Returns a list of all previous instructions/arguments of the current point in the data flow graph.
backward_df() -> List[Point]
The function works similarly to Instruction.backward_df(); the main difference is that in case of Instruction.backward_df() the function will return backward dataflow point for every point in the instruction, while Var.backward_df() returns only those connected with the current Var.
Like all other dataflow (DF) functions, it returns a list/set of Points, which can be instructions or "points" in the code, such as arguments, variables, etc.
Example
from glider import *
def query():
functions = Functions()\
.with_name('_transfer')\
.exec(1)
for function in functions:
for instruction in function.instructions().exec():
# print(instruction.source_code())
for operand in instruction.get_operands():
vars = operand.get_vars()
if len(vars)>0:
print("Var: "+vars[0].expression)
for point in vars[0].backward_df():
print(point.source_code())
return functions
Example Output
root":{3 items
"contract":string"0x31a13db822593aafef324ef2655e31b05b2577c0"
"contract_name":string"ERC20"
"sol_function":solidity
function _transfer(address sender,address recipient,uint256 amount) internal virtual {
require(sender != address(0),"ERC20: transfer from the zero address");
require(recipient != address(0),"ERC20: transfer to the zero address");
uint256 senderBalance = _balances[sender];
require(senderBalance >= amount,"ERC20: transfer amount exceeds balance");
_balances[sender] = senderBalance - amount;
_balances[recipient] += amount;
emit Transfer(sender,recipient,amount);
}
}
"root":{1 item
"print_output":[3 items
0:string"Var: sender"
1:string"require(sender != address(0),"ERC20: transfer from the zero address")"
2:string"address sender"
]
}
