Cut in:
1.About vec initialization(fixed, confusing capacity with length)
//panic code
let mut ext_key:Vec<u32>=Vec::with_capacity(44);
ext_key[0..4].copy_from_slice( &key[0..4]);
i intend to initialize ext_key with fixed length by key, which means ext_key has key at the first four words and ext_key has fixed length. Codes below would panic at ext_key[0..4], saying range end index 4 out of range for slice of length 0,which means ext_key 's length is 0.
However, it would fix if i use
let mut ext_key:Vec<u32>=vec![0;44];
ext_key[0..4].copy_from_slice( &key[0..4]);
don't know why.
2. iter of for(fixed, both pass,false alarm)
intend to create a loop in a reverse order.
//panic codes
for round in (1..=9).rev(){}
//pass
for round in (1..10).rev(){}
3.About performance using threads(it depends?)
In AES using CBC, conventionally, each block is en/decoded one by one. Also, AES is a computing process. These means AES may take long time to en/decoding if only occupying one thread. So, i'm thinking, would it be better if each block is assigned to a thread to en/decode? Or ,would there be a way to optimize using threads?
4.About closure performance
Muti-nested closure ,muti-nested function and muti-calling function ,which is better?Muti-nested closure is ungly, i know and i'm using.
5.Managing test
My project's tree
.
├── Cargo.lock
├── Cargo.toml
├── src
├── lib
│ └── aes.rs
├── lib.rs
└── main.rs
i wrote public test function in aes.rs, but i can't figure how to manage tests, especially when the test function need to access local function in aes.rs.
//things in lib.rs
pub mod lib {
pub mod aes;
}
// place for test funtion, haven't implemented
#[cfg(test)]
mod test{
#[test]
fn non(){
}
}
//aes.rs this is one test, don't know how to manage
pub fn unitied_test(mat:&mut Vec<u8>,key:&Vec<u32>){
//using private functions in aes.rs
}