class CompilationEnginge(object):
"""
lalala
"""
def __init__(self, input_file, output_file):
self.tokenizer = JackTokenizer(input_file)
self.out = open(output_file, 'w')
self.token = None
self.class_name = None
#######################
### PROJECT 11 CODE ###
#######################
self.symbol_table = SymbolTable()
self.vm_writer = VMWriter(output_file)
#######################
def analyze(self):
self.token = self.tokenizer.advance()
self.compile_class()
self.close()
print('CLASS TABLE:')
print(self.symbol_table.class_table)
def close(self):
if self.out:
self.out.close()
self.out = None
def advance(self):
self.token = self.tokenizer.advance()
def write_to_out(self):
pass
def format_line(self, defined_or_used=''):
token_type = self.tokenizer.token_type()
running_index = ''
if token_type == self.tokenizer.keyword_token:
meat = self.tokenizer.keyword()
defined_or_used=''
elif token_type == self.tokenizer.symbol_token:
meat = self.tokenizer.symbol()
defined_or_used=''
elif token_type == self.tokenizer.identifier_token:
meat = self.tokenizer.identifier()
#######################
### PROJECT 11 CODE ###
#######################
# Extending compilaiton engine to output <var/argument/static/field...> instead of <indentifier>
name = self.tokenizer.token
if self.symbol_table.kind_of(name):
token_type = self.symbol_table.kind_of(name)
running_index = str(self.symbol_table.index_of(name))
elif name[0].islower():
token_type = 'subroutine'
else:
token_type = 'class'
#######################
elif token_type == self.tokenizer.int_const:
meat = self.tokenizer.int_val()
defined_or_used=''
elif token_type == self.tokenizer.string_const:
meat = self.tokenizer.string_val()
defined_or_used=''
else:
raise ValueError('Something went wrong with token: {}'.format(self.token))
if defined_or_used != '':
defined_or_used += ' '
if running_index != '':
running_index = ' ' + running_index
formated_line = '<{2}{0}{3}> {1} </{2}{0}{3}>\n'.format(token_type, meat, defined_or_used, running_index)
return formated_line
#########################
### PROGARM STRUCTURE ###
#########################
def compile_class(self):
"""
####################################################################
### class: 'class' className '{' classVarDec* subroutineDec* '}' ###
####################################################################
"""
self.out.write('<class>\n')
# 'class'
keyword_line = self.format_line()
self.out.write(keyword_line)
# className
self.advance()
#######################
### PROJECT 11 CODE ###
#######################
self.class_name = self.tokenizer.token
#######################
identifier_line = self.format_line('defined')
self.out.write(identifier_line)
# '{'
self.advance()
symbol_line = self.format_line()
self.out.write(symbol_line)
### classVarDec* subroutineDec* ###
self.advance()
# classVarDec*
while self.tokenizer.token_type() == self.tokenizer.keyword_token and self.tokenizer.keyword() in [self.tokenizer.key_field, self.tokenizer.key_static]:
self.compile_class_var_dec()
# subroutineDec*
while self.tokenizer.token_type() == self.tokenizer.keyword_token and self.tokenizer.keyword() in [self.tokenizer.key_function, self.tokenizer.key_method, self.tokenizer.key_constructor]:
self.compile_subroutine()
# '}'
if self.tokenizer.token_type() == self.tokenizer.symbol_token:
# Class compilation is done
symbol_line = self.format_line()
self.out.write(symbol_line)
else:
raise ValueError('Something went wrong')
# Closing with </class>
self.out.write('</class>\n')
is_sucessfull = not(self.advance())
if is_sucessfull:
print('Compilation enginge succesfully finished')
else:
print('Something went wrong!')
def compile_class_var_dec(self):
"""
#######################################################################
### classVarDec: ('static'|'field') type varName (',' varName)* ';' ###
#######################################################################
"""
self.out.write('<classVarDec>\n')
#######################
### PROJECT 11 CODE ###
#######################
# Extract field or static
# field_or_static = re.match('<[a-z]*>', field_or_static_line)[0][1:-1]
field_or_static = self.tokenizer.token
#######################
# ('static' | 'field')
field_or_static_line = self.format_line()
self.out.write(field_or_static_line)
# type
self.advance()
#######################
### PROJECT 11 CODE ###
#######################
# Extract token type
type_ = self.tokenizer.token
#######################
type_line = self.format_line()
self.out.write(type_line)
# varName
self.advance()
#######################
### PROJECT 11 CODE ###
#######################
self.symbol_table.define(name=self.tokenizer.token, type_=type_, kind=field_or_static)
#######################
varname_line = self.format_line('defined')
self.out.write(varname_line)
# (',' varName)*
self.advance()
symbol = self.tokenizer.symbol()
while symbol == ',':
colon_line = self.format_line()
self.out.write(colon_line)
self.advance()
#######################
### PROJECT 11 CODE ###
#######################
self.symbol_table.define(name=self.tokenizer.token, type_=type_, kind=field_or_static)
#######################
varname_line = self.format_line('defined')
self.out.write(varname_line)
self.advance()
symbol = self.tokenizer.symbol()
# symbol == ';'
semicolon_line = self.format_line()
self.out.write(semicolon_line)
self.advance()
self.out.write('</classVarDec>\n')
def compile_subroutine(self):
"""
###########################################################################
### subroutineDec: ('constructor'|'function'|'method') ###
### ('void' | type) subroutineName '(' parameterList ')' ###
### subroutineBody ###
###########################################################################
"""
#######################
### PROJECT 11 CODE ###
#######################
print()
print('SUBROUTINE TABLE:')
print(self.symbol_table.subroutine_table)
print()
self.symbol_table.start_subroutine()
self.symbol_table.define(name='this', type_=self.class_name, kind='argument')
#######################
self.out.write('<subroutineDec>\n')
# ('constructor'|'function'|'method')
constructor_function_method_line = self.format_line()
self.out.write(constructor_function_method_line)
# ('void' | type)
self.advance()
void_or_type_line = self.format_line()
self.out.write(void_or_type_line)
# subroutineName
self.advance()
subroutine_name_line = self.format_line('defined')
self.out.write(subroutine_name_line)
# '('
self.advance()
symbol_line = self.format_line()
self.out.write(symbol_line)
# parameterList
self.advance()
self.compile_parameter_list()
# ')'
symbol_line = self.format_line()
self.out.write(symbol_line)
##################################################
### subroutineBody: '{' varDec* statements '}' ###
##################################################
self.out.write('<subroutineBody>\n')
# '{'
self.advance()
symbol_line = self.format_line()
self.out.write(symbol_line)
###############
### varDec* ###
###############
self.advance()
while self.tokenizer.token == self.tokenizer.key_var:
self.compile_var_dec()
##################
### statements ###
##################
self.compile_statements()
# '}'
symbol_line = self.format_line()
self.out.write(symbol_line)
self.advance()
self.out.write('</subroutineBody>\n')
self.out.write('</subroutineDec>\n')
#######################
### PROJECT 11 CODE ###
#######################
print()
print('SUBROUTINE TABLE:')
print(self.symbol_table.subroutine_table)
print()
#######################
def compile_parameter_list(self):
"""
############################################################
### parameterList: ((type varName) (',' type varName)*)? ###
############################################################
"""
self.out.write('<parameterList>\n')
# If token type is symbol then we have empty parameter list
# If we have symbol token then it means our parameter list is fully processed
if self.tokenizer.token_type() != self.tokenizer.symbol_token:
# type
#######################
### PROJECT 11 CODE ###
#######################
type_ = self.tokenizer.token
#######################
type_line = self.format_line()
self.out.write(type_line)
# varName
self.advance()
#######################
### PROJECT 11 CODE ###
#######################
self.symbol_table.define(name=self.tokenizer.token, type_=type_, kind='argument')
#######################
var_name_line = self.format_line('defined')
self.out.write(var_name_line)
# If next token is ',' we have more then one parameter
self.advance()
while self.tokenizer.token_type() == self.tokenizer.symbol_token and self.tokenizer.symbol() == ',':
# ','
comma_line = self.format_line()
self.out.write(comma_line)
# type
self.advance()
#######################
### PROJECT 11 CODE ###
#######################
type_ = self.tokenizer.token
#######################
type_line = self.format_line()
self.out.write(type_line)
# varName
self.advance()
#######################
### PROJECT 11 CODE ###
#######################
self.symbol_table.define(name=self.tokenizer.token, type_=type_, kind='argument')
# We are in new subroutine so add next nested scope
# self.symbol_table.start_subroutine()
#######################
var_name_line = self.format_line('defined')
self.out.write(var_name_line)
self.advance()
self.out.write('</parameterList>\n')
def compile_var_dec(self):
"""
#####################################################
### varDec: 'var' type varName (',' varName)* ';' ###
#####################################################
"""
self.out.write('<varDec>\n')
# var
var_line = self.format_line()
self.out.write(var_line)
# type
self.advance()
#######################
### PROJECT 11 CODE ###
#######################
type_ = self.tokenizer.token
#######################
type_line = self.format_line()
self.out.write(type_line)
# varName
self.advance()
#######################
### PROJECT 11 CODE ###
#######################
self.symbol_table.define(name=self.tokenizer.token, type_=type_, kind='local')
#######################
var_name_line = self.format_line('defined')
self.out.write(var_name_line)
# (',' varName)*
self.advance()
while self.tokenizer.symbol() == ',':
# ','
comma_line = self.format_line()
self.out.write(comma_line)
# varName
self.advance()
#######################
### PROJECT 11 CODE ###
#######################
self.symbol_table.define(name=self.tokenizer.token, type_=type_, kind='local')
#######################
var_name_line = self.format_line('defined')
self.out.write(var_name_line)
self.advance()
# ';'
semicolon_line = self.format_line()
self.out.write(semicolon_line)
self.advance()
self.out.write('</varDec>\n')
##################
### STATEMENTS ###
##################
def compile_statements(self):
"""
##############################
### statements: statement* ###
##############################
"""
self.out.write('<statements>\n')
while self.tokenizer.token_type() != self.tokenizer.symbol_token:
keyword = self.tokenizer.keyword()
# letStatement
if keyword == self.tokenizer.key_let:
self.compile_let()
# ifStatement
elif keyword == self.tokenizer.key_if:
self.compile_if()
# whileStatement
elif keyword == self.tokenizer.key_while:
self.compile_while()
# doStatement
elif keyword == self.tokenizer.key_do:
self.compile_do()
# returnStatement
elif keyword == self.tokenizer.key_return:
self.compile_return()
else:
raise ValueError('Wrong statement: {}'.format(keyword))
self.out.write('</statements>\n')
def compile_do(self):
"""
############################################
### doStatement: 'do' subroutineCall ';' ###
############################################
"""
self.out.write('<doStatement>\n')
# 'do'
do_line = self.format_line()
self.out.write(do_line)
# subroutineCall
self.advance()
self.compile_subroutine_call()
# ';'
symbol_line = self.format_line()
self.out.write(symbol_line)
self.advance()
self.out.write('</doStatement>\n')
def compile_let(self):
"""
############################################################################
### letStatement: 'let' varName ('[' expression ']')? '=' expression ';' ###
############################################################################
"""
self.out.write('<letStatement>\n')
# let
let_line = self.format_line()
self.out.write(let_line)
# varName
self.advance()
var_name_line = self.format_line('used')
self.out.write(var_name_line)
# Check if '[' or '='
self.advance()
if self.tokenizer.token == '[':
# '['
symbol_line = self.format_line()
self.out.write(symbol_line)
# expression
self.advance()
self.compile_expression()
# ']'
symbol_line = self.format_line()
self.out.write(symbol_line)
self.advance()
# '='
symbol_line = self.format_line()
self.out.write(symbol_line)
# expression
self.advance()
self.compile_expression()
# ';'
symbol_line = self.format_line()
self.out.write(symbol_line)
self.advance()
self.out.write('</letStatement>\n')
def compile_while(self):
"""
#####################################################################
### whileStatement: 'while' '(' expression ')' '{' statements '}' ###
#####################################################################
"""
self.out.write('<whileStatement>\n')
# 'while'
while_line = self.format_line()
self.out.write(while_line)
# '('
self.advance()
symbol_line = self.format_line()
self.out.write(symbol_line)
# expression
self.advance()
self.compile_expression()
# ')'
symbol_line = self.format_line()
self.out.write(symbol_line)
# '{'
self.advance()
symbol_line = self.format_line()
self.out.write(symbol_line)
# statements
self.advance()
self.compile_statements()
# '}'
symbol_line = self.format_line()
self.out.write(symbol_line)
self.advance()
self.out.write('</whileStatement>\n')
def compile_return(self):
"""
################################################
### ReturnStatement 'return' expression? ';' ###
################################################
"""
self.out.write('<returnStatement>\n')
# 'return'
return_line = self.format_line()
self.out.write(return_line)
# Ceck if expression
self.advance()
if self.tokenizer.token != ';':
# 'expression'
self.compile_expression()
# ';'
symbol_line = self.format_line()
self.out.write(symbol_line)
self.advance()
self.out.write('</returnStatement>\n')
def compile_if(self):
"""
###############################################################
### ifStatement: 'if' '(' expression ')' '{' statements '}' ###
### ('else' '{' statements '}')? ###
###############################################################
"""
self.out.write('<ifStatement>\n')
# 'if'
if_line = self.format_line()
self.out.write(if_line)
# '('
self.advance()
symbol_line = self.format_line()
self.out.write(symbol_line)
# expression
self.advance()
self.compile_expression()
# ')'
symbol_line = self.format_line()
self.out.write(symbol_line)
# '{'
self.advance()
symbol_line = self.format_line()
self.out.write(symbol_line)
# statements
self.advance()
self.compile_statements()
# '}'
symbol_line = self.format_line()
self.out.write(symbol_line)
# Check if there is 'else' part of ifStatement
self.advance()
if self.tokenizer.token_type() == self.tokenizer.keyword_token and self.tokenizer.keyword() == 'else':
# 'else'
else_line = self.format_line()
self.out.write(else_line)
# '{'
self.advance()
symbol_line = self.format_line()
self.out.write(symbol_line)
# statements
self.advance()
self.compile_statements()
# '}'
symbol_line = self.format_line()
self.out.write(symbol_line)
self.advance()
self.out.write('</ifStatement>\n')
###################
### EXPRESSIONS ###
###################
def compile_subroutine_call(self, skip_subroutine_name=False):
"""
############################################################################
### subroutineCall: subroutineName '(' expressionList ')' | (className | ###
### varName) '.' subroutineName '(' expressionList ')' ###
############################################################################
"""
if not skip_subroutine_name:
# subroutineName or className or varName
subroutine_class_var_name_line = self.format_line('used')
self.out.write(subroutine_class_var_name_line)
self.advance()
# Check '(' or '.'
if self.tokenizer.token == '.':
# '.'
symbol_line = self.format_line()
self.out.write(symbol_line)
# subroutineName
self.advance()
subroutine_name_line = self.format_line('used')
self.out.write(subroutine_name_line)
self.advance()
# '('
symbol_line = self.format_line()
self.out.write(symbol_line)
# expressionList
self.advance()
self.compile_expression_list()
# ')'
symbol_line = self.format_line()
self.out.write(symbol_line)
self.advance()
def compile_expression(self):
"""
###################################
### expression: term (op term)* ###
###################################
"""
self.out.write('<expression>\n')
ops = ['+', '-', '*', '/', '&', '|', '<', '>', '=']
# 'term'
self.compile_term()
# Check if there is (op term)* part
while self.tokenizer.token in ops:
# op
op_line = self.format_line()
self.out.write(op_line)
# term
self.advance()
self.compile_term()
self.out.write('</expression>\n')
def compile_term(self):
"""
################################################################
### integerConstant | stringConstant | keywordConstant | ###
### varName | varName '[' expression ']' | subroutineCall | ###
### '(' expression ')' | unaryOp term ###
################################################################
"""
self.out.write('<term>\n')
unary_ops = ['-', '~']
#############################################
### constant, name, expression or unaryOp ###
#############################################
# '(' expression ')'
if self.tokenizer.token == '(':
# '('
symbol_line = self.format_line()
self.out.write(symbol_line)
# expression
self.advance()
self.compile_expression()
# ')'
symbol_line = self.format_line()
self.out.write(symbol_line)
self.advance()
# unaryOp term
elif self.tokenizer.token in unary_ops:
# unaryOp
unary_op_line = self.format_line()
self.out.write(unary_op_line)
# term
self.advance()
self.compile_term()
# integerConstant | stringConstant | keywordConstant |
# varName | varName '[' expression ']' | subroutineCall
else:
# constant or name
constant_or_name = self.format_line('used')
self.out.write(constant_or_name)
# varName '[' expression ']' | subroutineCall or end of compile_term function
# Check if expression: '[', subroutineCall: '(' with parameter skip_subroutine_name = True,
# otherwise end of compile_term function
self.advance()
# '[' expression ']'
if self.tokenizer.token == '[':
# '['
symbol_line = self.format_line()
self.out.write(symbol_line)
# expression
self.advance()
self.compile_expression()
# ']'
symbol_line = self.format_line()
self.out.write(symbol_line)
self.advance()
# subroutineCall with skip_subroutine_name=True
elif self.tokenizer.token in ['(', '.']:
self.compile_subroutine_call(skip_subroutine_name=True)
self.out.write('</term>\n')
def compile_expression_list(self):
"""
########################################################
### expressionList: (expression (',' expression)* )? ###
########################################################
"""
self.out.write('<expressionList>\n')
# Check if token is ')', if so we got empty expression list
if self.tokenizer.token != ')':
# 'expression'
self.compile_expression()
# Check if token is ',', if so we got more expressions
while self.tokenizer.token == ',':
# ','
comma_line = self.format_line()
self.out.write(comma_line)
# expression
self.advance()
self.compile_expression()
self.out.write('</expressionList>\n')
class CompilationEngine(object):
# the destination file for writing
destination_file = None
# the tokenizer for the input file
tokenizer = None
# symbol table
symbol_table = None
# vm writer
vm_writer = None
# the class name
class_name = ""
# indicies for if and while loops
# start at -1 because we increment before use
while_index = -1
if_index = -1
# the constructor for compiling a single class
# the next method to be called after construction must be compile_class
# source_filename must be a single file, not a directory
def __init__(self, source_filename):
# destination filename
# if the original extension was .jack, then make the extension .vm
# if the original extension was not .jack, then append .vm
if source_filename.lower().endswith(".jack"):
destination_filename = source_filename[:-5] + ".vm"
else:
destination_filename = source_filename + ".vm"
# open the destination filename for writing
self.destination_file = open(destination_filename, 'w')
# create a tokenizer for the input file
self.tokenizer = JackTokenizer(source_filename)
# create the symbol table
self.symbol_table = SymbolTable()
# create the vm writer
self.vm_writer = VMWriter(self.destination_file)
# compiles a complete class and closes the output file
def compile_class(self):
# class keyword
tt, t = self._token_next(True, "KEYWORD", "class")
# name of class
tt, t = self._token_next(True, "IDENTIFIER")
self.class_name = t
# open brace
tt, t = self._token_next(True, "SYMBOL", "{")
# one or more variable declarations
self.tokenizer.advance()
while True:
tt, t = self._token_next(False)
if tt == "KEYWORD" and t in ["field", "static"]:
self.compile_class_var_dec()
else:
# stop trying to process variable declarations
break
# one or more subroutine declarations
while True:
tt, t = self._token_next(False)
if tt == "KEYWORD" and t in ["constructor", "function", "method"]:
self.compile_subroutine()
else:
# stop trying to process functions
break
# close brace
# do not advance because we already advanced upon exiting the last loop
tt, t = self._token_next(False, "SYMBOL", "}")
# done with compilation; close the output file
self.destination_file.close()
# compiles a static declaration or field declaration
def compile_class_var_dec(self):
# compile the variable declaration
# False means this is a class (not a subroutine)
self.compile_var_dec(False)
# compiles a complete method, function, or constructor
def compile_subroutine(self):
# start of subroutine
self.symbol_table.start_subroutine()
# constructor, function, or method keyword
tt, type = self._token_next(False, "KEYWORD")
# type of the return value
# can be either keyword (void) or an identifier (any type)
tt, t = self._token_next(True)
# name of the method/function/constructor
tt, name = self._token_next(True)
name = self.class_name + "." + name
# if the type is a method, "define" this as an argument, so the other
# argument indexes work correctly
if type == "method":
self.symbol_table.define("this", self.class_name, SymbolTable.ARG)
# opening parenthesis
tt, t = self._token_next(True, "SYMBOL", "(")
# arguments
self.tokenizer.advance()
self.compile_parameter_list()
# closing parenthesis
tt, t = self._token_next(False, "SYMBOL", ")")
# opening brace
tt, t = self._token_next(True, "SYMBOL", "{")
# variable declarations
self.tokenizer.advance()
while True:
tt, t = self._token_next(False)
if tt == "KEYWORD" and t == "var":
self.compile_var_dec()
else:
# stop trying to process variable declarations
break
# write the function
num_locals = self.symbol_table.var_count(self.symbol_table.VAR)
self.vm_writer.write_function(name, num_locals)
# write any special code at the top of the function
if type == "constructor":
# code to allocate memory and set "this"
size = self.symbol_table.var_count(self.symbol_table.FIELD)
self.vm_writer.write_push(self.vm_writer.CONST, size)
self.vm_writer.write_call("Memory.alloc", 1)
self.vm_writer.write_pop(self.vm_writer.POINTER, 0)
elif type == "function":
# nothing special
pass
elif type == "method":
# put argument 0 into pointer 0 (this)
self.vm_writer.write_push(self.vm_writer.ARG, 0)
self.vm_writer.write_pop(self.vm_writer.POINTER, 0)
else:
print "WARNING: Expected constructor, function, or name; got", type
# statements
self.compile_statements()
# closing brace
tt, t = self._token_next(False, "SYMBOL", "}")
self.tokenizer.advance()
# compiles a (possibly empty) parameter list, not including the enclosing
# parentheses
def compile_parameter_list(self):
# check for empty list
tt, t = self._token_next(False)
if tt == "SYMBOL" and t == ")":
# the parameter list was empty; do not process any more
pass
else:
# there are things in the parameter list
while True:
# keyword (variable type)
tt, type = self._token_next(False)
# identifier (variable name)
tt, name = self._token_next(True)
# the kind is always an arg, since these are all parameters to the
# function
kind = SymbolTable.ARG
# define the variable in the symbol table
self.symbol_table.define(name, type, kind)
# possible comma
tt, t = self._token_next(True)
if tt != "SYMBOL" or t != ",":
# not a comma; stop processing parameters
break
self.tokenizer.advance()
# compiles a var declaration
# if subroutine is true, only the var keyword can be used
# if subroutine is false, only the static and field keywords can be used
def compile_var_dec(self, subroutine=True):
# the keyword to start the declaration
tt, kind = self._token_next(False, "KEYWORD")
# check for required types
if subroutine:
if kind == "var":
kind = SymbolTable.VAR
else:
print "WARNING: expecting var, but received %s" % (str(kind))
else:
if kind == "static":
kind = SymbolTable.STATIC
elif kind == "field":
kind = SymbolTable.FIELD
else:
print "WARNING: expecting static or field, but received %s" % (str(kind))
# type of the declaration
# could be an identifier or a keyword (int, etc)
tt, type = self._token_next(True)
# name of the declaration
tt, name = self._token_next(True, "IDENTIFIER")
# define the variable in the symbol table
self.symbol_table.define(name, type, kind)
# can support more than one identifier name, to declare more than one
# variable, separated by commas; process the 2nd-infinite variables
self.tokenizer.advance()
while True:
tt, t = self._token_next(False)
if tt == "SYMBOL" and t == ",":
# another variable name follows
tt, name = self._token_next(True, "IDENTIFIER")
# define the variable in the symbol table
self.symbol_table.define(name, type, kind)
self.tokenizer.advance()
else:
# no more variable names
break
# should be on the semicolon at the end of the line
tt, t = self._token_next(False, "SYMBOL", ";")
self.tokenizer.advance()
# compiles a sequence of statements, not including the enclosing {}
def compile_statements(self):
while True:
tt, t = self._token_next(False)
if tt == "KEYWORD" and t in ["do", "let", "while", "return", "if"]:
# call compile_t, where t is the type of compilation we want
token = getattr(self, "compile_" + t)()
else:
# not a statement; stop processing statements
break
# compiles a do statement
def compile_do(self):
# do keyword
tt, t = self._token_next(False, "KEYWORD", "do")
# subroutine call
self.tokenizer.advance()
self.compile_subroutine_call()
# do statements do not have a return value, so eliminate the return
# off of the stack
self.vm_writer.write_pop(self.vm_writer.TEMP, 0)
# semicolon
tt, t = self._token_next(False, "SYMBOL", ";")
self.tokenizer.advance()
# compiles a let statement
def compile_let(self):
# let keyword
tt, t = self._token_next(False, "KEYWORD", "let")
# variable name
tt, name = self._token_next(True, "IDENTIFIER")
# possible brackets for array
tt, t = self._token_next(True)
if tt == "SYMBOL" and t == "[":
# array - write operation
array = True
# compile the offset expression
self.tokenizer.advance()
self.compile_expression()
# write the base address onto the stack
segment, index = self._resolve_symbol(name)
self.vm_writer.write_push(segment, index)
# add base and offset
self.vm_writer.write_arithmetic("add")
# we cannot yet put the result into pointer 1, since the read
# operation (which hasn't been parsed/computed yet) may use pointer 1
# to read from an arrya value
# closing bracket
tt, t = self._token_next(False, "SYMBOL", "]")
# advance to the next token, since we are expected to be on the = for
# the next line
self.tokenizer.advance()
else:
array = False
# equals sign
tt, t = self._token_next(False, "SYMBOL", "=")
# expression
self.tokenizer.advance()
self.compile_expression()
if array:
# our stack now looks like this:
# TOP OF STACK
# computed result to store
# address in which value should be stored
# ... previous stuff ...
# pop the computed value to temp 0
self.vm_writer.write_pop(self.vm_writer.TEMP, 0)
# pop the array address to pointer 1 (that)
self.vm_writer.write_pop(self.vm_writer.POINTER, 1)
# put the computed value back onto the stack
self.vm_writer.write_push(self.vm_writer.TEMP, 0)
# pop to the variable name or the array reference
self.vm_writer.write_pop(self.vm_writer.THAT, 0)
else:
# not an array - pop the expression to the variable
segment, index = self._resolve_symbol(name)
self.vm_writer.write_pop(segment, index)
# semicolon
tt, t = self._token_next(False, "SYMBOL", ";")
self.tokenizer.advance()
# compiles a while statement
def compile_while(self):
# labels for this while loop
self.while_index += 1
while_start = "WHILE_START_%d" % (self.while_index)
while_end = "WHILE_END_%d" % (self.while_index)
# while keyword
tt, t = self._token_next(False, "KEYWORD", "while")
# opening parenthesis
tt, t = self._token_next(True, "SYMBOL", "(")
# label for the start of the while statement
self.vm_writer.write_label(while_start)
# the expression that is the condition of the while statement
self.tokenizer.advance()
self.compile_expression()
# the closing parenthesis
tt, t = self._token_next(False, "SYMBOL", ")")
# the result of the evaluation is now on the stack
# if false, then goto to the end of the loop
# to do this, negate and then call if-goto
self.vm_writer.write_arithmetic("not")
self.vm_writer.write_if(while_end)
# the opening brace
tt, t = self._token_next(True, "SYMBOL", "{")
# the statments that is the body of the while loop
self.tokenizer.advance()
self.compile_statements()
# the closing brace
tt, t = self._token_next(False, "SYMBOL", "}")
# after the last statement of the while loop
# need to jump back up to the top of the loop to evaluate again
self.vm_writer.write_goto(while_start)
# label at the end of the loop
self.vm_writer.write_label(while_end)
self.tokenizer.advance()
# compiles a return statement
def compile_return(self):
# return keyword
tt, t = self._token_next(False, "KEYWORD", "return")
# possible expression to return
tt, t = self._token_next(True)
if tt != "SYMBOL" and t != ";":
self.compile_expression()
else:
# no return expression; return 0
self.vm_writer.write_push(self.vm_writer.CONST, 0)
# ending semicolon
tt, t = self._token_next(False, "SYMBOL", ";")
self.vm_writer.write_return()
self.tokenizer.advance()
# compiles a if statement, including a possible trailing else clause
def compile_if(self):
# it is more efficient in an if-else case to have the else portion first
# in the code when testing, but we use the less-efficient but
# easier-to-write true-false pattern here
# labels for this if statement
self.if_index += 1
if_false = "IF_FALSE_%d" % (self.if_index)
if_end = "IF_END_%d" % (self.if_index)
# if keyword
tt, t = self._token_next(False, "KEYWORD", "if")
# opening parenthesis
tt, t = self._token_next(True, "SYMBOL", "(")
# expression of if statement
self.tokenizer.advance()
self.compile_expression()
# closing parenthesis
tt, t = self._token_next(False, "SYMBOL", ")")
# the result of the evaluation is now on the stack
# if false, then goto the false label
# if true, fall through to executing code
# if there is no else, then false and end are the same, but having two
# labels does not increase code size
self.vm_writer.write_arithmetic("not")
self.vm_writer.write_if(if_false)
# opening brace
tt, t = self._token_next(True, "SYMBOL", "{")
# statements for true portion
self.tokenizer.advance()
self.compile_statements()
# closing brace
tt, t = self._token_next(False, "SYMBOL", "}")
tt, t = self._token_next(True)
if tt == "KEYWORD" and t == "else":
# else statement exists
# goto the end of the if statement at the end of the true portion
self.vm_writer.write_goto(if_end)
# label for the start of the false portion
self.vm_writer.write_label(if_false)
# opening brace
tt, t = self._token_next(True, "SYMBOL", "{")
# statements
self.tokenizer.advance()
self.compile_statements()
# closing brace
tt, t = self._token_next(False, "SYMBOL", "}")
# end label
self.vm_writer.write_label(if_end)
# advance tokenizer only if we are in the else, since otherwise the
# token was advanced by the else check
self.tokenizer.advance()
else:
# no else portion; only put in a label for false, since end is not
# used
self.vm_writer.write_label(if_false)
# compiles an expression (one or more terms connected by operators)
def compile_expression(self):
# the first term
self.compile_term()
# finish any number of operators followed by terms
while True:
tt, t = self._token_next(False)
if tt == "SYMBOL" and t in "+-*/&|<>=":
# found an operator
# postfix order - add the next term and then do the operator
# the next term
self.tokenizer.advance()
self.compile_term()
# the operator
if t == "+":
self.vm_writer.write_arithmetic("add")
if t == "-":
self.vm_writer.write_arithmetic("sub")
if t == "=":
self.vm_writer.write_arithmetic("eq")
if t == ">":
self.vm_writer.write_arithmetic("gt")
if t == "<":
self.vm_writer.write_arithmetic("lt")
if t == "&":
self.vm_writer.write_arithmetic("and")
if t == "|":
self.vm_writer.write_arithmetic("or")
if t == "*":
self.vm_writer.write_call("Math.multiply", 2)
if t == "/":
self.vm_writer.write_call("Math.divide", 2)
else:
# no term found; done parsing the expression
break
# compiles a term
# this routine is faced with a slight difficulty when trying to decide
# between some of the alternative parsing rules. specifically, if the
# current token is an identifier, the routine must distinguish between a
# variable, an array entry, and a subroutine call. a single lookahead token,
# which may be one of [, (, or ., suffices to distinguish between the three
# possibilities. any other token is not part of this term and should not
# be advanced over.
def compile_term(self):
# a term: integer_constant | string_constant | keyword_constant |
# varname | varname[expression] | subroutine_call | (expression) |
# unary_op term
tt, t = self._token_next(False)
if tt == "INT_CONST":
self.vm_writer.write_push(self.vm_writer.CONST, t)
# advance for the next statement
self.tokenizer.advance()
elif tt == "STRING_CONST":
# after this portion is run, a pointer to a string should be on the
# stack
# we create a new string of a certain size and then append characters
# one by one; each append operation returns the pointer to the same
# string
# create the string
# string is a len, data tuple; not null-terminated
size = len(t)
self.vm_writer.write_push(self.vm_writer.CONST, size)
self.vm_writer.write_call("String.new", 1)
# append each character
for char in t:
self.vm_writer.write_push(self.vm_writer.CONST, ord(char))
self.vm_writer.write_call("String.appendChar", 2)
# advance for the next statement
self.tokenizer.advance()
elif tt == "KEYWORD":
if t == "true":
# true is -1, which is 0 negated
self.vm_writer.write_push(self.vm_writer.CONST, 0)
self.vm_writer.write_arithmetic("not")
elif t == "false" or t == "null":
self.vm_writer.write_push(self.vm_writer.CONST, 0)
elif t == "this":
self.vm_writer.write_push(self.vm_writer.POINTER, 0)
# advance for the next statement
self.tokenizer.advance()
elif tt == "SYMBOL" and t == "(":
# ( expression )
# parse the expression
self.tokenizer.advance()
self.compile_expression()
# closing parenthesis
tt, t = self._token_next(False, "SYMBOL", ")")
# advance for the next statement
self.tokenizer.advance()
elif tt == "SYMBOL" and t in "-~":
# unary_op term
# postfix order - add the next term and then do the operator
# parse the rest of the term
self.tokenizer.advance()
self.compile_term()
# write the unary operation
if t == "-":
self.vm_writer.write_arithmetic("neg")
elif t == "~":
self.vm_writer.write_arithmetic("not")
elif tt == "IDENTIFIER":
# varname, varname[expression], subroutine_call
# do not write the identifer yet
# get the next bit of the expression
# if it is a [, then array; if it is a ( or ., then subroutine call
# if none of above, then pass over
tt2, t2 = self._token_next(True)
if tt2 == "SYMBOL" and t2 in "(.":
# subroutine call
# back up and then compile the subroutine call
self.tokenizer.retreat()
self.compile_subroutine_call()
elif tt2 == "SYMBOL" and t2 == "[":
# array - read operation
# write the base address onto the stack
segment, index = self._resolve_symbol(t)
self.vm_writer.write_push(segment, index)
# compile the offset expression
self.tokenizer.advance()
self.compile_expression()
# add base and offset
self.vm_writer.write_arithmetic("add")
# put the resulting address into pointer 1 (that)
self.vm_writer.write_pop(self.vm_writer.POINTER, 1)
# read from that 0 onto the stack
self.vm_writer.write_push(self.vm_writer.THAT, 0)
# closing bracket
tt, t = self._token_next(False, "SYMBOL", "]")
# advance for the next statement
self.tokenizer.advance()
else:
# none of above - just a single identifier
segment, index = self._resolve_symbol(t)
self.vm_writer.write_push(segment, index)
else:
# unknown
print "WARNING: Unknown term expression object:", tt, t
# compiles a (possible empty) comma-separated list of expressions
def compile_expression_list(self):
num_args = 0
# check for empty list
tt, t = self._token_next(False)
if tt == "SYMBOL" and t == ")":
# the parameter list was empty; do not process any more
pass
else:
# there are things in the parameter list
while True:
# expression to pass
self.compile_expression()
num_args += 1
# possible comma
tt, t = self._token_next(False)
if tt == "SYMBOL" and t == ",":
self.tokenizer.advance()
else:
# not a comma; stop processing parameters
break
return num_args
# compiles a subroutine call
# two cases:
# - subroutineName(expressionList)
# - (class|var).subroutineName(expressionList)
def compile_subroutine_call(self):
# first part of name
tt, name1 = self._token_next(False, "IDENTIFIER")
# a dot and another name may exist, or it could be a parenthesis
name2 = None
tt, t = self._token_next(True)
if tt == "SYMBOL" and t == ".":
# the name after the dot
tt, name2 = self._token_next(True, "IDENTIFIER")
# advance so that we are on the parenthesis
self.tokenizer.advance()
# determine if this is a method call
# three possibilities
# - class.func() - function call
# - var.func() - method call
# - func() - method call on current object
if self.symbol_table.contains(name1):
method_call = True
local_call = False
elif name2 == None:
method_call = True
local_call = True
else:
method_call = False
# if a method call, push variable name1
# this a method call if the symbol table contains name1 and name2 exists
# OR name1 is a method in the current object
if method_call and local_call:
# push the current object onto the stack as a hidden argument
self.vm_writer.write_push(self.vm_writer.POINTER, 0)
elif method_call and not local_call:
# push the variable onto the stack as a hidden argument
segment, index = self._resolve_symbol(name1)
self.vm_writer.write_push(segment, index)
# opening parenthesis
tt, t = self._token_next(False, "SYMBOL", "(")
# expression list
self.tokenizer.advance()
num_args = self.compile_expression_list()
# closing parenthesis
tt, t = self._token_next(False, "SYMBOL", ")")
# write the call
if method_call and local_call:
# methd + <blank>
# get the name of the vm function to call
classname = self.class_name
vm_function_name = classname + "." + name1
# increase arguments by 1, since there is the hidden "this"
num_args += 1
# make the call
self.vm_writer.write_call(vm_function_name, num_args)
elif method_call and not local_call:
# variable name + method
# get the name of the vm function to call
classname = self.symbol_table.get(name1)[1]
vm_function_name = classname + "." + name2
# increase arguments by 1, since there is the hidden "this"
num_args += 1
# make the call
self.vm_writer.write_call(vm_function_name, num_args)
else:
# get the name of the vm function to call
vm_function_name = name1 + "." + name2
# make the call
self.vm_writer.write_call(vm_function_name, num_args)
self.tokenizer.advance()
# returns the token_type and token of the next token after advancing the
# tokenizer before reading if advance is True
def _token_next(self, advance=False, expected_type=None, expected_value=None):
# advance the tokenizer, if requested
if advance:
self.tokenizer.advance()
# get the token type and the token itself
token_type = self.tokenizer.token_type()
token = str(getattr(self.tokenizer, token_type.lower())())
if expected_type and token_type != expected_type:
print "WARNING: Type", token_type, "found; expected", expected_type
import traceback, sys
traceback.print_stack()
sys.exit(1)
if expected_value and token != expected_value:
print "WARNING: Value", token, "found; expected", expected_value
import traceback, sys
traceback.print_stack()
sys.exit(1)
return token_type, token
# convets a symbol table type into a segment type
def _type_to_segment(self, type):
if type == self.symbol_table.STATIC:
return self.vm_writer.STATIC
elif type == self.symbol_table.FIELD:
return self.vm_writer.THIS
elif type == self.symbol_table.ARG:
return self.vm_writer.ARG
elif type == self.symbol_table.VAR:
return self.vm_writer.LOCAL
else:
print "ERROR: Bad type %s" % (str(type))
# resolves the symbol from the symbol table
# the segment and index is returned as a 2-tuple
def _resolve_symbol(self, name):
kind, type, index = self.symbol_table.get(name)
return self._type_to_segment(kind), index
class CompilationEngine:
CLASS_VAR_DEC_KEYWORDS = ['static', 'field']
SUBROUTINE_DEC_KEYWORDS = ['constructor', 'function', 'method']
BINARY_OPERATOR_SYMBOLS = ['+', '-', '*', '/', '&', '|', '<', '>', '=']
UNARY_OPERATOR_SYMBOLS = ['-', '~']
BINARY_OPERATORS_TO_COMMAND = {
'+': 'add',
'-': 'sub',
'=': 'eq',
'>': 'gt',
'<': 'lt',
'&': 'and',
'|': 'or'
}
UNARY_OPERATORS_TO_COMMAND = {'-': 'neg', '~': 'not'}
TYPE_TO_TAG = {
'STRING_CONST': 'stringConstant',
'INT_CONST': 'integerConstant',
'KEYWORD': 'keyword',
'IDENTIFIER': 'identifier',
'SYMBOL': 'symbol'
}
SYMBOLS_TO_XML_CONVENTION = {
'<': '<',
'>': '>',
'&': '&',
'"': '"'
}
def __init__(self, input_file_path, vm_writer: VMWriter):
self.jack_tokenizer = JackTokenizer(input_file_path)
self.symbol_table = SymbolTable()
self.vm_writer = vm_writer
if self.jack_tokenizer.has_more_tokens():
self.compile_class()
def compile_class(self):
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
self.symbol_table.define(self.jack_tokenizer.identifier(), 'NONE',
'CLASS')
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
if self.jack_tokenizer.key_word(
) in CompilationEngine.CLASS_VAR_DEC_KEYWORDS:
self.compile_class_var_dec()
if self.jack_tokenizer.key_word(
) in CompilationEngine.SUBROUTINE_DEC_KEYWORDS:
self.compile_subroutine()
self.jack_tokenizer.advance()
self.vm_writer.close()
def write_token(self, token_name):
type_tag = CompilationEngine.TYPE_TO_TAG[
self.jack_tokenizer.token_type()]
self.output_file.write('<{0}> {1} </{0}>\n'.format(
type_tag, token_name))
self.jack_tokenizer.advance()
def compile_class_var_dec(self):
while self.jack_tokenizer.key_word(
) in CompilationEngine.CLASS_VAR_DEC_KEYWORDS:
kind = ''
if self.jack_tokenizer.key_word() == 'field':
kind = 'FIELD'
elif self.jack_tokenizer.key_word() == 'static':
kind = 'STATIC'
self.jack_tokenizer.advance()
field_type = self.get_type()
self.jack_tokenizer.advance()
self.symbol_table.define(self.jack_tokenizer.identifier(),
field_type, kind)
self.jack_tokenizer.advance()
while self.jack_tokenizer.symbol() != ';':
self.jack_tokenizer.advance()
self.symbol_table.define(self.jack_tokenizer.identifier(),
field_type, kind)
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
def write_type(self):
if self.jack_tokenizer.token_type() == 'KEYWORD':
self.write_token(self.jack_tokenizer.key_word())
elif self.jack_tokenizer.token_type() == 'IDENTIFIER':
self.write_token(self.jack_tokenizer.identifier())
def compile_subroutine(self):
self.vm_writer.zero_branching_indexes()
while self.jack_tokenizer.key_word(
) in CompilationEngine.SUBROUTINE_DEC_KEYWORDS:
self.symbol_table.start_subroutine()
constructor = True if self.jack_tokenizer.key_word(
) == 'constructor' else False
method = False
if self.jack_tokenizer.key_word() == 'method':
method = True
self.symbol_table.define('this',
self.symbol_table.get_class_name(),
'ARG')
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
self.symbol_table.define(self.jack_tokenizer.identifier(), 'NONE',
'SUBROUTINE')
name = self.symbol_table.get_class_name(
) + '.' + self.jack_tokenizer.identifier()
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
self.compile_parameter_list()
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
var_num = 0
while self.jack_tokenizer.key_word() == 'var':
var_num += self.compile_var_dec()
self.vm_writer.write_function(name, var_num)
if method:
self.vm_writer.write_push('ARG', 0)
self.vm_writer.write_pop('POINTER', 0)
elif constructor:
field_count = self.symbol_table.var_count('FIELD')
self.vm_writer.write_push('CONST', field_count)
self.vm_writer.write_call('Memory.alloc', 1)
self.vm_writer.write_pop('POINTER', 0)
self.compile_statements()
self.jack_tokenizer.advance()
def compile_parameter_list(self):
if self.jack_tokenizer.symbol() != ')':
parameter_type = self.get_type()
self.jack_tokenizer.advance()
self.symbol_table.define(self.jack_tokenizer.identifier(),
parameter_type, 'ARG')
self.jack_tokenizer.advance()
while self.jack_tokenizer.symbol() == ",":
self.jack_tokenizer.advance()
parameter_type = self.get_type()
self.jack_tokenizer.advance()
self.symbol_table.define(self.jack_tokenizer.identifier(),
parameter_type, 'ARG')
self.jack_tokenizer.advance()
def get_type(self):
if self.jack_tokenizer.token_type() == 'KEYWORD':
parameter_type = self.jack_tokenizer.key_word()
elif self.jack_tokenizer.token_type() == 'IDENTIFIER':
parameter_type = self.jack_tokenizer.identifier()
return parameter_type
def compile_var_dec(self):
var_num = 1
self.jack_tokenizer.advance()
var_type = self.get_type()
self.jack_tokenizer.advance()
self.symbol_table.define(self.jack_tokenizer.identifier(), var_type,
'VAR')
self.jack_tokenizer.advance()
while self.jack_tokenizer.symbol() == ",":
var_num += 1
self.jack_tokenizer.advance()
self.symbol_table.define(self.jack_tokenizer.identifier(),
var_type, 'VAR')
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
return var_num
def compile_statements(self):
while self.jack_tokenizer.token_type() == 'KEYWORD':
if self.jack_tokenizer.key_word() == 'let':
self.compile_let()
elif self.jack_tokenizer.key_word() == 'if':
self.compile_if()
elif self.jack_tokenizer.key_word() == 'while':
self.compile_while()
elif self.jack_tokenizer.key_word() == 'do':
self.compile_do()
elif self.jack_tokenizer.key_word() == 'return':
self.compile_return()
def compile_do(self):
self.jack_tokenizer.advance()
name = self.jack_tokenizer.identifier()
self.jack_tokenizer.advance()
self.compile_subroutine_call(name)
# must dispose of void function return value
self.vm_writer.write_pop('TEMP', 0)
self.jack_tokenizer.advance()
def compile_subroutine_call(self, prefix_call=''):
if self.jack_tokenizer.symbol() == '(':
subroutine = False
# If not in symbol table - then subroutine
if not self.symbol_table.kind_of(
prefix_call) or self.symbol_table.kind_of(
prefix_call) == 'SUBROUTINE':
subroutine = True
self.jack_tokenizer.advance()
args_count = 0
if subroutine:
self.vm_writer.write_push('POINTER', 0)
args_count += 1
args_count += self.compile_expression_list()
if subroutine:
self.vm_writer.write_call(
self.symbol_table.get_class_name() + '.' + prefix_call,
args_count)
else:
self.vm_writer.write_call(prefix_call, args_count)
self.jack_tokenizer.advance()
elif self.jack_tokenizer.symbol() == '.':
variable = False
self.jack_tokenizer.advance()
if self.symbol_table.kind_of(prefix_call) in ['VAR', 'FIELD']:
variable = True
variable_name = prefix_call
prefix_call = self.symbol_table.type_of(prefix_call)
prefix_call += '.{0}'.format(self.jack_tokenizer.identifier())
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
args_count = 0
if variable:
self.vm_writer.write_push(
self.symbol_table.kind_of(variable_name),
self.symbol_table.index_of(variable_name))
args_count += 1
args_count += self.compile_expression_list()
self.vm_writer.write_call(prefix_call, args_count)
self.jack_tokenizer.advance()
def compile_let(self):
self.jack_tokenizer.advance()
var_name = self.jack_tokenizer.identifier()
self.jack_tokenizer.advance()
if self.jack_tokenizer.symbol() == '[':
self.vm_writer.write_push(self.symbol_table.kind_of(var_name),
self.symbol_table.index_of(var_name))
self.jack_tokenizer.advance()
self.compile_expression()
self.vm_writer.write_arithmetic("add")
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
self.compile_expression()
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.jack_tokenizer.advance()
self.compile_expression()
self.vm_writer.write_pop(self.symbol_table.kind_of(var_name),
self.symbol_table.index_of(var_name))
self.jack_tokenizer.advance()
def compile_while(self):
while_idx = self.vm_writer.get_next_label_index('while')
if_label = 'WHILE_IF_{0}'.format(while_idx)
end_label = 'WHILE_END_{0}'.format(while_idx)
self.vm_writer.write_label(if_label)
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
self.compile_expression()
self.vm_writer.write_arithmetic('not')
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
self.vm_writer.write_if(end_label)
self.compile_statements()
self.vm_writer.write_goto(if_label)
self.jack_tokenizer.advance()
self.vm_writer.write_label(end_label)
def compile_return(self):
self.jack_tokenizer.advance()
if self.jack_tokenizer.symbol() != ';':
self.compile_expression()
else:
self.vm_writer.write_push('CONST', 0)
self.vm_writer.write_return()
self.jack_tokenizer.advance()
def compile_if(self):
if_idx = self.vm_writer.get_next_label_index('if')
else_label = 'IF_ELSE_{0}'.format(if_idx)
end_label = 'IF_END_{0}'.format(if_idx)
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
self.compile_expression()
self.vm_writer.write_arithmetic('not')
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
self.vm_writer.write_if(else_label)
self.compile_statements()
self.jack_tokenizer.advance()
self.vm_writer.write_goto(end_label)
self.vm_writer.write_label(else_label)
if self.jack_tokenizer.key_word() == 'else':
self.jack_tokenizer.advance()
self.jack_tokenizer.advance()
self.compile_statements()
self.jack_tokenizer.advance()
self.vm_writer.write_label(end_label)
def compile_expression(self):
self.compile_term()
while self.jack_tokenizer.symbol(
) in CompilationEngine.BINARY_OPERATOR_SYMBOLS:
symbol = self.jack_tokenizer.symbol()
self.jack_tokenizer.advance()
self.compile_term()
if symbol in self.BINARY_OPERATORS_TO_COMMAND:
self.vm_writer.write_arithmetic(
self.BINARY_OPERATORS_TO_COMMAND[symbol])
elif symbol == '*':
self.vm_writer.write_call('Math.multiply', 2)
elif symbol == '/':
self.vm_writer.write_call('Math.divide', 2)
def compile_term(self):
token_type = self.jack_tokenizer.token_type()
if token_type == 'IDENTIFIER':
name = self.jack_tokenizer.identifier()
self.jack_tokenizer.advance()
if self.jack_tokenizer.symbol(
) == '(' or self.jack_tokenizer.symbol() == '.':
self.compile_subroutine_call(name)
elif self.jack_tokenizer.symbol() == '[':
self.vm_writer.write_push(self.symbol_table.kind_of(name),
self.symbol_table.index_of(name))
self.jack_tokenizer.advance()
self.compile_expression()
self.jack_tokenizer.advance()
self.vm_writer.write_arithmetic('add')
self.vm_writer.write_pop('POINTER', 1)
self.vm_writer.write_push('THAT', 0)
else:
kind = self.symbol_table.kind_of(name)
idx = self.symbol_table.index_of(name)
self.vm_writer.write_push(kind, idx)
elif token_type == 'STRING_CONST':
string_const = self.jack_tokenizer.string_val()
self.vm_writer.write_push("CONST", len(string_const))
self.vm_writer.write_call("String.new", 1)
for char in string_const:
self.vm_writer.write_push('CONST', ord(char))
self.vm_writer.write_call("String.appendChar", 2)
self.jack_tokenizer.advance()
elif token_type == 'KEYWORD':
keyword = self.jack_tokenizer.key_word()
if keyword == 'true':
self.vm_writer.write_push('CONST', 1)
self.vm_writer.write_arithmetic('neg')
elif keyword == 'false' or keyword == 'null':
self.vm_writer.write_push('CONST', 0)
elif keyword == 'this':
self.vm_writer.write_push('POINTER', 0)
self.jack_tokenizer.advance()
elif token_type == 'SYMBOL':
if self.jack_tokenizer.symbol() == '(':
self.jack_tokenizer.advance()
self.compile_expression()
self.jack_tokenizer.advance()
elif self.jack_tokenizer.symbol(
) in CompilationEngine.UNARY_OPERATOR_SYMBOLS:
command = CompilationEngine.UNARY_OPERATORS_TO_COMMAND[
self.jack_tokenizer.symbol()]
self.jack_tokenizer.advance()
self.compile_term()
self.vm_writer.write_arithmetic(command)
elif token_type == 'INT_CONST':
self.vm_writer.write_push('CONST', self.jack_tokenizer.int_val())
self.jack_tokenizer.advance()
def compile_expression_list(self):
expression_count = 0
if self.jack_tokenizer.symbol() != ')':
self.compile_expression()
expression_count += 1
while self.jack_tokenizer.symbol() == ',':
self.jack_tokenizer.advance()
self.compile_expression()
expression_count += 1
return expression_count
class CompilationEngine(object):
# the destination file for writing
destination_file = None
# the tokenizer for the input file
tokenizer = None
# current indentation level
indent = 0
# the constructor for compiling a single class
# the next method to be called after construction must be compile_class
# source_filename must be a single file, not a directory
def __init__(self, source_filename):
# destination filename
# if the original extension was .jack, then make the extension T.xml
# if the original extension was not .jack, then append T.xml
if source_filename.lower().endswith(".jack"):
destination_filename = source_filename[:-5] + ".xml"
else:
destination_filename = source_filename + ".xml"
# open the destination filename for writing
self.destination_file = open(destination_filename, 'w')
# create a tokenizer for the input file
self.tokenizer = JackTokenizer(source_filename)
# compiles a complete class and closes the output file
def compile_class(self):
# start the class
self._start_block("class")
# class
tt, t = self._token_next(True, "KEYWORD", "class")
self._write(tt, t)
# name of class
tt, t = self._token_next(True, "IDENTIFIER")
self._write(tt, t)
# open brace
tt, t = self._token_next(True, "SYMBOL", "{")
self._write(tt, t)
# one or more variable declarations
self.tokenizer.advance()
while True:
tt, t = self._token_next(False)
if tt == "KEYWORD" and t in ["field", "static"]:
self.compile_class_var_dec()
else:
# stop trying to process variable declarations
break
# one or more subroutine declarations
while True:
tt, t = self._token_next(False)
if tt == "KEYWORD" and t in ["constructor", "function", "method"]:
self.compile_subroutine()
else:
# stop trying to process functions
break
# close brace
# do not advance because we already advanced upon exiting the last loop
tt, t = self._token_next(False, "SYMBOL", "}")
self._write(tt, t)
# end the class
self._end_block("class")
# close the output file
self.destination_file.close()
# compiles a static declaration or field declaration
def compile_class_var_dec(self):
# start variable declaration
self._start_block("classVarDec")
# compile the variable declaration
# False means don't print the tags
self.compile_var_dec(False)
# end variable declaration
self._end_block("classVarDec")
# compiles a complete method, function, or constructor
def compile_subroutine(self):
# start subroutine declaration
self._start_block("subroutineDec")
# constructor, function, or name keyword
tt, t = self._token_next(False, "KEYWORD")
self._write(tt, t)
# type of the return value
# can be either keyword (void) or an identifier (any type)
tt, t = self._token_next(True)
self._write(tt, t)
# name of the method/function/constructor
tt, t = self._token_next(True)
self._write(tt, t)
# opening parenthesis
tt, t = self._token_next(True, "SYMBOL", "(")
self._write(tt, t)
# arguments
self.tokenizer.advance()
self.compile_parameter_list()
# closing parenthesis
tt, t = self._token_next(False, "SYMBOL", ")")
self._write(tt, t)
# start body of subroutine
self._start_block("subroutineBody")
# opening brace
tt, t = self._token_next(True, "SYMBOL", "{")
self._write(tt, t)
# variable declarations
self.tokenizer.advance()
while True:
tt, t = self._token_next(False)
if tt == "KEYWORD" and t == "var":
self.compile_var_dec()
else:
# stop trying to process variable declarations
break
# statements
self.compile_statements()
# closing brace
tt, t = self._token_next(False, "SYMBOL", "}")
self._write(tt, t)
# end body of subroutine
self._end_block("subroutineBody")
# finish subroutine declaration
self._end_block("subroutineDec")
self.tokenizer.advance()
# compiles a (possibly empty) parameter list, not including the enclosing
# parentheses
def compile_parameter_list(self):
self._start_block("parameterList")
# check for empty list
tt, t = self._token_next(False)
if tt == "SYMBOL" and t == ")":
# the parameter list was empty; do not process any more
pass
else:
# there are things in the parameter list
while True:
# keyword (variable type)
tt, t = self._token_next(False)
self._write(tt, t)
# identifier (variable name)
tt, t = self._token_next(True)
self._write(tt, t)
# possible comma
tt, t = self._token_next(True)
if tt == "SYMBOL" and t == ",":
self._write(tt, t)
else:
# not a comma; stop processing parameters
break
self.tokenizer.advance()
self._end_block("parameterList")
# compiles a var declaration
def compile_var_dec(self, print_tags=True):
if print_tags:
self._start_block("varDec")
# the keyword to start the declaration
tt, t = self._token_next(False, "KEYWORD")
self._write(tt, t)
# type of the declaration
# could be an identifier or a keyword (int, etc)
tt, t = self._token_next(True)
self._write(tt, t)
# name of the declaration
tt, t = self._token_next(True, "IDENTIFIER")
self._write(tt, t)
# can support more than one identifier name, to declare more than one
# variable, separated by commas; process the 2nd-infinite variables
self.tokenizer.advance()
while True:
tt, t = self._token_next(False)
if tt == "SYMBOL" and t == ",":
# write the comma
self._write(tt, t)
# another variable name follows
tt, t = self._token_next(True, "IDENTIFIER")
self._write(tt, t)
self.tokenizer.advance()
else:
# no more variable names
break
# should be on the semicolon at the end of the line
tt, t = self._token_next(False, "SYMBOL", ";")
self._write(tt, t)
self.tokenizer.advance()
if print_tags:
self._end_block("varDec")
# compiles a sequence of statements, not including the enclosing {}
def compile_statements(self):
self._start_block("statements")
while True:
tt, t = self._token_next(False)
if tt == "KEYWORD" and t in ["do", "let", "while", "return", "if"]:
# call compile_t, where t is the type of compilation we want
token = getattr(self, "compile_" + t)()
else:
# not a statement; stop processing statements
break
self._end_block("statements")
# compiles a do statement
def compile_do(self):
self._start_block("doStatement")
# do keyword
tt, t = self._token_next(False, "KEYWORD", "do")
self._write(tt, t)
# subroutine call
self.tokenizer.advance()
self.compile_subroutine_call()
# semicolon
tt, t = self._token_next(False, "SYMBOL", ";")
self._write(tt, t)
self._end_block("doStatement")
self.tokenizer.advance()
# compiles a let statement
def compile_let(self):
self._start_block("letStatement")
# let keyword
tt, t = self._token_next(False, "KEYWORD", "let")
self._write(tt, t)
# variable name
tt, t = self._token_next(True, "IDENTIFIER")
self._write(tt, t)
# possible brackets for array
tt, t = self._token_next(True)
if tt == "SYMBOL" and t == "[":
# write bracket
self._write(tt, t)
# compile the expression
self.tokenizer.advance()
self.compile_expression()
# closing bracket
tt, t = self._token_next(False, "SYMBOL", "]")
self._write(tt, t)
# advance to the next token, since we are expected to be on the = for
# the next line
self.tokenizer.advance()
# equals sign
tt, t = self._token_next(False, "SYMBOL", "=")
self._write(tt, t)
# expression
self.tokenizer.advance()
self.compile_expression()
# semicolon
tt, t = self._token_next(False, "SYMBOL", ";")
self._write(tt, t)
self._end_block("letStatement")
self.tokenizer.advance()
# compiles a while statement
def compile_while(self):
self._start_block("whileStatement")
# while keyword
tt, t = self._token_next(False, "KEYWORD", "while")
self._write(tt, t)
# opening parenthesis
tt, t = self._token_next(True, "SYMBOL", "(")
self._write(tt, t)
# the expression that is the condition of the while statement
self.tokenizer.advance()
self.compile_expression()
# the closing parenthesis
tt, t = self._token_next(False, "SYMBOL", ")")
self._write(tt, t)
# the opening brace
tt, t = self._token_next(True, "SYMBOL", "{")
self._write(tt, t)
# the statments that is the body of the while loop
self.tokenizer.advance()
self.compile_statements()
# the closing brace
tt, t = self._token_next(False, "SYMBOL", "}")
self._write(tt, t)
self._end_block("whileStatement")
self.tokenizer.advance()
# compiles a return statement
def compile_return(self):
self._start_block("returnStatement")
# return keyword
tt, t = self._token_next(False, "KEYWORD", "return")
self._write(tt, t)
# possible expression to return
tt, t = self._token_next(True)
if tt != "SYMBOL" and t != ";":
self.compile_expression()
# ending semicolon
tt, t = self._token_next(False, "SYMBOL", ";")
self._write(tt, t)
self._end_block("returnStatement")
self.tokenizer.advance()
# compiles a if statement, including a possible trailing else clause
def compile_if(self):
self._start_block("ifStatement")
# if keyword
tt, t = self._token_next(False, "KEYWORD", "if")
self._write(tt, t)
# opening parenthesis
tt, t = self._token_next(True, "SYMBOL", "(")
self._write(tt, t)
# expression of if statement
self.tokenizer.advance()
self.compile_expression()
# closing parenthesis
tt, t = self._token_next(False, "SYMBOL", ")")
self._write(tt, t)
# opening brace
tt, t = self._token_next(True, "SYMBOL", "{")
self._write(tt, t)
# statements
self.tokenizer.advance()
self.compile_statements()
# closing brace
tt, t = self._token_next(False, "SYMBOL", "}")
self._write(tt, t)
tt, t = self._token_next(True)
if tt == "KEYWORD" and t == "else":
# else statement exists
# write else
seld._write(tt, t)
# opening brace
tt, t = self._token_next(False, "SYMBOL", "{")
self._write(tt, t)
# statements
self.compile_statements()
# closing brace
tt, t = self._token_next(False, "SYMBOL", "}")
self._write(tt, t)
# advance tokenizer only if we are in the else, since otherwise the
# token was advanced by the else check
self.tokenizer.advance()
self._end_block("ifStatement")
# compiles an expression (one or more terms connected by operators)
def compile_expression(self):
self._start_block("expression")
# the first term
self.compile_term()
# finish any number of operators followed by terms
while True:
tt, t = self._token_next(False)
if tt == "SYMBOL" and t in "+-*/&|<>=":
# found an operator
self._write(tt, t)
# the next term
self.tokenizer.advance()
self.compile_term()
else:
# no term found; done parsing the expression
break
self._end_block("expression")
# compiles a term
# this routine is faced with a slight difficulty when trying to decide
# between some of the alternative parsing rules. specifically, if the
# current token is an identifier, the routine must distinguish between a
# variable, an array entry, and a subroutine call. a single lookahead token,
# which may be one of [, (, or ., suffices to distinguish between the three
# possibilities. any other token is not part of this term and should not
# be advanced over.
def compile_term(self):
self._start_block("term")
# a term: integer_constant | string_constant | keyword_constant |
# varname | varname[expression] | subroutine_call | (expression) |
# unary_op term
tt, t = self._token_next(False)
if tt in ["INT_CONST", "STRING_CONST", "KEYWORD"]:
self._write(tt, t)
# advance for the next statement
self.tokenizer.advance()
elif tt == "SYMBOL" and t == "(":
# ( expression )
# write the opening parenthesis
self._write(tt, t)
# parse the expression
self.tokenizer.advance()
self.compile_expression()
# closing parenthesis
tt, t = self._token_next(False, "SYMBOL", ")")
self._write(tt, t)
# advance for the next statement
self.tokenizer.advance()
elif tt == "SYMBOL" and t in "-~":
# unary_op term
# write the unary operation
self._write(tt, t)
# parse the rest of the term
self.tokenizer.advance()
self.compile_term()
elif tt == "IDENTIFIER":
# varname, varname[expression], subroutine_call
# do not write the identiifer yet
# get the next bit of the expression
# if it is a [, then array; if it is a ( or ., then subroutine call
# if none of above, then pass over
tt2, t2 = self._token_next(True)
if tt2 == "SYMBOL" and t2 in "(.":
# subroutine call
# back up and then compile the subroutine call
self.tokenizer.retreat()
self.compile_subroutine_call()
elif tt2 == "SYMBOL" and t2 == "[":
# array
# write identifier
self._write(tt, t)
# write bracket
self._write(tt2, t2)
# compile the expression
self.tokenizer.advance()
self.compile_expression()
# closing bracket
tt, t = self._token_next(False, "SYMBOL", "]")
self._write(tt, t)
# advance for the next statement
self.tokenizer.advance()
else:
# none of above - just a single identifier
self._write(tt, t)
else:
# unknown
print "WARNING: Unknown term expression object:", tt, t
self._end_block("term")
# compiles a (possible empty) comma-separated list of expressions
def compile_expression_list(self):
self._start_block("expressionList")
# check for empty list
tt, t = self._token_next(False)
if tt == "SYMBOL" and t == ")":
# the parameter list was empty; do not process any more
pass
else:
# there are things in the parameter list
while True:
# expression to pass
self.compile_expression()
# possible comma
tt, t = self._token_next(False)
if tt == "SYMBOL" and t == ",":
self._write(tt, t)
self.tokenizer.advance()
else:
# not a comma; stop processing parameters
break
self._end_block("expressionList")
# compiles a subroutine call
# two cases:
# - subroutineName(expressionList)
# - (class|var).subroutineName(expressionList)
def compile_subroutine_call(self):
# first part of name
tt, t = self._token_next(False, "IDENTIFIER")
self._write(tt, t)
# a dot and another name may exist, or it could be a parenthesis
tt, t = self._token_next(True)
if tt == "SYMBOL" and t == ".":
self._write(tt, t)
# the name after the dot
tt, t = self._token_next(True, "IDENTIFIER")
self._write(tt, t)
# advance so that we are on the parenthesis
self.tokenizer.advance()
# opening parenthesis
tt, t = self._token_next(False, "SYMBOL", "(")
self._write(tt, t)
# expression list
self.tokenizer.advance()
self.compile_expression_list()
# closing parenthesis
tt, t = self._token_next(False, "SYMBOL", ")")
self._write(tt, t)
self.tokenizer.advance()
# returns the token_type and token of the next token after advancing the
# tokenizer before reading if advance is True
def _token_next(self, advance=False, expected_type=None, expected_value=None):
# advance the tokenizer, if requested
if advance:
self.tokenizer.advance()
# get the token type and the token itself
token_type = self.tokenizer.token_type()
token = str(getattr(self.tokenizer, token_type.lower())())
if expected_type and token_type != expected_type:
print "WARNING: Type", token_type, "found; expected", expected_type
import traceback, sys
traceback.print_stack()
sys.exit(1)
if expected_value and token != expected_value:
print "WARNING: Value", token, "found; expected", expected_value
import traceback, sys
traceback.print_stack()
sys.exit(1)
return token_type, token
# writes the given token to the output file
def _write(self, token_type, token):
# lowercase for tag name
token_type = token_type.lower()
# special types
token_type = token_type.replace("int_const", "integerConstant")
token_type = token_type.replace("string_const", "stringConstant")
# special values to replace for output
s = {"<": "<", ">": ">", '"': """, "&": "&"}
for s, r in s.iteritems():
token = token.replace(s, r)
# print the token type and token to the file
output = ['<', token_type, '>', ' ', token, ' ', '</', token_type,
'>', '\n']
self.destination_file.write(self._indent("".join(output)))
# starts an XML block
def _start_block(self, block_name):
self.destination_file.write(self._indent("<" + block_name + ">\n"))
self.indent += 2
# ends an XML block
def _end_block(self, block_name):
self.indent -= 2
self.destination_file.write(self._indent("</" + block_name + ">\n"))
# indents a single line of text at the current indentation level
def _indent(self, text):
return " " * self.indent + text
class CompilationEngine:
CLASS_VAR_DEC_KEYWORDS = ['static', 'field']
SUBROUTINE_DEC_KEYWORDS = ['constructor', 'function', 'method']
BINARY_OPERATOR_SYMBOLS = ['+', '-', '*', '/', '&', '|', '<', '>', '=']
UNARY_OPERATOR_SYMBOLS = ['-', '~']
TYPE_TO_TAG = {'STRING_CONST': 'stringConstant', 'INT_CONST': 'integerConstant', 'KEYWORD': 'keyword',
'IDENTIFIER': 'identifier', 'SYMBOL': 'symbol'}
SYMBOLS_TO_XML_CONVENTION = {'<': '<', '>': '>', '&': '&', '"': '"'}
def __init__(self, input_file_path, output_file_path):
self.output_file = open(output_file_path, 'w')
self.jack_tokenizer = JackTokenizer(input_file_path)
if self.jack_tokenizer.has_more_tokens():
self.compile_class()
def compile_class(self):
self.output_file.write('<class>\n') # get first token
self.jack_tokenizer.advance()
self.write_token(self.jack_tokenizer.key_word())
self.write_token(self.jack_tokenizer.identifier())
self.write_token(self.jack_tokenizer.symbol())
if self.jack_tokenizer.key_word() in CompilationEngine.CLASS_VAR_DEC_KEYWORDS:
self.compile_class_var_dec()
if self.jack_tokenizer.key_word() in CompilationEngine.SUBROUTINE_DEC_KEYWORDS:
self.compile_subroutine()
self.write_token(self.jack_tokenizer.symbol())
self.output_file.write('</class>')
self.output_file.close()
def write_token(self, token_name):
type_tag = CompilationEngine.TYPE_TO_TAG[self.jack_tokenizer.token_type()]
self.output_file.write('<{0}> {1} </{0}>\n'.format(
type_tag, token_name))
self.jack_tokenizer.advance()
def compile_class_var_dec(self):
while self.jack_tokenizer.key_word() in CompilationEngine.CLASS_VAR_DEC_KEYWORDS:
self.output_file.write('<classVarDec>\n')
self.write_token(self.jack_tokenizer.key_word())
self.write_type()
self.write_token(self.jack_tokenizer.identifier())
while self.jack_tokenizer.symbol() != ';':
self.write_token(self.jack_tokenizer.symbol())
self.write_token(self.jack_tokenizer.identifier())
self.write_token(self.jack_tokenizer.symbol())
self.output_file.write('</classVarDec>\n')
def write_type(self):
if self.jack_tokenizer.token_type() == 'KEYWORD':
self.write_token(self.jack_tokenizer.key_word())
elif self.jack_tokenizer.token_type() == 'IDENTIFIER':
self.write_token(self.jack_tokenizer.identifier())
def compile_subroutine(self):
while self.jack_tokenizer.key_word() in CompilationEngine.SUBROUTINE_DEC_KEYWORDS:
self.output_file.write('<subroutineDec>\n')
self.write_token(self.jack_tokenizer.key_word())
self.write_type()
self.write_token(self.jack_tokenizer.identifier())
self.write_token(self.jack_tokenizer.symbol())
self.compile_parameter_list()
self.write_token(self.jack_tokenizer.symbol())
self.output_file.write('<subroutineBody>\n')
self.write_token(self.jack_tokenizer.symbol())
while self.jack_tokenizer.key_word() == 'var':
self.compile_var_dec()
self.compile_statements()
self.write_token(self.jack_tokenizer.symbol())
self.output_file.write('</subroutineBody>\n')
self.output_file.write('</subroutineDec>\n')
def compile_parameter_list(self):
self.output_file.write('<parameterList>\n')
if self.jack_tokenizer.symbol() != ')':
self.write_type()
self.write_token(self.jack_tokenizer.identifier())
while self.jack_tokenizer.symbol() == ",":
self.write_token(self.jack_tokenizer.symbol())
self.write_type()
self.write_token(self.jack_tokenizer.identifier())
self.output_file.write('</parameterList>\n')
def compile_var_dec(self):
self.output_file.write('<varDec>\n')
self.write_token(self.jack_tokenizer.key_word())
self.write_type()
self.write_token(self.jack_tokenizer.identifier())
while self.jack_tokenizer.symbol() == ",":
self.write_token(self.jack_tokenizer.symbol())
self.write_token(self.jack_tokenizer.identifier())
self.write_token(self.jack_tokenizer.symbol())
self.output_file.write('</varDec>\n')
def compile_statements(self):
self.output_file.write('<statements>\n')
while self.jack_tokenizer.token_type() == 'KEYWORD':
if self.jack_tokenizer.key_word() == 'let':
self.compile_let()
elif self.jack_tokenizer.key_word() == 'if':
self.compile_if()
elif self.jack_tokenizer.key_word() == 'while':
self.compile_while()
elif self.jack_tokenizer.key_word() == 'do':
self.compile_do()
elif self.jack_tokenizer.key_word() == 'return':
self.compile_return()
self.output_file.write('</statements>\n')
def compile_do(self):
self.output_file.write('<doStatement>\n')
self.write_token(self.jack_tokenizer.key_word())
self.write_token(self.jack_tokenizer.identifier())
self.compile_subroutine_call()
self.write_token(self.jack_tokenizer.symbol())
self.output_file.write('</doStatement>\n')
def compile_subroutine_call(self):
if self.jack_tokenizer.symbol() == '(':
self.write_token(self.jack_tokenizer.symbol())
self.compile_expression_list()
self.write_token(self.jack_tokenizer.symbol())
elif self.jack_tokenizer.symbol() == '.':
self.write_token(self.jack_tokenizer.symbol())
self.write_token(self.jack_tokenizer.identifier())
self.write_token(self.jack_tokenizer.symbol())
self.compile_expression_list()
self.write_token(self.jack_tokenizer.symbol())
def compile_let(self):
self.output_file.write('<letStatement>\n')
self.write_token(self.jack_tokenizer.key_word())
self.write_token(self.jack_tokenizer.identifier())
if self.jack_tokenizer.symbol() == '[':
self.write_token(self.jack_tokenizer.symbol())
self.compile_expression()
self.write_token(self.jack_tokenizer.symbol())
self.write_token(self.jack_tokenizer.symbol())
self.compile_expression()
self.write_token(self.jack_tokenizer.symbol())
self.output_file.write('</letStatement>\n')
def compile_while(self):
self.output_file.write('<whileStatement>\n')
self.write_token(self.jack_tokenizer.identifier())
self.write_token(self.jack_tokenizer.symbol())
self.compile_expression()
self.write_token(self.jack_tokenizer.symbol())
self.write_token(self.jack_tokenizer.symbol())
self.compile_statements()
self.write_token(self.jack_tokenizer.symbol())
self.output_file.write('</whileStatement>\n')
def compile_return(self):
self.output_file.write('<returnStatement>\n')
self.write_token(self.jack_tokenizer.key_word())
if self.jack_tokenizer.symbol() != ';':
self.compile_expression()
self.write_token(self.jack_tokenizer.symbol())
self.output_file.write('</returnStatement>\n')
def compile_if(self):
self.output_file.write('<ifStatement>\n')
self.write_token(self.jack_tokenizer.key_word())
self.write_token(self.jack_tokenizer.symbol())
self.compile_expression()
self.write_token(self.jack_tokenizer.symbol())
self.write_token(self.jack_tokenizer.symbol())
self.compile_statements()
self.write_token(self.jack_tokenizer.symbol())
if self.jack_tokenizer.key_word() == 'else':
self.write_token(self.jack_tokenizer.key_word())
self.write_token(self.jack_tokenizer.symbol())
self.compile_statements()
self.write_token(self.jack_tokenizer.symbol())
self.output_file.write('</ifStatement>\n')
def compile_expression(self):
self.output_file.write('<expression>\n')
self.compile_term()
while self.jack_tokenizer.symbol() in CompilationEngine.BINARY_OPERATOR_SYMBOLS:
symbol = self.jack_tokenizer.symbol()
if symbol in CompilationEngine.SYMBOLS_TO_XML_CONVENTION:
symbol = CompilationEngine.SYMBOLS_TO_XML_CONVENTION[symbol]
self.write_token(symbol)
self.compile_term()
self.output_file.write('</expression>\n')
def compile_term(self):
self.output_file.write('<term>\n')
token_type = self.jack_tokenizer.token_type()
if token_type == 'IDENTIFIER':
self.write_token(self.jack_tokenizer.identifier())
if self.jack_tokenizer.symbol() == '[':
self.write_token(self.jack_tokenizer.symbol())
self.compile_expression()
self.write_token(self.jack_tokenizer.symbol())
elif self.jack_tokenizer.symbol() == '(' or self.jack_tokenizer.symbol() == '.':
self.compile_subroutine_call()
elif token_type == 'STRING_CONST':
self.write_token(self.jack_tokenizer.string_val())
elif token_type == 'INT_CONST':
self.write_token(self.jack_tokenizer.int_val())
elif token_type == 'KEYWORD':
self.write_token(self.jack_tokenizer.key_word())
elif token_type == 'INT_CONST':
self.write_token(self.jack_tokenizer.int_val())
elif token_type == 'SYMBOL':
if self.jack_tokenizer.symbol() == '(':
self.write_token(self.jack_tokenizer.symbol())
self.compile_expression()
self.write_token(self.jack_tokenizer.symbol())
elif self.jack_tokenizer.symbol() in CompilationEngine.UNARY_OPERATOR_SYMBOLS:
self.write_token(self.jack_tokenizer.symbol())
self.compile_term()
self.output_file.write('</term>\n')
def compile_expression_list(self):
self.output_file.write('<expressionList>\n')
if self.jack_tokenizer.symbol() != ')':
self.compile_expression()
while self.jack_tokenizer.symbol() == ',':
self.write_token(self.jack_tokenizer.symbol())
self.compile_expression()
self.output_file.write('</expressionList>\n')
