile_expression()
self.eat(CompilationEngine._CLOSE_PARENTHESIS)
# Case: unaryOp term
elif self.peek_token("-|~"):
self.eat("-|~")
self.compile_term()
else:
print("Error: Incorrect Term")
exit(-1)
self.wrap("term", term)
def __var_name_array(self):
"""
Handles the case of varName[expression]
:return:
"""
self.eat(NAME_REG)
self.eat(CompilationEngine._OPEN_BRACKET)
self.compile_expression()
self.eat(CompilationEngine._CLOSE_BRACKET)
def is_term(self):
curr_type = self.peek_type()
return curr_type == STRING_CONST or curr_type == INT_CONST or \
curr_type == KEYWORD or curr_type == IDENTIFIER or \
self.peek_token(CompilationEngine._OPEN_PARENTHESIS) or \
self.peek_token(CompilationEngine._OPS)
def __subroutine_call(self):
if self.curr_token.get_type() == IDENTIFIER:
# self.write(self.curr_token.get_xml_wrap())
# self.__advance_token()
if self.peek_next(CompilationEngine._OPEN_PARENTHESIS):
self.__subroutine_name()
elif self.peek_next(CompilationEngine._DOT):
self.__object_subroutine_call()
else:
print("Error: ( or . expected")
exit(-1)
def __object_subroutine_call(self):
self.eat(NAME_REG)
self.eat(CompilationEngine._DOT)
self.__subroutine_name()
def __subroutine_name(self):
"""
Handles the case of subroutineName(expressionList)
:return:
"""
if self.curr_token.get_type() == IDENTIFIER:
self.eat(NAME_REG)
self.eat(CompilationEngine._OPEN_PARENTHESIS)
self.compile_expression_list()
self.eat(CompilationEngine._CLOSE_PARENTHESIS)
def compile_expression_list(self):
"""
Compiles a possibly empty list of comma separated expressions
:return:
"""
def exp_list():
if self.is_term():
self.compile_expression()
while self.peek_token(","):
self.eat(",")
self.compile_expression()
self.wrap("expressionList", exp_list)
def wrap(self, section_name, func):
"""
Wraps a program structure block with the section_name, and executes
its function
:param section_name: The name of the section
:param func: The function to perform
:return:
"""
self.write("<{}>".format(section_name))
self.indent += 2
func()
self.indent -= 2
self.write("</{}>".format(section_name))
def eat(self, token):
"""
Handles advancing and writing terminal tokens.
Will exit the program if an error occurs.
:param token: The regex of the token to compare
:return:
"""
if re.match(token, self.curr_token.get_token()):
self.write(self.curr_token.get_xml_wrap())
self.__advance_token()
else:
# if self.tokenizer.get_current_token() != token:
print("Error: Expected " + token)
exit(-1)
def peek_token(self, compare_next):
"""
:param compare_next: The regex to compare.
:return: True if the current token matches the regex, False otherwise.
"""
if self.curr_token:
return re.match(compare_next, self.curr_token.get_token())
return False
def peek_type(self):
"""
:return: the type of the current token
"""
return self.curr_token.get_type()
def peek_next(self, comp):
next_token = self.tokenizer.get_next_token()
# Case: There actually is a next token
if next_token:
return re.match(comp, self.tokenizer.get_next_token().get_token())
return False
def __advance_token(self):
self.tokenizer.advance()
if self.tokenizer.has_more_tokens():
self.curr_token = self.tokenizer.get_current_token()
class CompilationEngine:
_OPEN_PARENTHESIS = "\("
_CLOSE_PARENTHESIS = "\)"
_OPEN_BRACKET = "\["
_CLOSE_BRACKET = "\]"
_DOT = "\."
_OPS = "\+|-|\*|\/|&|\||<|>|="
def __init__(self, in_address):
self.tokenizer = Tokenizer(in_address)
self.symbol_table = SymbolTable()
self.vm_writer = VMWriter(in_address.replace(".jack", ".vm"))
self.curr_token = self.tokenizer.get_current_token()
self.out_address = in_address.replace(".jack", ".xml")
self.output = ""
self.indent = 0
self.label_count = -1
self.class_name = ""
self.compile_class()
def write_file(self):
# with open(self.out_address, 'w') as f:
# f.write(self.output)
self.vm_writer.write_file()
def write(self, to_write):
"""
Writes to the output, with indentation.
:param to_write: The string to write
"""
self.output += (self.indent * " ") + to_write + "\n"
# ========== Compilation Methods ========== #
def compile_class(self):
"""
Compiles a complete class.
"""
def comp_class():
self.eat("class")
self.class_name = self.eat(NAME_REG)
self.eat("{")
self.compile_class_var_dec()
self.compile_subroutine()
self.eat("}")
self.wrap("class", comp_class)
def compile_class_var_dec(self):
"""
Compiles a static or field declaration.
:return:
"""
var_type_reg = "static|field"
if self.peek_token(var_type_reg):
self.wrap("classVarDec", self.__class_var_dec)
self.compile_class_var_dec()
def compile_subroutine(self):
"""
Compiles a complete method, function or constructor.
:return:
"""
sub_regex = "(constructor|function|method)"
self.symbol_table.start_subroutine()
kind = self.eat(sub_regex)
self.__compile_type(True)
# subroutine name
name = self.__compile_name()
self.eat(CompilationEngine._OPEN_PARENTHESIS)
self.compile_parameter_list(kind)
self.eat(CompilationEngine._CLOSE_PARENTHESIS)
self.eat("{")
if self.peek_token("var"):
self.compile_var_dec()
num_locals = self.symbol_table.var_count("local")
self.vm_writer.write_function("{}.{}".format(self.class_name, name),
num_locals)
self.__set_pointer(kind)
self.compile_statements()
self.eat("}")
# def subroutine_dec():
# kind = self.eat(sub_regex)
# self.__compile_type(True)
# # subroutine name
# name = self.__compile_name()
# self.eat(CompilationEngine._OPEN_PARENTHESIS)
# self.compile_parameter_list(kind)
# self.eat(CompilationEngine._CLOSE_PARENTHESIS)
# subroutine_body(name)
# # self.wrap("subroutineBody", subroutine_body)
#
# def subroutine_body(name):
# self.eat("{")
# num_locals = 0
# if self.peek_token("var"):
# num_locals = self.compile_var_dec()
# self.vm_writer.write_function("{}.{}".format(self.class_name,
# name), num_locals)
#
# self.compile_statements()
# # if sub_type == "void":
# # self.vm_writer.write_push("constant", 0)
# self.eat("}")
# Handle next subroutine if there is one
if self.peek_token(sub_regex):
self.compile_subroutine()
def compile_parameter_list(self, kind):
"""
Compiles a possibly empty parameter list, not including the
enclosing ()
:return:
"""
if kind == "method":
self.symbol_table.define("this", self.class_name, "argument")
type_reg = r"int|char|boolean|[A-Za-z_]\w*"
while self.peek_token(type_reg):
self.__params()
def compile_var_dec(self):
"""
Compiles a var declaration.
:return:
"""
# self.wrap("varDec", self.__comp_var_dec)
self.eat("var")
var_type = self.__compile_type(False)
self.__var_declare(var_type, "var")
self.eat(";")
if self.peek_token("var"):
self.compile_var_dec()
def compile_statements(self):
"""
Compiles a sequence of statements, not including the enclosing {}
:return:
"""
statement_reg = "let|if|while|do|return"
if self.peek_token(statement_reg):
if self.peek_token("let"):
self.compile_let()
elif self.peek_token("if"):
self.compile_if()
elif self.peek_token("while"):
self.compile_while()
elif self.peek_token("do"):
self.compile_do()
elif self.peek_token("return"):
self.compile_return()
self.compile_statements()
def compile_do(self):
"""
Compiles a do statement
"""
self.eat("do")
self.__subroutine_call()
# Since we don't use the return value, we pop it to temp
self.vm_writer.write_pop("temp", 0)
self.eat(";")
def compile_let(self):
"""
Compiles a let statement
"""
self.eat("let")
name = self.__compile_name()
is_array = False
# Determine [expression]
if self.peek_token(CompilationEngine._OPEN_BRACKET):
is_array = True
self.__handle_array(name)
self.eat("=")
self.compile_expression()
# Pop the value to the spot in the memory
if is_array:
self.vm_writer.write_pop("temp", 0)
self.vm_writer.write_pop("pointer", 1)
self.vm_writer.write_push("temp", 0)
self.vm_writer.write_pop("that", 0)
else:
self.__write_pop(name)
self.eat(";")
def compile_while(self):
"""
Compiles a while statement.
:return:
"""
self.eat("while")
loop_label = self.__get_label("WHILE_START")
exit_label = self.__get_label("WHILE_END")
self.vm_writer.write_label(loop_label)
self.eat(CompilationEngine._OPEN_PARENTHESIS)
# Compute ~condition
self.compile_expression()
self.vm_writer.write_arithmetic("~")
# if ~condition exit loop
self.vm_writer.write_if(exit_label)
self.eat(CompilationEngine._CLOSE_PARENTHESIS)
self.eat("{")
self.compile_statements()
self.vm_writer.write_goto(loop_label)
self.vm_writer.write_label(exit_label)
self.eat("}")
def compile_return(self):
"""
Compiles a return statement.
"""
self.eat("return")
# if next is expression:
if self.__is_term():
self.compile_expression()
else:
# Void function - push 0
self.vm_writer.write_push(CONSTANT, 0)
self.vm_writer.write_return()
self.eat(";")
def compile_if(self):
"""
Compiles an if statement, possibly with a trailing else clause.
:return:
"""
self.eat("if")
self.eat(CompilationEngine._OPEN_PARENTHESIS)
# ~cond
self.compile_expression()
# self.vm_writer.write_arithmetic("~")
self.eat(CompilationEngine._CLOSE_PARENTHESIS)
self.eat("{")
if_true = self.__get_label("IF_TRUE")
self.vm_writer.write_if(if_true)
if_false = self.__get_label("IF_FALSE")
self.vm_writer.write_goto(if_false)
self.vm_writer.write_label(if_true)
self.compile_statements()
self.eat("}")
# Handle else:
if self.peek_token("else"):
if_end = self.__get_label("IF_END")
self.vm_writer.write_goto(if_end)
self.vm_writer.write_label(if_false)
self.eat("else")
self.eat("{")
self.compile_statements()
self.eat("}")
self.vm_writer.write_label(if_end)
else:
self.vm_writer.write_label(if_false)
def compile_expression(self):
"""
Compiles an expression.
:return:
"""
def comp_expression():
self.compile_term()
# Case: term op term
if self.peek_token(CompilationEngine._OPS):
operation = self.eat(CompilationEngine._OPS)
self.compile_term()
self.vm_writer.write_arithmetic(operation)
self.wrap("expression", comp_expression)
def compile_term(self):
"""
Compiles a term.
:return:
"""
def term():
curr_type = self.peek_type()
val = self.curr_token.get_token()
# Handle integer constant
if curr_type == INT_CONST:
self.vm_writer.write_push(CONSTANT, int(val))
self.__advance_token()
# Handle String constant
elif curr_type == STRING_CONST:
self.__handle_string_constant(val)
self.__advance_token()
# Handle Keyword constant
elif curr_type == KEYWORD:
self.__handle_keyword_constant(val)
self.__advance_token()
# Case: token is a varName or a subroutineName
elif curr_type == IDENTIFIER:
self.__handle_identifier()
# Case: ( expression )
elif self.peek_token(CompilationEngine._OPEN_PARENTHESIS):
self.eat(CompilationEngine._OPEN_PARENTHESIS)
self.compile_expression()
self.eat(CompilationEngine._CLOSE_PARENTHESIS)
# Case: unaryOp term
elif self.peek_token("-|~"):
self.__handle_unary_op()
else:
print("Error: Incorrect Term")
exit(-1)
term()
# self.wrap("term", term)
def compile_expression_list(self):
"""
Compiles a possibly empty list of comma separated expressions
:return:
"""
def exp_list():
count = 0
if self.__is_term():
self.compile_expression()
count += 1
while self.peek_token(","):
self.eat(",")
self.compile_expression()
count += 1
return count
return exp_list()
# self.wrap("expressionList", exp_list)
# ========== Compilation Helper ========== #
def __class_var_dec(self):
"""
Compiles a single class var declaration.
"""
var_type_reg = "static|field"
# (static|field)
kind = self.eat(var_type_reg)
# type
var_type = self.__compile_type(False)
# Compile varName combo until no more ","
self.__var_declare(var_type, kind)
self.eat(";")
def __var_declare(self, var_type, kind):
name = self.eat(NAME_REG)
self.symbol_table.define(name, var_type, kind)
if self.peek_token(","):
self.eat(",")
self.__var_declare(var_type, kind)
def __compile_type(self, for_function):
"""
Compiles a type for a function or variable, determined by
a received boolean value.
:param for_function: True if is type of function, false otherwise.
:return:
"""
type_reg = r"int|char|boolean|[A-Za-z_]\w*"
if for_function:
type_reg += "|void"
return self.eat(type_reg)
def __set_pointer(self, kind):
if kind == "method":
self.vm_writer.write_push("argument", 0)
self.vm_writer.write_pop("pointer", 0)
elif kind == "constructor":
self.__handle_constructor()
def __handle_constructor(self):
# Allocate memory for the new object
var_num = self.symbol_table.var_count("this")
self.vm_writer.write_push(CONSTANT, var_num)
self.vm_writer.write_call("Memory.alloc", 1)
# Set the new memory spot to this
self.vm_writer.write_pop("pointer", 0)
def __compile_name(self):
if self.peek_type() == IDENTIFIER:
return self.eat(NAME_REG)
else:
print("ERROR: Identifier Expected")
exit(-1)
def __params(self):
var_type = self.__compile_type(False)
name = self.eat(NAME_REG)
self.symbol_table.define(name, var_type, "argument")
if self.peek_token(","):
self.eat(",")
def __handle_unary_op(self):
command = self.eat("-|~")
self.compile_term()
if command == "-":
self.vm_writer.write_arithmetic("neg")
else:
self.vm_writer.write_arithmetic(command)
def __handle_identifier(self):
"""
Handles the case of an identifier given as a term
"""
# Case: varName [ expression ]
if self.peek_next(CompilationEngine._OPEN_BRACKET):
name = self.__compile_name()
self.__handle_array(name)
self.vm_writer.write_pop("pointer", 1)
self.vm_writer.write_push("that", 0)
# self.__var_name_array()
# Case: subroutineCall:
elif self.peek_next(CompilationEngine._OPEN_PARENTHESIS) or \
self.peek_next(CompilationEngine._DOT):
self.__subroutine_call()
else:
name = self.eat(NAME_REG)
self.__write_push(name)
def __handle_string_constant(self, string):
"""
Handles the case of a string constant in a term
:param string: the constant
"""
self.vm_writer.write_push(CONSTANT, len(string))
self.vm_writer.write_call("String.new", 1)
for char in string:
self.vm_writer.write_push(CONSTANT, ord(char))
self.vm_writer.write_call("String.appendChar", 2)
def __handle_keyword_constant(self, word):
"""
Handles the case of a keyword constant given in a term.
If the word is not valid the program prints a relevant message and
exits.
:param word: The keyword
"""
if word == "this":
self.vm_writer.write_push("pointer", 0)
else:
self.vm_writer.write_push(CONSTANT, 0)
if word == "true":
self.vm_writer.write_arithmetic("~")
def __is_term(self):
curr_type = self.peek_type()
return curr_type == STRING_CONST or curr_type == INT_CONST or \
curr_type == KEYWORD or curr_type == IDENTIFIER or \
self.peek_token(CompilationEngine._OPEN_PARENTHESIS) or \
self.peek_token(CompilationEngine._OPS)
def __subroutine_call(self):
if self.curr_token.get_type() == IDENTIFIER:
if self.peek_next(CompilationEngine._OPEN_PARENTHESIS):
self.vm_writer.write_push("pointer", 0)
self.__subroutine_name(self.class_name, 1)
elif self.peek_next(CompilationEngine._DOT):
self.__object_subroutine_call()
else:
print("Error: ( or . expected")
exit(-1)
def __object_subroutine_call(self):
name = self.eat(NAME_REG)
n_args = 0
# Push the object reference to the stack
if self.symbol_table.kind_of(name):
self.__write_push(name)
name = self.symbol_table.type_of(name)
n_args = 1
self.eat(CompilationEngine._DOT)
self.__subroutine_name(name, n_args)
def __subroutine_name(self, type_name, n_args):
"""
Handles the case of subroutineName(expressionList)
:return:
"""
name = self.eat(NAME_REG)
self.eat(CompilationEngine._OPEN_PARENTHESIS)
nargs = self.compile_expression_list()
self.eat(CompilationEngine._CLOSE_PARENTHESIS)
self.vm_writer.write_call("{}.{}".format(type_name, name),
nargs + n_args)
def __handle_array(self, name):
self.eat(CompilationEngine._OPEN_BRACKET)
self.compile_expression()
self.eat(CompilationEngine._CLOSE_BRACKET)
self.__write_push(name)
self.vm_writer.write_arithmetic("+")
# ========== XML Handling ========== #
def wrap(self, section_name, func):
"""
Wraps a program structure block with the section_name, and executes
its function
:param section_name: The name of the section
:param func: The function to perform
:return:
"""
self.write("<{}>".format(section_name))
self.indent += 2
func()
self.indent -= 2
self.write("</{}>".format(section_name))
# ========== Token Handling ========== #
def eat(self, token):
"""
Handles advancing and writing terminal tokens.
Will exit the program if an error occurs.
:param token: The regex of the token to compare
:return:
"""
ctoken = self.curr_token.get_token()
if re.match(token, self.curr_token.get_token()):
# self.write(self.curr_token.get_xml_wrap())
self.__advance_token()
return ctoken
# else:
# # if self.tokenizer.get_current_token() != token:
# print("Error: Expected " + token)
# exit(-1)
def peek_token(self, compare_next):
"""
:param compare_next: The regex to compare.
:return: True if the current token matches the regex, False otherwise.
"""
if self.curr_token:
return re.match(compare_next, self.curr_token.get_token())
return False
def peek_type(self):
"""
:return: the type of the current token
"""
return self.curr_token.get_type()
def peek_next(self, comp):
next_token = self.tokenizer.get_next_token()
# Case: There actually is a next token
if next_token:
return re.match(comp, self.tokenizer.get_next_token().get_token())
return False
def __advance_token(self):
self.tokenizer.advance()
if self.tokenizer.has_more_tokens():
self.curr_token = self.tokenizer.get_current_token()
# ========== VM Helper ========== #
def __get_label(self, label):
self.label_count += 1
return "{}{}".format(label, str(self.label_count))
def __write_pop(self, name):
self.vm_writer.write_pop(self.symbol_table.kind_of(name),
self.symbol_table.index_of(name))
def __write_push(self, name):
self.vm_writer.write_push(self.symbol_table.kind_of(name),
self.symbol_table.index_of(name))
# Developed for project 10 of nand2tetris course
from JackParser import Parser
from JackTokenizer import Tokenizer
from pathlib import Path
import sys
if __name__ == '__main__':
input_arg = sys.argv[1]
input_path = Path(input_arg).resolve()
files = input_path.glob('*.jack')
for file in files:
print(file)
output_path = str(file.with_name(file.stem + '.xml'))
tokenizer = Tokenizer(input_file=str(file)).tokenize()
parser = Parser(tokenizer)
parser.parse_to_file(output_path)
class CompilationEngine:
def __init__(self):
self.local_symbol_table = None
self.class_symbol_tables = {}
self.type_size_map = {"int": 1, "bool": 1, "char": 1}
self.unique_label_index = 0
def SetClass(self, input_path, output_path):
self.tokenizer = Tokenizer(input_path)
self.output_file = open("{0}.xml".format(output_path), 'w')
self.code_file = open(output_path, 'w')
self.tokenizer.advance()
self.indent_level = 0
self.current_class_name = None
self.current_sub_name = None
def CompileClass(self):
"""
Compiles a complete class.
"""
self.EnterScope("class")
self.ConsumeKeyword([Keyword.CLASS])
self.ConsumeDeclaration("class", None)
self.class_symbol_tables[self.current_class_name] = SymbolTable()
self.ConsumeSymbol('{')
totalSize = 0
while (self.IsKeyword([Keyword.STATIC, Keyword.FIELD])):
totalSize += self.CompileClassVarDec()
self.type_size_map[self.current_class_name] = totalSize
# subroutineDec*
while (self.IsKeyword(subroutine_types)):
self.CompileSubroutine()
self.ConsumeSymbol('}')
self.ExitScope("class")
self.output_file.close()
def CompileClassVarDec(self):
"""
Compiles a static declaration or a field declaration.
"""
self.EnterScope("classVarDec")
amount = 0
category = self.tokenizer.keyword()
self.ConsumeKeyword([Keyword.STATIC, Keyword.FIELD])
varType = self.ConsumeType()
self.ConsumeDeclaration(category, varType)
amount += 1
while (self.IsSymbol([','])):
self.ConsumeSymbol(',')
self.ConsumeDeclaration(category, varType)
amount += 1
self.ConsumeSymbol(';')
self.ExitScope("classVarDec")
return amount
def CompileSubroutine(self):
"""
Compiles a complete method, function, or constructor.
"""
self.EnterScope("subroutineDec")
self.local_symbol_table = SymbolTable()
subType = self.tokenizer.keyword()
self.ConsumeKeyword([Keyword.CONSTRUCTOR, Keyword.FUNCTION,
Keyword.METHOD])
if (self.IsKeyword([Keyword.VOID])):
self.ConsumeKeyword([Keyword.VOID])
else:
self.ConsumeType()
# The first param is converted to internal rep. the second is preserved
self.ConsumeDeclaration(subType, subType)
if subType == "method":
self.local_symbol_table.indexList[Categories.ARGUMENT[0]] += 1
self.ConsumeSymbol('(')
self.CompileParameterList()
self.ConsumeSymbol(')')
self.CompileSubroutineBody()
self.ExitScope("subroutineDec")
def CompileSubroutineBody(self):
self.EnterScope("subroutineBody")
nVars = 0
self.ConsumeSymbol('{')
while (self.IsKeyword([Keyword.VAR])):
nVars += self.CompileVarDec()
self.WriteCode("function {0}.{1} {2}".format(self.current_class_name,
self.current_sub_name,
str(nVars)))
entry = self.SymbolTableLookup(self.current_sub_name)
if entry.type == "constructor":
self.WriteCode("push constant {0}".
format(self.type_size_map[self.current_class_name]))
self.WriteCode("call Memory.alloc 1")
self.WriteCode("pop pointer 0")
elif entry.type == "method":
self.WriteCode("push argument 0")
self.WriteCode("pop pointer 0")
self.CompileStatements()
self.ConsumeSymbol('}')
self.ExitScope("subroutineBody")
def ConsumeDeclaration(self, category, entry_type):
entry = SymbolTableEntry()
entry.SetCategory(category)
entry.name = self.ConsumeIdentifier()
entry.type = entry_type
local_categories = [Categories.VAR, Categories.ARGUMENT]
class_categories = [Categories.SUBROUTINE, Categories.FIELD,
Categories.STATIC]
# Updating current class / subroutine names
if entry.category == Categories.CLASS:
self.current_class_name = entry.name
elif entry.category == Categories.SUBROUTINE:
self.current_sub_name = entry.name
# Updating local / class symbol tables
if entry.category in local_categories:
self.local_symbol_table.InsertEntry(entry)
elif entry.category in class_categories:
self.class_symbol_tables[self.current_class_name].\
InsertEntry(entry)
def CompileParameterList(self):
"""
Compiles a (possibly empty) parameter list,
not including the enclosing "()".
"""
self.EnterScope("parameterList")
nVars = 0
if (not self.IsSymbol([')'])):
varType = self.ConsumeType()
self.ConsumeDeclaration("argument", varType)
nVars += 1
while(self.IsSymbol([','])):
self.ConsumeSymbol(',')
varType = self.ConsumeType()
self.ConsumeDeclaration("argument", varType)
nVars += 1
self.ExitScope("parameterList")
return nVars
def CompileVarDec(self):
"""
Compiles a var declaration.
"""
self.EnterScope("varDec")
nVars = 0
self.ConsumeKeyword([Keyword.VAR])
varType = self.ConsumeType()
self.ConsumeDeclaration("var", varType)
nVars += 1
while (self.IsSymbol([','])):
self.ConsumeSymbol(',')
self.ConsumeDeclaration("var", varType)
nVars += 1
self.ConsumeSymbol(';')
self.ExitScope("varDec")
return nVars
def CompileStatements(self):
"""
Compiles a sequence of statements, not including the
enclosing "{}".
"""
self.EnterScope("statements")
while self.IsKeyword([Keyword.LET, Keyword.IF, Keyword.WHILE,
Keyword.DO, Keyword.RETURN]):
if self.IsKeyword([Keyword.LET]):
self.CompileLet()
if self.IsKeyword([Keyword.IF]):
self.CompileIf()
if self.IsKeyword([Keyword.WHILE]):
self.CompileWhile()
if self.IsKeyword([Keyword.DO]):
self.CompileDo()
if self.IsKeyword([Keyword.RETURN]):
self.CompileReturn()
self.ExitScope("statements")
def CompileDo(self):
"""
Compiles a do statement.
"""
self.EnterScope("doStatement")
self.ConsumeKeyword([Keyword.DO])
prefix = self.ConsumeIdentifier()
calleeLocation = None
subName = None
if self.IsSymbol(['.']):
self.ConsumeSymbol('.')
entry = self.SymbolTableLookup(prefix)
if entry is not None and entry.category != Categories.CLASS:
calleeLocation = "{0} {1}".format(entry.segment, entry.index)
prefix = entry.type
postfix = self.ConsumeIdentifier()
subName = "{0}.{1}".format(prefix, postfix)
else:
subName = "{0}.{1}".format(self.current_class_name, prefix)
calleeLocation = "pointer 0"
nArgs = 0
# This means we are calling an instance method, so we push it first
if calleeLocation is not None:
self.WriteCode("push {0} //Pushing callee".format(calleeLocation))
nArgs += 1
self.ConsumeSymbol('(')
nArgs += self.CompileExpressionList()
self.ConsumeSymbol(')')
self.ConsumeSymbol(';')
self.WriteCode("call {0} {1}".format(subName, nArgs))
# Get rid of the return value (garbage)
self.WriteCode("pop temp 0")
self.ExitScope("doStatement")
def CompileLet(self):
"""
Compiles a let statement.
"""
self.EnterScope("letStatement")
self.ConsumeKeyword([Keyword.LET])
varName = self.ConsumeIdentifier()
entry = self.SymbolTableLookup(varName)
isArray = False
if self.IsSymbol(['[']):
isArray = True
self.ConsumeSymbol('[')
self.CompileExpression()
self.WriteCode("push {0} {1}".
format(entry.segment, entry.index)) # array base
self.WriteCode("add") # Add offset
self.ConsumeSymbol(']')
self.ConsumeSymbol('=')
self.CompileExpression()
self.ConsumeSymbol(';')
if isArray:
self.WriteCode("pop temp 0") # Save the expression result
self.WriteCode("pop pointer 1") # Align THAT
self.WriteCode("push temp 0") # Push the exp result
# Put the exp result in the array position
self.WriteCode("pop that 0")
else:
self.WriteCode("pop {0} {1}".format(entry.segment, entry.index))
self.ExitScope("letStatement")
def CompileWhile(self):
"""
Compiles a while statement.
"""
self.EnterScope("whileStatement")
self.ConsumeKeyword([Keyword.WHILE])
L1 = self.GenerateUniqueLabel()
L2 = self.GenerateUniqueLabel()
# While entry point
self.WriteCode("label {0}".format(L1))
# while loop condition
self.ConsumeSymbol('(')
self.CompileExpression()
self.ConsumeSymbol(')')
# Jump to L2 if condition doesn't hold
self.WriteCode("not")
self.WriteCode("if-goto {0}".format(L2))
# While loop logic
self.ConsumeSymbol('{')
self.CompileStatements()
self.ConsumeSymbol('}')
# Go back to L1 for another iteration
self.WriteCode("goto {0}".format(L1))
# While termination point
self.WriteCode("label {0}".format(L2))
self.ExitScope("whileStatement")
def CompileReturn(self):
"""
Compiles a return statement.
"""
self.EnterScope("returnStatement")
self.ConsumeKeyword([Keyword.RETURN])
if not self.IsSymbol([';']):
self.CompileExpression()
else:
self.WriteCode("push constant 0")
self.ConsumeSymbol(';')
self.WriteCode("return")
self.ExitScope("returnStatement")
def CompileIf(self):
"""
Compiles an if statement, possibly with a trailing
else clause.
"""
self.EnterScope("ifStatement")
self.ConsumeKeyword([Keyword.IF])
IF_TRUE = self.GenerateUniqueLabel()
IF_FALSE = self.GenerateUniqueLabel()
IF_END = self.GenerateUniqueLabel()
# The if statement condition
self.ConsumeSymbol('(')
self.CompileExpression()
self.ConsumeSymbol(')')
# Jump to L1 if condition doesn't hold
self.WriteCode("if-goto {0}".format(IF_TRUE))
self.WriteCode("goto {0}".format(IF_FALSE))
self.WriteCode("label {0}".format(IF_TRUE))
self.ConsumeSymbol('{')
self.CompileStatements()
self.ConsumeSymbol('}')
self.WriteCode("goto {0}".format(IF_END))
self.WriteCode("label {0}".format(IF_FALSE))
if self.IsKeyword([Keyword.ELSE]):
self.ConsumeKeyword([Keyword.ELSE])
self.ConsumeSymbol('{')
self.CompileStatements()
self.ConsumeSymbol('}')
self.WriteCode("label {0}".format(IF_END))
self.ExitScope("ifStatement")
def CompileExpression(self):
"""
Compiles an expression.
"""
self.EnterScope("expression")
self.CompileTerm()
while (self.IsSymbol(op_symbols.keys())):
op = self.ConsumeSymbol(self.tokenizer.symbol())
self.CompileTerm()
self.WriteCode(op_symbols[op])
self.ExitScope("expression")
def CompileTerm(self):
"""
Compiles a term.
"""
self.EnterScope("term")
keyword_constants = [Keyword.TRUE, Keyword.FALSE, Keyword.NULL,
Keyword.THIS]
termName = None
if self.IsType(TokenType.INT_CONST):
self.WriteCode("push constant {0}".
format(self.ConsumeIntegerConstant()))
elif self.IsType(TokenType.STRING_CONST):
self.ConsumeStringConstant()
elif self.IsKeyword(keyword_constants):
keyword = self.ConsumeKeyword(keyword_constants)
if keyword == "false":
self.WriteCode("push constant 0")
elif keyword == "true":
self.WriteCode("push constant 0")
self.WriteCode("not")
elif keyword == "this":
self.WriteCode("push pointer 0")
elif keyword == "null":
self.WriteCode("push constant 0")
elif self.IsSymbol(['(']):
self.ConsumeSymbol('(')
self.CompileExpression()
self.ConsumeSymbol(')')
elif self.IsSymbol(unary_symbols.keys()):
symbol = self.ConsumeSymbol(self.tokenizer.symbol())
self.CompileTerm()
self.WriteCode(unary_symbols[symbol])
else:
termName = self.ConsumeIdentifier()
entry = self.SymbolTableLookup(termName)
if entry is not None:
if CategoryUtils.IsIndexed(entry.category):
self.WriteCode("push {0} {1} //{2}".
format(CategoryUtils.
GetSegment(entry.category),
entry.index, termName))
if self.IsSymbol(['[']): # varName '[' expression ']'
self.ConsumeSymbol('[')
self.CompileExpression()
self.WriteCode("add")
self.WriteCode("pop pointer 1")
self.WriteCode("push that 0")
self.ConsumeSymbol(']')
elif self.IsSymbol(['(']): # subroutineCall
self.ConsumeSymbol('(')
self.WriteCode("call {0} {1}".
format(termName, self.CompileExpressionList()))
self.ConsumeSymbol(')')
elif self.IsSymbol(['.']):
self.ConsumeSymbol('.')
funcName = self.ConsumeIdentifier()
entry = self.GetSubroutineEntry(termName, funcName)
extraParam = 0
if entry is not None and entry.type == "method":
termName = self.SymbolTableLookup(termName).type
extraParam = 1
self.ConsumeSymbol('(')
self.WriteCode("call {0}.{1} {2}".
format(termName, funcName,
self.CompileExpressionList() +
extraParam))
self.ConsumeSymbol(')')
self.ExitScope("term")
return termName
def GetSubroutineEntry(self, prefix, postfix):
entry = self.SymbolTableLookup(postfix)
if entry is not None:
return entry
varEntry = self.SymbolTableLookup(prefix)
if varEntry is not None:
return self.ClassSymbolTableLookup(postfix, varEntry.type)
return None
def CompileExpressionList(self):
"""
Compiles a (possibly empty) comma-separated
list of expressions.
"""
self.EnterScope("expressionList")
nArgs = 0
if not self.IsSymbol(')'):
self.CompileExpression()
nArgs += 1
while self.IsSymbol([',']):
self.ConsumeSymbol(',')
self.CompileExpression()
nArgs += 1
self.ExitScope("expressionList")
return nArgs
def IsKeyword(self, keyword_list):
return (self.IsType(TokenType.KEYWORD) and
self.tokenizer.keyword() in keyword_list)
def IsSymbol(self, symbol_list):
return (self.IsType(TokenType.SYMBOL) and
self.tokenizer.symbol() in symbol_list)
def IsType(self, tokenType):
return self.tokenizer.tokenType() == tokenType
def ConsumeType(self):
if (self.tokenizer.tokenType() == TokenType.IDENTIFIER):
return self.ConsumeIdentifier()
else:
return self.ConsumeKeyword([Keyword.INT, Keyword.CHAR,
Keyword.BOOLEAN])
def ConsumeKeyword(self, keywordList):
self.VerifyTokenType(TokenType.KEYWORD)
actual = self.tokenizer.keyword()
if actual not in keywordList:
raise Exception("Expected keywords: {}, Actual: {}".
format(keywordList, actual))
self.OutputTag("keyword", actual)
if self.tokenizer.hasMoreTokens():
self.tokenizer.advance()
return actual
def ConsumeSymbol(self, symbol):
self.VerifyTokenType(TokenType.SYMBOL)
actual = self.tokenizer.symbol()
if actual != symbol:
raise Exception("Expected symbol: {}, Actual: {}".
format(symbol, actual))
self.OutputTag("symbol", actual)
if self.tokenizer.hasMoreTokens():
self.tokenizer.advance()
return actual
def ConsumeIntegerConstant(self):
self.VerifyTokenType(TokenType.INT_CONST)
actual = self.tokenizer.intVal()
self.OutputTag("integerConstant", self.tokenizer.intVal())
if self.tokenizer.hasMoreTokens():
self.tokenizer.advance()
return actual
def ConsumeStringConstant(self):
self.VerifyTokenType(TokenType.STRING_CONST)
actual = self.tokenizer.stringVal()
self.WriteCode("push constant {0}".format(len(actual)))
self.WriteCode("call String.new 1")
for c in actual:
self.WriteCode("push constant {0}".format(ord(c)))
self.WriteCode("call String.appendChar 2")
self.OutputTag("stringConstant", self.tokenizer.stringVal())
if self.tokenizer.hasMoreTokens():
self.tokenizer.advance()
return actual
def ConsumeIdentifier(self):
self.VerifyTokenType(TokenType.IDENTIFIER)
actual = self.tokenizer.identifier()
self.OutputTag("identifierName", self.tokenizer.identifier())
if self.tokenizer.hasMoreTokens():
self.tokenizer.advance()
return actual
def VerifyTokenType(self, tokenType):
actual = self.tokenizer.tokenType()
if actual != tokenType:
raise Exception("Expected token type: {}, Actual: {}".
format(tokenType, actual))
def EnterScope(self, name):
self.Output("<{}>".format(name))
self.indent_level += 1
def ExitScope(self, name):
self.indent_level -= 1
self.Output("</{}>".format(name))
def ClassSymbolTableLookup(self, name, containingClass):
return self.class_symbol_tables[containingClass].GetEntry(name)
def SymbolTableLookup(self, name):
entry = self.local_symbol_table.GetEntry(name)
if entry is not None:
return entry
else:
return self.ClassSymbolTableLookup(name, self.current_class_name)
def WriteCode(self, line):
self.code_file.write(line + '\n')
def OutputTag(self, tag, value):
self.Output("<{}> {} </{}>".format(tag, value, tag))
def Output(self, text):
self.output_file.write((" " * self.indent_level) + text + '\n')
def GenerateUniqueLabel(self):
self.unique_label_index += 1
return "pfl{0}".format(self.unique_label_index - 1)
class CompilationEngine:
def __init__(self, inpath, outpath):
self.tokenizer = Tokenizer(inpath)
self.symboltable = SymbolTable()
self.vmwriter = VMWriter(outpath)
self._class_name = None
if self.tokenizer.has_more_tokens():
self.compile_class()
self.vmwriter.close()
print("{0} completed.".format(outpath))
def _subroutine_init(self):
self._sub_kind = None
self._sub_name = None
self._ret_type = None
def _advance(self):
self._check_EOF()
self.tokenizer.advance()
@property
def _current_token(self):
t_type = self.tokenizer.token_type
return (self.tokenizer.keyword if t_type == T_KEYWORD else
self.tokenizer.symbol if t_type == T_SYMBOL else
self.tokenizer.identifier if t_type == T_ID else self.tokenizer
.intval if t_type == T_INTEGER else self.tokenizer.stringval)
@property
def _current_tok_type(self):
return self.tokenizer.token_type
@property
def _current_tok_tag(self):
return token_tags[self._current_tok_type]
@property
def _next_token(self):
"""return raw next_token in the tokenizer"""
return str(self.tokenizer.next_token)
def _require_token(self, tok_type, token=None):
"""Check whether the next_token(terminal) in the tokenizer meets the
requirement (specific token or just token type). If meets, tokenizer
advances (update current_token and next_token) and terminal will be
writed into outfile; If not, report an error."""
self._advance()
if token and self._current_token != token:
return self._error(expect_toks=(token, ))
elif self._current_tok_type != tok_type:
return self._error(expect_types=(tok_type, ))
def _require_id(self):
self._require_token(T_ID)
def _require_kw(self, token):
return self._require_token(T_KEYWORD, token=token)
def _require_sym(self, token):
return self._require_token(T_SYMBOL, token=token)
def _require_brackets(self, brackets, procedure):
front, back = brackets
self._require_sym(front)
procedure()
self._require_sym(back)
def _fol_by_class_vardec(self):
return self._next_token in (KW_STATIC, KW_FIELD)
def _fol_by_subroutine(self):
return self._next_token in (KW_CONSTRUCTOR, KW_FUNCTION, KW_METHOD)
def _fol_by_vardec(self):
return self._next_token == KW_VAR
#########################
# structure compilation #
#########################
def compile_class_name(self):
self._require_id()
self._class_name = self._current_token
def compile_subroutine_name(self):
self._require_id()
self._sub_name = self._current_token
def compile_var_name(self, kind=None, type=None, declare=False):
self._require_id()
name = self._current_token
if declare is True: # kind and type are not None
self.symboltable.define(name, type, kind)
else:
self.check_var_name(name, type)
def check_var_name(self, name, type=None):
recorded_kind = self.symboltable.kindof(name)
if recorded_kind is None:
self._traceback('name used before declared: {0}'.format(name))
elif type is not None:
recorded_type = self.symboltable.typeof(name)
if recorded_type != type:
get = '{0} "{1}"'.format(recorded_type, name)
self._error(expect_types=(type, ), get=get)
def compile_type(self, advanced=False, expect='type'):
# int, string, boolean or identifier(className)
if advanced is False:
self._advance()
if (self._current_token not in SymbolTable.builtIn_types
and self._current_tok_type != T_ID):
return self._error(expect=expect)
def compile_return_type(self):
# void or type
self._advance()
if self._current_token != KW_VOID:
self.compile_type(True, '"void" or type')
self._ret_type = self._current_token
if self._sub_kind == KW_CONSTRUCTOR and self._ret_type != self._class_name:
me = 'constructor expect current class as return type'
self._traceback(me)
@record_non_terminal('class')
def compile_class(self):
# 'class' className '{' classVarDec* subroutineDec* '}'
self._require_kw(KW_CLASS)
self.compile_class_name()
self._require_sym('{')
while self._fol_by_class_vardec():
self.compile_class_vardec()
while self._fol_by_subroutine():
self.compile_subroutine()
self._advance()
if self._current_token != '}':
self._traceback("Except classVarDec first, subroutineDec second.")
if self.tokenizer.has_more_tokens():
if self._next_token == KW_CLASS:
self._traceback('Only expect one classDec.')
self._traceback('Unexpected extra tokens.')
def compile_declare(self):
self._advance()
id_kind = self._current_token # ('static | field | var')
# type varName (',' varName)* ';'
self.compile_type()
id_type = self._current_token
self.compile_var_name(id_kind, id_type, declare=True)
# compile ',' or ';'
self._advance()
while self._current_token == ',':
self.compile_var_name(id_kind, id_type, declare=True)
self._advance()
if self._current_token != ';':
return self._error((',', ';'))
@record_non_terminal('classVarDec')
def compile_class_vardec(self):
# ('static|field') type varName (',' varName)* ';'
self.compile_declare()
@record_non_terminal('subroutineDec')
def compile_subroutine(self):
# ('constructor'|'function'|'method')
# ('void'|type) subroutineName '(' parameterList ')' subroutineBody
self._subroutine_init()
self.symboltable.start_subroutine()
self._advance()
self._sub_kind = self._current_token
if self._sub_kind == KW_METHOD:
self.symboltable.define('this', self._class_name, 'argument')
self.compile_return_type()
self.compile_subroutine_name()
self._require_brackets('()', self.compile_parameter_list)
self.compile_subroutine_body()
@record_non_terminal('parameterList')
def compile_parameter_list(self):
# ((type varName) (',' type varName)*)?
if self._next_token == ')':
return
self.compile_type()
self.compile_var_name('argument', self._current_token, True)
while self._next_token != ')':
self._require_sym(',')
self.compile_type()
self.compile_var_name('argument', self._current_token, True)
@record_non_terminal('subroutineBody')
def compile_subroutine_body(self):
# '{' varDec* statements '}'
self._require_sym('{')
while self._fol_by_vardec():
self.compile_vardec()
self.compile_function()
self.compile_statements()
self._require_sym('}')
def compile_function(self):
fn_name = '.'.join((self._class_name, self._sub_name))
num_locals = self.symboltable.varcount(KW_VAR)
self.vmwriter.write_function(fn_name,
num_locals) # function fn_name num_locals
# set up pointer this
if self._sub_kind == KW_CONSTRUCTOR:
num_fields = self.symboltable.varcount(KW_FIELD)
self.vmwriter.write_push('constant', num_fields)
self.vmwriter.write_call('Memory.alloc', 1)
self.vmwriter.write_pop('pointer', 0)
elif self._sub_kind == KW_METHOD:
self.vmwriter.write_push('argument', 0)
self.vmwriter.write_pop('pointer', 0)
@record_non_terminal('varDec')
def compile_vardec(self):
# 'var' type varName (',' varName)* ';'
self.compile_declare()
#########################
# statement compilation #
#########################
@record_non_terminal('statements')
def compile_statements(self):
# (letStatement | ifStatement | whileStatement | doStatement |
# returnStatement)*
last_statement = None
while self._next_token != '}':
self._advance()
last_statement = self._current_token
if last_statement == 'do':
self.compile_do()
elif last_statement == 'let':
self.compile_let()
elif last_statement == 'while':
self.compile_while()
elif last_statement == 'return':
self.compile_return()
elif last_statement == 'if':
self.compile_if()
else:
return self._error(expect='statement expression')
#if STACK[-2] == 'subroutineBody' and last_statement != 'return':
# self._error(expect='return statement', get=last_statement)
@record_non_terminal('doStatement')
def compile_do(self):
# 'do' subroutineCall ';'
self._advance()
self.compile_subroutine_call()
self.vmwriter.write_pop('temp', 0) # temp[0] store useless value
self._require_sym(';')
@record_non_terminal('letStatement')
def compile_let(self):
# 'let' varName ('[' expression ']')? '=' expression ';'
self.compile_var_name()
var_name = self._current_token
array = (self._next_token == '[')
if array:
self.compile_array_subscript(
var_name) # push (array base + subscript)
self._require_sym('=')
self.compile_expression() # push expression value
self._require_sym(';')
if array:
self.vmwriter.write_pop('temp', 1) # pop exp value to temp[1]
self.vmwriter.write_pop('pointer',
1) # that = array base + subscript
self.vmwriter.write_push('temp', 1)
self.vmwriter.write_pop('that', 0)
else:
self.assign_variable(var_name)
kind_segment = {
'static': 'static',
'field': 'this',
'argument': 'argument',
'var': 'local'
}
def assign_variable(self, name):
kind = self.symboltable.kindof(name)
index = self.symboltable.indexof(name)
self.vmwriter.write_pop(self.kind_segment[kind], index)
def load_variable(self, name):
kind = self.symboltable.kindof(name)
index = self.symboltable.indexof(name)
self.vmwriter.write_push(self.kind_segment[kind], index)
label_num = 0
@record_non_terminal('whileStatement')
def compile_while(self):
# 'while' '(' expression ')' '{' statements '}'
start_label = 'WHILE_START_' + str(self.label_num)
end_label = 'WHILE_END_' + str(self.label_num)
self.label_num += 1
self.vmwriter.write_label(start_label)
self.compile_cond_expression(start_label, end_label)
@record_non_terminal('ifStatement')
def compile_if(self):
# 'if' '(' expression ')' '{' statements '}'
# ('else' '{' statements '}')?
else_label = 'IF_ELSE_' + str(self.label_num)
end_label = 'IF_END_' + str(self.label_num)
self.label_num += 1
self.compile_cond_expression(end_label, else_label)
# else clause
if self._next_token == KW_ELSE:
self._require_kw(KW_ELSE)
self._require_brackets('{}', self.compile_statements)
self.vmwriter.write_label(end_label)
def compile_cond_expression(self, goto_label, end_label):
self._require_brackets('()', self.compile_expression)
self.vmwriter.write_arithmetic('not')
self.vmwriter.write_if(end_label)
self._require_brackets('{}', self.compile_statements)
self.vmwriter.write_goto(goto_label) # meet
self.vmwriter.write_label(end_label)
@record_non_terminal('returnStatement')
def compile_return(self):
# 'return' expression? ';'
if self._sub_kind == KW_CONSTRUCTOR:
self._require_kw(KW_THIS) # constructor must return 'this'
self.vmwriter.write_push('pointer', 0)
elif self._next_token != ';':
self.compile_expression()
else:
if self._ret_type != KW_VOID:
self._traceback('expect return ' + self._ret_type)
self.vmwriter.write_push('constant', 0)
self._require_sym(';')
self.vmwriter.write_return()
##########################
# expression compilation #
##########################
unary_ops = {'-': 'neg', '~': 'not'}
binary_ops = {
'+': 'add',
'-': 'sub',
'*': None,
'/': None,
'&': 'and',
'|': 'or',
'<': 'lt',
'>': 'gt',
'=': 'eq'
}
@record_non_terminal('expression')
def compile_expression(self):
# term (op term)*
self.compile_term()
while self._next_token in self.binary_ops:
self._advance()
if self._current_tok_type != T_SYMBOL:
self._error(expect_types=(T_SYMBOL, ))
op = self._current_token
self.compile_term()
self.compile_binaryop(op)
def compile_binaryop(self, op):
if op == '*':
self.vmwriter.write_call('Math.multiply', 2)
elif op == '/':
self.vmwriter.write_call('Math.divide', 2)
else:
self.vmwriter.write_arithmetic(self.binary_ops[op])
kw_consts = (KW_TRUE, KW_FALSE, KW_NULL, KW_THIS)
@record_non_terminal('term')
def compile_term(self):
# integerConstant | stringConstant | keywordConstant |
# varName | varName '[' expression ']' | subroutineCall |
# '(' expression ')' | unaryOp term
if self._next_token == '(':
self._require_brackets('()', self.compile_expression)
else:
self._advance()
tok = self._current_token
tok_type = self._current_tok_type
if tok_type == T_KEYWORD and tok in self.kw_consts:
self.compile_kw_consts(tok)
elif tok_type == T_INTEGER:
self.vmwriter.write_push('constant', tok)
elif tok_type == T_STRING:
self.compile_string(tok)
elif tok_type == T_ID:
if self._next_token in '(.':
self.compile_subroutine_call()
elif self._next_token == '[':
self.check_var_name(tok)
self.compile_array_subscript(tok)
self.vmwriter.write_pop('pointer', 1)
self.vmwriter.write_push('that', 0)
else:
self.check_var_name(tok)
self.load_variable(tok)
elif tok_type == T_SYMBOL and tok in self.unary_ops:
self.compile_term()
self.vmwriter.write_arithmetic(self.unary_ops[tok])
else:
self._error(expect='term')
# keywordConstant: 'true' | 'false' | 'null' | 'this'
def compile_kw_consts(self, kw):
if kw == KW_THIS:
self.vmwriter.write_push('pointer', 0)
elif kw == KW_TRUE:
self.vmwriter.write_push('constant', 1)
self.vmwriter.write_arithmetic('neg')
else:
self.vmwriter.write_push('constant', 0)
def compile_string(self, string):
self.vmwriter.write_push('constant', len(string))
self.vmwriter.write_call('String.new', 1)
for char in string:
self.vmwriter.write_push('constant', ord(char))
self.vmwriter.write_call('String.appendChar', 2)
def compile_subroutine_call(self):
# subroutineName '(' expressionList ')' |
# (className | varName) '.' subroutineName '(' expressionList ')'
## the first element of structure has already been compiled.
fn_name, num_args = self.compile_call_name()
self._require_sym('(')
num_args = self.compile_expressionlist(num_args)
self._require_sym(')')
self.vmwriter.write_call(fn_name, num_args)
def compile_call_name(self):
# the fisrt name of subroutine call could be (className or varName) if
# it is followed by '.', or subroutineName if followed by '('.
# return name of function call and num_args (1: means pushing this, 0:
# means don't)
if self._current_tok_type != T_ID:
self._error(expect_types=(T_ID, ))
name = self._current_token
if self._next_token == '.':
self._require_sym('.')
self.compile_subroutine_name()
sub_name = self._current_token
if (name in self.symboltable.all_class_types()
or name in SymbolTable.builtIn_class
or name == self._class_name):
return '.'.join((name, sub_name)), 0 # className
else:
self.check_var_name(name) # varName with class type
type = self.symboltable.typeof(name)
if type in SymbolTable.builtIn_types:
return self._error(expect='class instance or class',
get=type)
self.load_variable(name)
return '.'.join((type, sub_name)), 1
elif self._next_token == '(':
self.vmwriter.write_push('pointer', 0) # push this to be 1st arg
return '.'.join((self._class_name, name)), 1 # subroutineName
@record_non_terminal('expressionList')
def compile_expressionlist(self, num_args):
# (expression (',' expression)*)?
if self._next_token != ')':
self.compile_expression()
num_args += 1
while self._next_token != ')':
self._require_sym(',')
self.compile_expression()
num_args += 1
return num_args
def compile_array_subscript(self, var_name):
# varName '[' expression ']'
self.check_var_name(var_name, 'Array')
self._require_brackets(
'[]', self.compile_expression) # push expression value
self.load_variable(var_name)
self.vmwriter.write_arithmetic('add') # base + subscript
def _check_EOF(self):
if not self.tokenizer.has_more_tokens():
self._traceback("Unexpected EOF.")
def _error(self, expect_toks=(), expect_types=(), expect=None, get=None):
if expect is None:
exp_tok = ' or '.join(('"{0}"'.format(t) for t in expect_toks))
exp_types = ('type {0}'.format(token_tags[t])
for t in expect_types)
exp_type = ' or '.join(exp_types)
if exp_tok and exp_type:
expect = ' or '.join(exp_tok, exp_type)
else:
expect = exp_tok + exp_type
if get is None:
get = self._current_token
me = 'Expect {0} but get "{1}"'.format(expect, get)
return self._traceback(me)
def _traceback(self, message):
if DEBUG:
print('--------------------------------------------')
print(self.symboltable)
print(self.symboltable.all_class_types())
print('--------------------------------------------')
file_info = 'file: "{0}"'.format(self.tokenizer.filename)
line_info = 'line {0}'.format(self.tokenizer.line_count)
raise CompileError("{0}, {1}: {2}".format(file_info, line_info,
message))
def __init__(self, inputPath, outputPath):
self.tokenizer = Tokenizer(inputPath)
self.outputFile = open(outputPath, 'w')
self.tokenizer.advance()
self.indentLevel = 0
class CompilationEngine:
def __init__(self, inputPath, outputPath):
self.tokenizer = Tokenizer(inputPath)
self.outputFile = open(outputPath, 'w')
self.tokenizer.advance()
self.indentLevel = 0
def CompileClass(self):
"""
Compiles a complete class.
"""
self.EnterScope("class")
self.ConsumeKeyword([Keyword.CLASS])
self.ConsumeIdentifier() # className
self.ConsumeSymbol('{')
while (self.IsKeyword([Keyword.STATIC, Keyword.FIELD])):
self.CompileClassVarDec()
# subroutineDec*
while (self.IsKeyword([Keyword.CONSTRUCTOR, Keyword.FUNCTION,
Keyword.METHOD])):
self.CompileSubroutine()
self.ConsumeSymbol('}')
self.ExitScope("class")
self.outputFile.close()
def CompileClassVarDec(self):
"""
Compiles a static declaration or a field declaration.
"""
self.EnterScope("classVarDec")
self.ConsumeKeyword([Keyword.STATIC, Keyword.FIELD])
self.ConsumeType()
self.ConsumeIdentifier() # varName
while (self.IsSymbol([','])):
self.ConsumeSymbol(',')
self.ConsumeIdentifier() # varName
self.ConsumeSymbol(';')
self.ExitScope("classVarDec")
def CompileSubroutine(self):
"""
Compiles a complete method, function, or constructor.
"""
self.EnterScope("subroutineDec")
self.ConsumeKeyword([Keyword.CONSTRUCTOR, Keyword.FUNCTION,
Keyword.METHOD])
if (self.IsKeyword([Keyword.VOID])):
self.ConsumeKeyword([Keyword.VOID])
else:
self.ConsumeType()
self.ConsumeIdentifier() # subroutineName
self.ConsumeSymbol('(')
self.CompileParameterList()
self.ConsumeSymbol(')')
self.CompileSubroutineBody()
self.ExitScope("subroutineDec")
def CompileSubroutineBody(self):
self.EnterScope("subroutineBody")
self.ConsumeSymbol('{')
while (self.IsKeyword([Keyword.VAR])):
self.CompileVarDec()
self.CompileStatements()
self.ConsumeSymbol('}')
self.ExitScope("subroutineBody")
def CompileParameterList(self):
"""
Compiles a (possibly empty) parameter list,
not including the enclosing "()".
"""
self.EnterScope("parameterList")
if (not self.IsSymbol([')'])):
self.ConsumeType()
self.ConsumeIdentifier()
while(self.IsSymbol([','])):
self.ConsumeSymbol(',')
self.ConsumeType()
self.ConsumeIdentifier()
self.ExitScope("parameterList")
def CompileVarDec(self):
"""
Compiles a var declaration.
"""
self.EnterScope("varDec")
self.ConsumeKeyword([Keyword.VAR])
self.ConsumeType()
self.ConsumeIdentifier() # varName
while (self.IsSymbol([','])):
self.ConsumeSymbol(',')
self.ConsumeIdentifier() # varName
self.ConsumeSymbol(';')
self.ExitScope("varDec")
def CompileStatements(self):
"""
Compiles a sequence of statements, not including the
enclosing "{}".
"""
self.EnterScope("statements")
while self.IsKeyword([Keyword.LET, Keyword.IF, Keyword.WHILE,
Keyword.DO, Keyword.RETURN]):
if self.IsKeyword([Keyword.LET]):
self.CompileLet()
if self.IsKeyword([Keyword.IF]):
self.CompileIf()
if self.IsKeyword([Keyword.WHILE]):
self.CompileWhile()
if self.IsKeyword([Keyword.DO]):
self.CompileDo()
if self.IsKeyword([Keyword.RETURN]):
self.CompileReturn()
self.ExitScope("statements")
def CompileDo(self):
"""
Compiles a do statement.
"""
self.EnterScope("doStatement")
self.ConsumeKeyword([Keyword.DO])
self.ConsumeIdentifier()
if self.IsSymbol(['.']):
self.ConsumeSymbol('.')
self.ConsumeIdentifier()
self.ConsumeSymbol('(')
self.CompileExpressionList()
self.ConsumeSymbol(')')
self.ConsumeSymbol(';')
self.ExitScope("doStatement")
def CompileLet(self):
"""
Compiles a let statement.
"""
self.EnterScope("letStatement")
self.ConsumeKeyword([Keyword.LET])
self.ConsumeIdentifier()
if self.IsSymbol(['[']):
self.ConsumeSymbol('[')
self.CompileExpression()
self.ConsumeSymbol(']')
self.ConsumeSymbol('=')
self.CompileExpression()
self.ConsumeSymbol(';')
self.ExitScope("letStatement")
def CompileWhile(self):
"""
Compiles a while statement.
"""
self.EnterScope("whileStatement")
self.ConsumeKeyword([Keyword.WHILE])
self.ConsumeSymbol('(')
self.CompileExpression()
self.ConsumeSymbol(')')
self.ConsumeSymbol('{')
self.CompileStatements()
self.ConsumeSymbol('}')
self.ExitScope("whileStatement")
def CompileReturn(self):
"""
Compiles a return statement.
"""
self.EnterScope("returnStatement")
self.ConsumeKeyword([Keyword.RETURN])
if not self.IsSymbol([';']):
self.CompileExpression()
self.ConsumeSymbol(';')
self.ExitScope("returnStatement")
def CompileIf(self):
"""
Compiles an if statement, possibly with a trailing
else clause.
"""
self.EnterScope("ifStatement")
self.ConsumeKeyword([Keyword.IF])
self.ConsumeSymbol('(')
self.CompileExpression()
self.ConsumeSymbol(')')
self.ConsumeSymbol('{')
self.CompileStatements()
self.ConsumeSymbol('}')
if self.IsKeyword([Keyword.ELSE]):
self.ConsumeKeyword([Keyword.ELSE])
self.ConsumeSymbol('{')
self.CompileStatements()
self.ConsumeSymbol('}')
self.ExitScope("ifStatement")
def CompileExpression(self):
"""
Compiles an expression.
"""
self.EnterScope("expression")
op_symbols = ['+', '-', '*', '/', '&', '|', "<", ">", '=']
self.CompileTerm()
while (self.IsSymbol(op_symbols)):
self.ConsumeSymbol(self.tokenizer.symbol())
self.CompileTerm()
self.ExitScope("expression")
def CompileTerm(self):
"""
Compiles a term.
"""
self.EnterScope("term")
keyword_constants = [Keyword.TRUE, Keyword.FALSE, Keyword.NULL,
Keyword.THIS]
unary_symbols = ['-', '~']
if self.IsType(TokenType.INT_CONST):
self.ConsumeIntegerConstant()
elif self.IsType(TokenType.STRING_CONST):
self.ConsumeStringConstant()
elif self.IsKeyword(keyword_constants):
self.ConsumeKeyword(keyword_constants)
elif self.IsSymbol(['(']):
self.ConsumeSymbol('(')
self.CompileExpression()
self.ConsumeSymbol(')')
elif self.IsSymbol(unary_symbols):
self.ConsumeSymbol(self.tokenizer.symbol())
self.CompileTerm()
else:
self.ConsumeIdentifier()
if self.IsSymbol(['[']): # varName '[' expression ']'
self.ConsumeSymbol('[')
self.CompileExpression()
self.ConsumeSymbol(']')
elif self.IsSymbol(['(']): # subroutineCall
self.ConsumeSymbol('(')
self.CompileExpressionList()
self.ConsumeSymbol(')')
elif self.IsSymbol(['.']):
self.ConsumeSymbol('.')
self.ConsumeIdentifier()
self.ConsumeSymbol('(')
self.CompileExpressionList()
self.ConsumeSymbol(')')
self.ExitScope("term")
def CompileExpressionList(self):
"""
Compiles a (possibly empty) comma-separated
list of expressions.
"""
self.EnterScope("expressionList")
if not self.IsSymbol(')'):
self.CompileExpression()
while self.IsSymbol([',']):
self.ConsumeSymbol(',')
self.CompileExpression()
self.ExitScope("expressionList")
def IsKeyword(self, keyword_list):
return (self.IsType(TokenType.KEYWORD) and
self.tokenizer.keyword() in keyword_list)
def IsSymbol(self, symbol_list):
return (self.IsType(TokenType.SYMBOL) and
self.tokenizer.symbol() in symbol_list)
def IsType(self, tokenType):
return self.tokenizer.tokenType() == tokenType
def ConsumeType(self):
if (self.tokenizer.tokenType() == TokenType.IDENTIFIER):
self.ConsumeIdentifier()
else:
self.ConsumeKeyword([Keyword.INT, Keyword.CHAR, Keyword.BOOLEAN])
def ConsumeKeyword(self, keywordList):
self.VerifyTokenType(TokenType.KEYWORD)
actual = self.tokenizer.keyword()
if actual not in keywordList:
raise Exception("Expected keywords: {}, Actual: {}".
format(keywordList, actual))
self.OutputTag("keyword", actual)
if self.tokenizer.hasMoreTokens():
self.tokenizer.advance()
def ConsumeSymbol(self, symbol):
self.VerifyTokenType(TokenType.SYMBOL)
actual = self.tokenizer.symbol()
if actual != symbol:
raise Exception("Expected symbol: {}, Actual: {}".
format(symbol, actual))
self.OutputTag("symbol", actual)
if self.tokenizer.hasMoreTokens():
self.tokenizer.advance()
def ConsumeIntegerConstant(self):
self.VerifyTokenType(TokenType.INT_CONST)
self.OutputTag("integerConstant", self.tokenizer.intVal())
if self.tokenizer.hasMoreTokens():
self.tokenizer.advance()
def ConsumeStringConstant(self):
self.VerifyTokenType(TokenType.STRING_CONST)
self.OutputTag("stringConstant", self.tokenizer.stringVal())
if self.tokenizer.hasMoreTokens():
self.tokenizer.advance()
def ConsumeIdentifier(self):
self.VerifyTokenType(TokenType.IDENTIFIER)
self.OutputTag("identifier", self.tokenizer.identifier())
if self.tokenizer.hasMoreTokens():
self.tokenizer.advance()
def VerifyTokenType(self, tokenType):
actual = self.tokenizer.tokenType()
if actual != tokenType:
raise Exception("Expected token type: {}, Actual: {}".
format(tokenType, actual))
def EnterScope(self, name):
self.Output("<{}>".format(name))
self.indentLevel += 1
def ExitScope(self, name):
self.indentLevel -= 1
self.Output("</{}>".format(name))
def OutputTag(self, tag, value):
self.Output("<{}> {} </{}>".format(tag, value, tag))
def Output(self, text):
self.outputFile.write((" " * self.indentLevel) + text + '\n')
token.text = token.text.strip()
value, type = token.text, token.tag
return value, type
def _format_string(self, s):
if s[0] == '<':
if s[1] == '/':
return s
else:
return '\n' + s
elif s[0] == '>':
return s
else:
return ' {} '.format(s)
def _dump_tree(self, filename='/Users/rsenseman/Desktop/dump.xml'):
with open(filename, 'w') as f:
f.write(tostring(self.tree.getroot(), encoding='utf8').decode('utf8'))
return None
if __name__ == '__main__':
input_arg = sys.argv[1]
input_path = Path(input_arg).resolve()
output_path = input_path.with_name(input_path.stem + '_parsed.xml')
tokenizer = Tokenizer(input_file=input_path).tokenize()
parser = Parser(tokenizer)
parser.parse_to_file(output_path)
class CompilationEngine(object):
def __init__(self, inFile):
self.t = Tokenizer(inFile)
self.symTable = SymbolTable()
self.vmName = inFile.rstrip('.jack') + '.vm'
self.vm = VMWriter(self.vmName)
self.className = ''
self.types = ['int', 'char', 'boolean', 'void']
self.stmnt = ['do', 'let', 'if', 'while', 'return']
self.subroutType = ''
self.whileIndex = 0
self.ifIndex = 0
self.fieldNum = 0
def compile_class(self):
self.t.advance()
self.validator('class')
self.className = self.t.current_token()
self.t.advance()
self.validator('{')
self.fieldNum = self.compile_class_var_dec()
while self.t.symbol() != '}': # subroutines
self.compile_subroutine()
self.validator('}')
self.vm.close()
return
def compile_class_var_dec(self):
varKeyWords = ['field', 'static']
name = ''
kind = ''
varType = ''
counter = 0
while self.t.keyword() in varKeyWords:
kind = self.t.current_token()
self.validator(varKeyWords)
# variable type
varType = self.t.current_token()
self.validator(['int', 'char', 'boolean', 'IDENTIFIER'])
name = self.t.current_token()
self.symTable.define(name, varType, kind)
self.t.advance()
if kind == 'field':
counter += 1
while self.t.symbol() != ';': # checks multiple vars
self.validator(',')
name = self.t.current_token()
self.symTable.define(name, varType, kind)
self.t.advance()
if kind == 'field':
counter += 1
self.validator(';')
return counter
def compile_subroutine(self):
current_subrout_scope = self.symTable.subDict
self.symTable.start_subroutine()
subroutKword = self.t.current_token()
self.validator(['constructor', 'function', 'method'])
self.subroutType = self.t.current_token()
self.validator(['int', 'char', 'boolean', 'void', 'IDENTIFIER'])
name = self.t.current_token()
subroutName = self.className + '.' + name
self.t.advance()
self.validator('(')
if subroutKword == 'method':
self.compile_parameter_list(method=True)
else:
self.compile_parameter_list()
self.validator(')')
self.validator('{')
if self.t.symbol() == '}':
self.t.advance()
return
self.validator(['var', 'let', 'do', 'if', 'while', 'return'],
advance=False)
numLocals = 0
if self.t.keyword() == 'var':
numLocals = self.compile_var_dec()
self.vm.write_function(subroutName, numLocals)
if subroutKword == 'constructor':
self.vm.write_push('constant', self.fieldNum)
self.vm.write_call('Memory.alloc', 1)
self.vm.write_pop('pointer', 0)
elif subroutKword == 'method':
self.vm.write_push('argument', 0)
self.vm.write_pop('pointer', 0)
if self.t.keyword() in self.stmnt:
self.compile_statements()
self.validator('}')
self.symTable.subDict = current_subrout_scope
self.whileIndex = 0
self.ifIndex = 0
return
def compile_parameter_list(self, method=False):
name = ''
varType = ''
kind = ''
counter = 0
if self.t.symbol() == ')':
return counter
varType = self.t.current_token()
self.validator(['int', 'char', 'boolean', 'void', 'IDENTIFIER'])
kind = 'arg'
name = self.t.current_token()
if method:
self.symTable.define(name, varType, kind, method=True)
else:
self.symTable.define(name, varType, kind)
self.t.advance()
counter += 1
while self.t.symbol() == ',':
self.validator(',')
self.validator(['int', 'char', 'boolean', 'void', 'IDENTIFIER'])
kind = 'arg'
name = self.t.current_token()
self.symTable.define(name, varType, kind)
self.t.advance()
counter += 1
return counter
def compile_var_dec(self):
name = ''
kind = ''
varType = ''
counter = 0
while self.t.keyword() == 'var': # check multiple lines of var
kind = 'var'
self.t.advance()
varType = self.t.current_token()
self.validator(['int', 'char', 'boolean', 'void', 'IDENTIFIER'])
name = self.t.current_token()
self.symTable.define(name, varType, kind)
self.t.advance()
counter += 1
while self.t.symbol() == ',': # multiple varNames
self.t.advance()
name = self.t.current_token()
self.symTable.define(name, varType, kind)
self.t.advance()
counter += 1
self.validator(';')
return counter
def compile_statements(self):
while self.t.keyword() in self.stmnt:
if self.t.keyword() == 'let':
self.compile_let()
elif self.t.keyword() == 'do':
self.compile_do()
elif self.t.keyword() == 'if':
self.compile_if()
elif self.t.keyword() == 'while':
self.compile_while()
elif self.t.keyword() == 'return':
self.compile_return()
else:
raise Exception(self.t.current_token() + ' is not valid')
return
def compile_do(self):
lookAhead = ''
self.t.advance() # do
lookAhead = self.t.tokens[self.t.tokenIndex + 1]
if lookAhead == '(': # subroutineName(exprlist)
subroutName = self.className + '.' + self.t.current_token()
self.t.advance()
self.validator('(')
self.vm.write_push('pointer', 0)
numArgs = self.compile_expression_list()
self.vm.write_call(subroutName, numArgs + 1) # add 1 for 'this'
self.validator(')')
self.validator(';')
self.vm.write_pop('temp', 0) # throws away returned value
return
else:
className = self.t.current_token()
self.t.advance()
self.validator('.') # name.subroutine(exprList)
subroutName = self.t.current_token()
self.t.advance()
self.validator('(')
if self.symTable.kind_of(className) in [
'this', 'static', 'local', 'argument'
]:
# used 'this' for 'field'
typeName = self.symTable.type_of(className)
subroutName = typeName + '.' + subroutName
segment = self.symTable.kind_of(className)
index = self.symTable.index_of(className)
self.vm.write_push(segment, index)
numArgs = self.compile_expression_list()
self.vm.write_call(subroutName, numArgs + 1)
else:
subroutName = className + '.' + subroutName
numArgs = self.compile_expression_list()
self.vm.write_call(subroutName, numArgs)
self.validator(')')
self.validator(';')
self.vm.write_pop('temp', 0)
return
def compile_let(self):
name = ''
kind = ''
array = False
self.t.advance() # let
while self.t.symbol() != ';':
name = self.t.identifier()
kind = self.symTable.kind_of(name)
index = self.symTable.index_of(name)
if name in self.symTable.classDict:
self.t.advance()
elif name in self.symTable.subDict:
self.t.advance()
else:
raise Exception(self.t.identifier() + ' is not defined')
if self.t.symbol() == '[': # array index
array = True
self.vm.write_push(kind, index)
self.validator('[')
self.compile_expression()
self.validator(']')
self.vm.write_arithmetic('+')
self.validator('=')
self.compile_expression()
if array:
self.vm.write_pop('temp', 0)
self.vm.write_pop('pointer', 1)
self.vm.write_push('temp', 0)
self.vm.write_pop('that', 0)
else:
self.vm.write_pop(kind, index)
self.validator(';')
return
def compile_while(self):
currentWhile = 'WHILE' + str(self.whileIndex)
self.vm.write_label(currentWhile)
self.whileIndex += 1
self.t.advance() # while
self.validator('(')
self.compile_expression()
self.vm.write_arithmetic('~')
self.vm.write_if('END' + currentWhile)
self.validator(')')
self.validator('{')
self.compile_statements()
self.vm.write_goto(currentWhile)
self.validator('}')
self.vm.write_label('END' + currentWhile)
return
def compile_return(self):
self.t.advance() # return
if self.t.symbol() == ';':
self.vm.write_push('constant', '0')
self.vm.write_return()
self.t.advance()
else:
self.compile_expression()
self.validator(';')
self.vm.write_return()
return
def compile_if(self):
endIf = 'END_IF' + str(self.ifIndex)
currentElse = 'IF_ELSE' + str(self.ifIndex)
self.ifIndex += 1
self.t.advance() # if
self.validator('(')
self.compile_expression()
self.vm.write_arithmetic('~')
self.vm.write_if(currentElse)
self.validator(')')
self.validator('{')
self.compile_statements()
self.vm.write_goto(endIf)
self.validator('}')
self.vm.write_label(currentElse)
if self.t.keyword() == 'else':
self.t.advance() # else
self.validator('{')
self.compile_statements()
self.validator('}')
self.vm.write_label(endIf)
return
def compile_expression(self):
op = ['+', '-', '*', '/', '&', '|', '<', '>', '=']
self.compile_term()
while self.t.symbol() in op:
opToken = self.t.current_token()
self.t.advance()
self.compile_term()
self.vm.write_arithmetic(opToken)
return
def compile_term(self):
keyConst = ['true', 'false', 'null', 'this']
unOps = ['-', '~']
lookAhead = ''
name = ''
current_subrout_scope = ''
if self.t.token_type() == 'INT_CONST':
self.vm.write_push('constant', self.t.int_val())
self.t.advance()
elif self.t.token_type() == 'STRING_CONST':
string = self.t.string_val()
length = len(string)
self.vm.write_push('constant', length)
self.vm.write_call('String.new', 1)
for char in string:
char = ord(char) # gives the ASCII number
self.vm.write_push('constant', char)
self.vm.write_call('String.appendChar', 2)
self.t.advance()
elif self.t.token_type() == 'KEYWORD':
self.validator(keyConst, advance=False)
if self.t.current_token() in ['false', 'null']:
self.t.advance()
self.vm.write_push('constant', '0')
elif self.t.current_token() == 'true':
self.vm.write_push('constant', '1')
self.vm.write_arithmetic('-', neg=True)
self.t.advance()
else:
self.vm.write_push('pointer', '0')
self.t.advance()
elif self.t.token_type() == 'SYMBOL':
if self.t.symbol() in unOps: # unary operator
unOpToken = self.t.current_token()
self.t.advance()
self.compile_term()
self.vm.write_arithmetic(unOpToken, neg=True)
elif self.t.symbol() == '(': # (expression))
self.t.advance()
self.compile_expression()
self.t.advance()
else:
raise Exception(self.t.current_token() + ' is not valid')
elif self.t.token_type() == 'IDENTIFIER': # varName, array, or subcall
lookAhead = self.t.tokens[self.t.tokenIndex + 1]
if lookAhead == '[': # array item
name = self.t.identifier()
kind = self.symTable.kind_of(name)
index = self.symTable.index_of(name)
if name in self.symTable.classDict:
self.t.advance()
elif name in self.symTable.subDict:
self.t.advance()
else:
raise Exception(self.t.identifier() + ' is not defined')
self.vm.write_push(kind, index)
self.validator('[')
self.compile_expression()
self.vm.write_arithmetic('+')
self.vm.write_pop('pointer', 1)
self.vm.write_push('that', 0)
self.validator(']')
elif lookAhead == '(': # subcall
current_subrout_scope = self.symTable.subDict
name = self.className + '.' + self.t.current_token()
self.t.advance()
self.validator('(')
numArgs = self.compile_expression_list()
self.vm.write_call(name, numArgs + 1)
self.validator(')')
self.symTable.subDict = current_subrout_scope
elif lookAhead == '.': # name.subroutName(expressList)
current_subrout_scope = self.symTable.subDict
className = self.t.current_token()
self.t.advance()
self.validator('.')
subroutName = self.t.current_token()
self.validator('IDENTIFIER')
name = className + '.' + subroutName
self.validator('(')
if self.symTable.kind_of(className) in [
'this', 'static', 'local', 'argument'
]:
# used 'this' for 'field'
classType = self.symTable.type_of(className)
name = classType + '.' + subroutName
kind = self.symTable.kind_of(className)
index = self.symTable.index_of(className)
self.vm.write_push(kind, index)
numArgs = self.compile_expression_list()
self.vm.write_call(name, numArgs + 1)
else:
numArgs = self.compile_expression_list()
self.vm.write_call(name, numArgs)
self.validator(')')
self.symTable.subDict = current_subrout_scope
else:
name = self.t.identifier() # varName
kind = self.symTable.kind_of(name)
index = self.symTable.index_of(name)
self.vm.write_push(kind, index)
self.t.advance()
else:
raise Exception(self.t.current_token() + ' is not valid')
return
def compile_expression_list(self): # only in subroutineCall
counter = 0
if self.t.symbol() == ')':
return counter
else:
self.compile_expression()
counter += 1
while self.t.symbol() == ',':
self.t.advance()
self.compile_expression()
counter += 1
return counter
def validator(self, syntax, advance=True):
tokenType = self.t.token_type()
token = self.t.current_token()
if advance:
self.t.advance()
if type(syntax) != list:
syntax = [syntax]
for item in syntax:
if item in [tokenType, token]:
return True
raise Exception(self.t.current_token() + ' is not valid')
def __init__(self, inpath, outpath):
self.tokenizer = Tokenizer(inpath)
XMLWriter.set_filepath(outpath)
if self.tokenizer.has_more_tokens():
self.compile_class()
XMLWriter.close()
class Compiler:
def __init__(self, inpath, outpath):
self.tokenizer = Tokenizer(inpath)
XMLWriter.set_filepath(outpath)
if self.tokenizer.has_more_tokens():
self.compile_class()
XMLWriter.close()
def _write_current_terminal(self):
XMLWriter.write_terminal(self._current_token, self._current_tok_tag)
def _advance(self):
self._check_EOF()
self.tokenizer.advance()
type_kws = (KW_INT, KW_CHAR, KW_BOOLEAN)
kw_consts = (KW_TRUE, KW_FALSE, KW_NULL, KW_THIS)
@property
def _current_token(self):
t_type = self.tokenizer.token_type
return (self.tokenizer.keyword if t_type == T_KEYWORD else
self.tokenizer.symbol if t_type == T_SYMBOL else
self.tokenizer.identifier if t_type == T_ID else
self.tokenizer.intval if t_type == T_INTEGER else
self.tokenizer.stringval)
@property
def _current_tok_type(self):
return self.tokenizer.token_type
@property
def _current_tok_tag(self):
return token_tags[self._current_tok_type]
@property
def _next_token(self):
"""return raw next_token in the tokenizer"""
return str(self.tokenizer.next_token)
def _require_token(self, tok_type, token=None):
"""Check whether the next_token(terminal) in the tokenizer meets the
requirement (specific token or just token type). If meets, tokenizer
advances (update current_token and next_token) and terminal will be
writed into outfile; If not, report an error."""
self._advance()
if token and self._current_token != token:
return self._error(expect_toks=(token,))
elif self._current_tok_type != tok_type:
return self._error(expect_types=(tok_type,))
self._write_current_terminal()
def _require_id(self):
return self._require_token(T_ID)
def _require_kw(self, token):
return self._require_token(T_KEYWORD, token=token)
def _require_sym(self, token):
return self._require_token(T_SYMBOL, token=token)
def _require_brackets(self, brackets, procedure):
front, back = brackets
self._require_sym(front)
procedure()
self._require_sym(back)
def _fol_by_class_vardec(self):
return self._next_token in (KW_STATIC, KW_FIELD)
def _fol_by_subroutine(self):
return self._next_token in (KW_CONSTRUCTOR, KW_FUNCTION, KW_METHOD)
def _fol_by_vardec(self):
return self._next_token == KW_VAR
#########################
# structure compilation #
#########################
# the compilation of three types of name might seem redundant here, but
# it was for abstraction and later code generation in project 11.
def compile_class_name(self):
self._require_id()
def compile_subroutine_name(self):
self._require_id()
def compile_var_name(self):
self._require_id()
def compile_type(self, advanced=False, expect='type'):
# int, string, boolean or identifier(className)
if advanced is False:
self._advance()
if self._current_token in self.type_kws:
return self._write_current_terminal()
elif self._current_tok_type == T_ID:
return self._write_current_terminal()
else:
return self._error(expect=expect)
def compile_void_or_type(self):
# void or type
self._advance()
if self._current_token == KW_VOID:
self._write_current_terminal()
else:
self.compile_type(True, '"void" or type')
@record_non_terminal('class')
def compile_class(self):
# 'class' className '{' classVarDec* subroutineDec* '}'
self._require_kw(KW_CLASS)
self.compile_class_name()
self._require_sym('{')
while self._fol_by_class_vardec():
self.compile_class_vardec()
while self._fol_by_subroutine():
self.compile_subroutine()
self._advance()
if self._current_token != '}':
self._traceback("Except classVarDec or subroutineDec.")
self._write_current_terminal()
def compile_declare(self):
self._advance()
self._write_current_terminal()
# type varName (',' varName)* ';'
self.compile_type()
self.compile_var_name()
# compile ',' or ';'
self._advance()
while self._current_token == ',':
self._write_current_terminal()
self.compile_var_name()
self._advance()
if self._current_token != ';':
return self._error((',', ';'))
self._write_current_terminal()
@record_non_terminal('classVarDec')
def compile_class_vardec(self):
# ('static|field') type varName (',' varName)* ';'
self.compile_declare()
@record_non_terminal('subroutineDec')
def compile_subroutine(self):
# ('constructor'|'function'|'method')
# ('void'|type) subroutineName '(' parameterList ')' subroutineBody
self._advance()
self._write_current_terminal() # ('constructor'|'function'|'method')
self.compile_void_or_type()
self.compile_subroutine_name()
self._require_brackets('()', self.compile_parameter_list)
self.compile_subroutine_body()
@record_non_terminal('parameterList')
def compile_parameter_list(self):
# ((type varName) (',' type varName)*)?
if self._next_token == ')':
return
self.compile_type()
self.compile_var_name()
while self._next_token != ')':
self._require_sym(',')
self.compile_type()
self.compile_var_name()
@record_non_terminal('subroutineBody')
def compile_subroutine_body(self):
# '{' varDec* statements '}'
self._require_sym('{')
while self._fol_by_vardec():
self.compile_vardec()
self.compile_statements()
self._require_sym('}')
@record_non_terminal('varDec')
def compile_vardec(self):
# 'var' type varName (',' varName)* ';'
self.compile_declare()
#########################
# statement compilation #
#########################
@record_non_terminal('statements')
def compile_statements(self):
# (letStatement | ifStatement | whileStatement | doStatement |
# returnStatement)*
last_statement = None
while self._next_token != '}':
self._advance()
last_statement = self._current_token
if last_statement == 'do':
self.compile_do()
elif last_statement == 'let':
self.compile_let()
elif last_statement == 'while':
self.compile_while()
elif last_statement == 'return':
self.compile_return()
elif last_statement == 'if':
self.compile_if()
else:
return self._error(expect='statement expression')
if STACK[-2] == 'subroutineBody' and last_statement != 'return':
self._error(expect='return statement', get=last_statement)
@record_non_terminal('doStatement')
def compile_do(self):
# 'do' subroutineCall ';'
self._write_current_terminal()
# compile identifier first
self._advance()
self.compile_subroutine_call()
self._require_sym(';')
@record_non_terminal('letStatement')
def compile_let(self):
# 'let' varName ('[' expression ']')? '=' expression ';'
self._write_current_terminal()
self.compile_var_name()
if self._next_token == '[':
self._compile_array_subscript()
self._require_sym('=')
self.compile_expression()
self._require_sym(';')
@record_non_terminal('whileStatement')
def compile_while(self):
# 'while' '(' expression ')' '{' statements '}'
self._write_current_terminal()
self._require_brackets('()', self.compile_expression)
self._require_brackets('{}', self.compile_statements)
@record_non_terminal('returnStatement')
def compile_return(self):
# 'return' expression? ';'
self._write_current_terminal()
if self._next_token != ';':
self.compile_expression()
self._require_sym(';')
@record_non_terminal('ifStatement')
def compile_if(self):
# 'if' '(' expression ')' '{' statements '}'
# ('else' '{' statements '}')?
self._write_current_terminal()
self._require_brackets('()', self.compile_expression)
self._require_brackets('{}', self.compile_statements)
# else clause
if self._next_token == KW_ELSE:
self._require_kw(KW_ELSE)
self._require_brackets('{}', self.compile_statements)
##########################
# expression compilation #
##########################
@record_non_terminal('expression')
def compile_expression(self):
# term (op term)*
self.compile_term()
while is_op(self._next_token):
self.compile_op()
self.compile_term()
@record_non_terminal('term')
def compile_term(self):
# integerConstant | stringConstant | keywordConstant |
# varName | varName '[' expression ']' | subroutineCall |
# '(' expression ')' | unaryOp term
if self._next_token == '(':
self._require_brackets('()', self.compile_expression)
elif self._next_token in set('-~'):
self.compile_unaryop()
else:
self._advance()
tok = self._current_token
tok_type = self._current_tok_type
if tok in self.kw_consts or tok_type in (T_INTEGER, T_STRING):
self._write_current_terminal()
elif tok_type == T_ID:
if self._next_token in '(.':
self.compile_subroutine_call()
else:
self._write_current_terminal()
if self._next_token == '[':
self._compile_array_subscript()
else:
self._error(expect='term')
def compile_call_name(self):
# the fisrt name of subroutine call could be (className or varName) if
# it is followed by '.', or subroutineName if followed by '('.
if self._current_tok_type != T_ID:
self._error(expect_types=(T_ID,))
self._write_current_terminal()
# just write it without analysis.
# this method will be extended to decide which kind the name is.
def compile_subroutine_call(self):
# subroutineName '(' expressionList ')' | (className |
# varName) '.' subroutineName '(' expressionList ')'
## the first element of structure has already been compiled.
self.compile_call_name()
if self._next_token == '.':
self._require_sym('.')
self.compile_subroutine_name()
self._require_brackets('()', self.compile_expressionlist)
@record_non_terminal('expressionList')
def compile_expressionlist(self):
# (expression (',' expression)*)?
if self._next_token != ')':
self.compile_expression()
while self._next_token != ')':
self._require_sym(',')
self.compile_expression()
def compile_op(self):
# exclude '~'
self._advance()
if self._current_token == '~':
self._traceback('Unexpected operator: ~')
self._write_current_terminal()
def compile_unaryop(self):
self._advance()
self._write_current_terminal() # symbol: - or ~
self.compile_term()
def _compile_array_subscript(self):
# '[' expression ']'
self._require_brackets('[]', self.compile_expression)
def _check_EOF(self):
if not self.tokenizer.has_more_tokens():
self._traceback("Unexpected EOF.")
def _error(self, expect_toks=(), expect_types=(), expect=None, get=None):
if expect is None:
exp_tok = ' or '.join(('"{0}"'.format(t) for t in expect_toks))
exp_types = ('type {0}'.format(token_tags[t]) for t in expect_types)
exp_type = ' or '.join(exp_types)
if exp_tok and exp_type:
expect = ' or '.join(expect_tok, expect_type)
else:
expect = exp_toks + exp_types
if get is None:
get = self._current_token
me = 'Expect {0} but get "{1}"'.format(expect, get)
return self._traceback(me)
def _traceback(self, message):
file_info = 'file: "{0}"'.format(self.tokenizer.filename)
line_info = 'line {0}'.format(self.tokenizer.line_count)
raise CompileError("{0}, {1}: {2}".format(file_info, line_info, message))
