[][src]Crate tx_padding

Padding and unpadding of messages with random prepended bytes and trailing zeros

This crate provides a padding scheme compatible with block-padding crate in which random bytes are prepended and zeros are appended.

For a message of length size, a buffer of length block_size * ((size + 1) / block_size) + 2 * block_size is required for padding. Let pad_len be (-size - 2) % block_size + 2. Apparently pad_len is the number of bytes to pad the message into multiple of block_size. The padding scheme appends pad_len + 1 bytes at the front of the message, where the lower log(block_size) bits of the first byte stores pad_len - 2 and the rest of the bits in the padding are random. At this point the message needs block_size - 1 more bytes to form multiple of block_size and we will just pad \0 at the end.

So TxPadding<N> comes with a type parameter N which specify the block size to use which is essential for unpadding. N must be a power of 2.

use tx_padding::{TxPadding, Padding};
use tx_padding::consts::{U8};

let msg = b"test";
let n = msg.len();
let mut buffer = [0xff; 16];
let padded_msg = TxPadding::<U8>::pad(&mut buffer, n, 8).unwrap();
assert_eq!(&padded_msg[5..], b"test\x00\x00\x00\x00\x00\x00\x00");
assert_eq!((padded_msg[0] & 0x7) + 2, 4);
assert_eq!(TxPadding::<U8>::unpad(&padded_msg).unwrap(), msg);
use tx_padding::{TxPadding, Padding};
use tx_padding::consts::{U8};
let mut buffer = [0xff; 8];
assert!(TxPadding::<U8>::pad(&mut buffer, 5, 8).is_err());

pad_block will always return PadError since it is not intended to be called. pad will return PadError if block_size > 511, block_size mismatch type parameter N or buffer is not sufficiently large, which is stricter than the requirement of the Padding trait.



Type aliases for many constants.



Error for indicating failed padding operation


Error for indicating failed unpadding operation





Trait for padding messages divided into blocks