def _get_format_expression(self, lexer: JavaTokenizer):
token_type, token = lexer.peek()
if token_type != 'var':
self.log.info(f'Unexpected BOF token: {token_type} ({token})')
stack = []
while not lexer.eof():
token_type, token = lexer.peek()
if token_type == 'punc' and token == '(':
return stack
stack.append(token)
lexer.next()
def _read_variable(self, lexer: JavaTokenizer):
stack = []
variable_name = []
previous_was_var = False
while not lexer.eof():
token_type, token = lexer.peek()
if token_type == 'punc' and token == ',' and not stack:
return 'simple', variable_name, None
elif token_type == 'op' and token == '+' and not stack:
return 'simple', variable_name, None
elif token_type == 'var':
previous_was_var = True
# Function Call
elif token_type == 'punc' and token == '(' and previous_was_var:
previous_was_var = False
func_type, default_args = self._map_function(variable_name[-1])
# If it is another formatting call, follow it
if func_type in self._processing_map.keys():
_new_stream = Stream(lexer.input.s[lexer.input.pos - 1:])
_new_lexer = JavaTokenizer(_new_stream)
args = self._count_arguments(_new_lexer)
param_mapping = self._create_params_mapping(
default_args, args)
func = self._processing_map[func_type]
msg, variables = func(self, lexer, param_mapping)
return 'nested', msg, variables
# Elsewise we handle as normal expression bracket
else:
stack.append('(')
# Bracket expression
elif token_type == 'punc' and token == '(':
previous_was_var = False
stack.append('(')
elif token_type == 'punc' and token == ')':
previous_was_var = False
# If there's a closing bracket without an opening one it is not ours again and we release
if not stack:
return 'simple', variable_name, None
# In this case it is our bracket and we pop it
else:
stack.pop()
else:
previous_was_var = False
variable_name.append(token)
lexer.next()
raise ValueError('Unexpected EOF')
def _count_arguments(self, lexer: JavaTokenizer):
previous_token_type = None
previous_token = None
while not lexer.eof():
token_type, token = lexer.peek()
if token_type == 'punc' and token == '(':
previous_token_type = token_type
previous_token = token
lexer.next()
break
lexer.next()
if lexer.eof():
raise ValueError(f'Does not contain a function call')
stack = []
argument_count = 1
while not lexer.eof():
token_type, token = lexer.peek()
if token_type == 'punc' and token == ')' and not stack:
if previous_token_type and previous_token_type == 'punc' and previous_token and previous_token == '(':
return 0
return argument_count
elif token_type == 'punc' and token == '(':
stack.append('(')
elif token_type == 'punc' and token == ')':
stack.pop()
elif token_type == 'punc' and token == ',' and not stack:
argument_count += 1
previous_token_type = token_type
previous_token = token
lexer.next()
return argument_count
def _parse(self, inp) -> Tuple[str, List[str]]:
character_stream = Stream(inp)
lexer = JavaTokenizer(character_stream)
log_string = ""
arguments = []
argument_mode = False
first_token = True
# We basically have two modes: Parse string concatenation and parse arguments
# If we find a String.format we know what to look for. If we don't we assume that the first occurrence of ','
# is the delimiter between string concatenation and arguments for that string
while not lexer.eof():
current_type, current_token = lexer.peek()
if current_type == 'str':
if first_token:
argument_mode = True
log_string += current_token
lexer.next()
elif current_type == 'op' and current_token == '+':
lexer.next()
current_type, current_token = lexer.peek()
if current_type == 'str':
log_string += current_token
lexer.next()
elif current_type == 'op' and not lexer.is_unary_ops(
current_token):
raise ValueError(
f'Operator {current_token} may not follow a +')
elif current_type == 'op':
lexer.next()
elif current_type == 'punc' and not current_token == '(':
raise ValueError(f'"{current_token}" may not follow a +')
elif current_type == 'punc' and current_token == '(':
hints, _, string_only = self._read_expression(lexer)
if string_only:
pass
if argument_mode:
log_string += '{}'
arguments.append(hints[0])
else:
arguments.append(hints[0])
elif current_type == 'var':
variable = self._read_var(lexer)
if argument_mode:
log_string += '{}'
arguments.append(variable)
else:
arguments.append(variable)
elif current_type == 'punc' and current_token == ',':
argument_mode = False
lexer.next()
elif current_type == 'op' and lexer.is_unary_ops(current_token):
lexer.next()
elif current_type == 'var':
_, expression, _ = self._read_expression(lexer)
if 'String.format' in expression:
expression = expression.replace("String.format(", '')
expression = expression[:expression.rindex(')')]
tmp = self._parse(expression)
return tmp
# handle this here:
if argument_mode:
log_string += '{}'
else:
arguments.append(expression)
elif current_type == 'num':
dtype, value = self._check_number(current_token)
if argument_mode:
log_string += '{}'
arguments.append('{!Integer}' if dtype ==
'int' else '{!Float}')
else:
arguments.append('{!Integer}' if dtype ==
'int' else '{!Float}')
lexer.next()
elif current_type == 'punc' and current_token == '(':
hints, output, string_only = self._read_expression(lexer)
if string_only:
stream = JavaTokenizer(Stream(output))
constructed_token = ""
while not stream.eof():
if (token := stream.next())[0] == 'str':
constructed_token += token[1]
log_string += constructed_token
elif argument_mode:
log_string += '{}'
else:
arguments.append(hints[0])
else:
print(
f'Weird behavio for token {current_token}<{current_type}>')
lexer.next()
def _parse_format(self, lexer: JavaTokenizer, params: List[str]):
if not params:
raise ValueError(
"Trying to parse format without argument. Aborting...")
param_offset = 0
param_type = params[param_offset]
message = ''
variables = []
statement_stack = []
while not lexer.eof():
token_type, token = lexer.peek()
# Advance argument
if token_type == 'punc' and token == ',' and param_type != '...':
param_offset = self._increase_index(param_offset, len(params))
param_type = params[param_offset]
# If the argument shall be skipped, e.g. when it's the log level argument or so on.
elif param_type == 'skip':
pass
# New expression
elif token_type == 'punc' and token == '(':
statement_stack.append(token)
# Closing expression
elif token_type == 'punc' and token == ')':
statement_stack.pop()
# No expressions left
if not statement_stack:
break
# String literal
elif token_type == 'str' and param_type == '...':
variables.append(token)
elif token_type == 'str':
message += self._parse_format_string(token)
elif token_type == 'num' and param_type == 'str':
message += str(token)
elif token_type == 'num' and param_type == '...':
variables.append(token)
# Variable
elif token_type == 'var' or (token_type == 'op'
and lexer.is_unary_ops(token)):
# If it is not, handle as normal variable
var_type, tokens, arguments = self._read_variable(lexer)
if var_type == 'simple':
variables.append(''.join(tokens))
if param_type == 'str':
message += '{}'
if var_type == 'nested':
message += tokens
variables.append(arguments)
# Operator '+' on string
elif param_type == 'str' and token_type == 'op' and token == '+':
lexer.next()
token_type, token = lexer.peek()
if token_type == 'str':
message += token
elif token_type == 'var':
tmp_mode, tmp_message, tmp_variables = self._read_variable(
lexer)
if tmp_mode == 'simple' and tmp_message:
message += '{}'
variables += tmp_message
elif tmp_mode == 'nested':
pass
lexer.next()
self._parse_format_string(message)
return message, variables
def _read_expression(self, lexer: JavaTokenizer):
value_hint = []
stack = []
string_only = True
original_string = "" # current_token
while not lexer.eof():
next_type, next_token = lexer.peek()
if next_type == 'str':
original_string += f'"{next_token}"'
else:
original_string += next_token
if next_type == 'punc' and next_token == '(':
lexer.next()
stack.append(True)
elif next_type == 'punc' and next_token == ')':
lexer.next()
stack.pop()
elif next_type == 'punc' and next_token == ')':
lexer.next()
stack.pop()
string_only = False
elif not stack and (next_type == 'punc' and next_token == ','):
break
elif not stack and (next_type == 'punc' and next_token == ';'):
break
elif next_type == 'var':
value_hint.append(next_token)
lexer.next()
string_only = False
elif next_type == 'num':
dtype, value = self._check_number(next_token)
value_hint.append(dtype)
lexer.next()
string_only = False
elif next_type == 'str':
lexer.next()
else:
lexer.next()
string_only = False
# print(f'Expression parsed, I would suggest it be: {value_hint}')
return value_hint, original_string, string_only
def _read_var(self, lexer: JavaTokenizer):
initial_type, var_tokens = lexer.peek()
if initial_type != 'var':
raise ValueError(
'Called _read_var on a stream that\'s not pointing to a var')
stack = []
current_type, current_token = lexer.next()
while not lexer.eof():
next_type, next_token = lexer.peek()
if current_type == 'var' and (next_type == 'op'
and next_token == '.'):
var_tokens += next_token
lexer.next()
elif not stack and (current_type == 'var' and
(next_type == 'op' and next_token == '+')):
break
elif not stack and (current_type == 'var' and
(next_type == 'punc' and next_token == ',')):
break
elif next_type == 'punc' and next_token == '(':
lexer.next()
stack.append(True)
elif next_type == 'punc' and next_token == ')':
lexer.next()
stack.pop()
elif current_type != 'var' and next_type == 'var':
var_tokens += next_token
lexer.next()
else:
var_tokens += next_token
lexer.next()
return var_tokens