Repeating pattern in encryption | Together C & C++ | Page 1

smoky pulsar Oct 19, 2022, 7:37 PM

#

Hey! I need some help with my first encryption algorithm.
I'm getting a repeating pattern after a while and i don't get why. Can anyone help me?

pearl chasmBOT Oct 19, 2022, 7:37 PM

#

When your question is answered use !solved to mark the question as resolved.

Remember to ask specific questions, provide necessary details, and reduce your question to its simplest form. For tips on how to ask a good question run !howto ask.

smoky pulsar Oct 19, 2022, 7:37 PM

#

    {
        int u = SL/2 + i;
        char L = S[i];
        char R = S[u];

        for (short int k = 0; k < 4; k++)
        {
            short int T = A[k][k + 1] + i;

            L ^= T + K;
            R ^= T + C(K);
        }
        L ^= C(K);
        R ^= A[0][1] * K;

        if (L == 0 || L == 10 || L == 32 || L == 8 || L == 7 || L == 9 || L == 13) L <<= 1;
        if (R == 0 || R == 10 || R == 32 || R == 8 || R == 7 || R == 9 || R == 13) R <<= 1;

        S[i] = L;
        S[u] = R;
    }```

#

this is for both encryption and decryption

#

{
    int B = A[0][X >> 32 & 8] + A[3][X >> 16 & 4];
    return (B ^ A[1][X >> 8 & 2]) >> 12;
}```

little nymph Oct 19, 2022, 7:42 PM

#

the real answer is to never roll your own encryption

smoky pulsar Oct 19, 2022, 7:43 PM

#

:(

little nymph Oct 19, 2022, 7:43 PM

#

I don't know what S and A are

smoky pulsar Oct 19, 2022, 7:44 PM

#

S is the string

#

        {07321,13908,86778,44329,53288,81847,82277,56246},
        {31701,87817,72960,23541,30606,72136,04600,79250},
        {87402,56425,34708,79089,13729,17228,28611,51591},
        {69054,65937,46161,93238,40363,89313,13643,27137}
};```

little nymph Oct 19, 2022, 7:45 PM

#

and you only read it down the minor diagonal, or the first two rows?

#

why

weary finch Oct 19, 2022, 7:45 PM

#

i wouldn't discourage it if it's just for fun, but if you're actually trying to encrypt something real, you should use an existing encryption algorithm because most likely yours will be cracked easily

little nymph Oct 19, 2022, 7:45 PM

#

I don't even think this is possible to decrypt

#

there are bitshifts

smoky pulsar Oct 19, 2022, 7:45 PM

#

the encryption algorithm works for decypriotn as well

smoky pulsar Oct 19, 2022, 7:46 PM

#

little nymph and you only read it down the minor diagonal, or the first two rows?

im not

#

A[k][k + 1]```

smoky pulsar Oct 19, 2022, 7:46 PM

#

weary finch i wouldn't discourage it if it's just for fun, but if you're actually trying to ...

i mean yeah mine is just for fun

little nymph Oct 19, 2022, 7:46 PM

#

is a diagonal

smoky pulsar Oct 19, 2022, 7:47 PM

#

oh yeah

#

i changed the code

#

before it was k + i % 8

little nymph Oct 19, 2022, 7:47 PM

#

it's not A[k][k] which is a major diagonal, but it is one off

#

so

#

If you get a repeating pattern in the output, I suspect it is because of a repeating pattern in the input

smoky pulsar Oct 19, 2022, 7:48 PM

#

yeah, but that shouldn't happen

little nymph Oct 19, 2022, 7:48 PM

#

why not?

smoky pulsar Oct 19, 2022, 7:48 PM

#

because that's not my goal

#

same words shoulnd't be the same in the encryption

little nymph Oct 19, 2022, 7:49 PM

#

by the way, if you actually want to prevent it from being read, you will need to have a key

smoky pulsar Oct 19, 2022, 7:49 PM

#

Input:
Hello world! Hello world! Hello world! Hello world!Hello world! Hello world! Hello world! Hello world!Hello world! Hello world! Hello world! Hello world!Hello world! Hello world! Hello world! Hello world!Hello world! Hello world! Hello world! Hello world!Hello world! Hello world! Hello world! Hello world!Hello world! Hello world! Hello world! Hello world!
Ouput:

ı~6ä[<Æ═åÉ§É²âÞ─^mZH⌂p6GÛ½òéµÞ╗ê‗)6:female_sign:§<mGXP¥ÐBë÷:diamonds:ı¼═/(~:arrow_up_down:Û¶ƒ\µú:ƒı■Â:diamonds:[─¥ü║¼Ðä¨þ:diamonds:↑☻a♫\ktZïÂù▒ÂÔýÎ∟#L☻{nOY▬zñƒ:diamonds:ã¡→ÏIá↓{╝t:arrow_up_down:┼
vÞ{ÞÔ'Íú)RB/X¥:relaxed:║¼Ð:diamonds:yþ:diamonds:↑éaµqÔ¶Y─ß☼¼(Û)V£#Lé¹¯O┘û·ñƒäF-→XI@↓{╝¶Æ┼÷Þ¹ÞÔ'V#)RBp/░ôê║»×Ëß·→►G`fqbtY─ßÅ¼¿Û:registered:Í∟#L☻{nOY▬zñƒ:diamonds:ã¡→Ï
Iá↓{╝t:arrow_up_down:┼```

little nymph Oct 19, 2022, 7:49 PM

#

otherwise someone reading the code will reverse engineer your algorithm

smoky pulsar Oct 19, 2022, 7:49 PM

#

i have a key

#

K is a key

little nymph Oct 19, 2022, 7:49 PM

#

K is a 32 bit int, so we're shooting for weak encryption I guess

smoky pulsar Oct 19, 2022, 7:50 PM

#

what?

little nymph Oct 19, 2022, 7:50 PM

#

I see

#

I don't understand how the bitshifts don't ruin everything

smoky pulsar Oct 19, 2022, 7:51 PM

#

yeah K is a 32 bit int

#

why would the bit shiftr uin everything?

little nymph Oct 19, 2022, 7:51 PM

#

I guess you've only tested it on ascii

smoky pulsar Oct 19, 2022, 7:51 PM

#

yeah i did

#

bit shift are inside a function which always run the same code with the same input

little nymph Oct 19, 2022, 7:52 PM

#

because if you reach the if statements with bad values of L, like, L = 10 and L = 20, the results are indistinguishable in the output

smoky pulsar Oct 19, 2022, 7:52 PM

#

and by shifting it by 12 bits i get a 2 digits integer

smoky pulsar Oct 19, 2022, 7:52 PM

#

little nymph because if you reach the if statements with bad values of L, like, L = 10 and L ...

oh thattt

#

i haven't tested it yet

#

but yeah, that probably does ruin things

#

guess im not gonna bit shift there tho

#

maybe add the key length to the character

little nymph Oct 19, 2022, 7:53 PM

#

I don't understand what the purpose of this is

smoky pulsar Oct 19, 2022, 7:53 PM

#

some of those characters are annoying

#

you think i should leave them in?

little nymph Oct 19, 2022, 7:54 PM

#

are you trying to protect real data from real adversaries? if so, your only bet is to go download a good C encryption library

smoky pulsar Oct 19, 2022, 7:54 PM

#

im not

weary finch Oct 19, 2022, 7:54 PM

#

smoky pulsar yeah, but that shouldn't happen

If you want this, you should make the previous bytes affect the encryption of the current plaintext

#

Instead of encrypting everything individually

little nymph Oct 19, 2022, 7:55 PM

#

instead of each one being a function of just i and the next 4 chars

#

yeah

smoky pulsar Oct 19, 2022, 7:55 PM

#

oh you're right supsm

#

wait what do you mean by previous bytes?

#

S[(i - 1) * (i > 0)]

little nymph Oct 19, 2022, 7:56 PM

#

that is only the immediate previous byte

#

but I guess that works

smoky pulsar Oct 19, 2022, 7:56 PM

#

what do you mean by "previous"

#

make an example

little nymph Oct 19, 2022, 7:57 PM

#

ie l, if you encrypt

aaaaaa z

#

or
bbbbbb z

#

should z correspond to the same output in the ciphertext?

#

if you take into account the previous part of the message, the answer would be no

smoky pulsar Oct 19, 2022, 7:59 PM

#

alright

#

thanks both of you :)

weary finch Oct 19, 2022, 8:00 PM

#

smoky pulsar what do you mean by "previous"

https://en.wikipedia.org/wiki/Block_cipher_mode_of_operation
see this, it has some stuff you can take inspiration from

smoky pulsar Oct 19, 2022, 8:01 PM

#

thanks! I will look into it

pearl chasmBOT Oct 19, 2022, 8:06 PM

#

@smoky pulsar Has your question been resolved? If so, run !solved :)

smoky pulsar Oct 19, 2022, 8:06 PM

#

shut the fuck up, never tell me what to do

#

!solved

pearl chasmBOT Oct 19, 2022, 8:06 PM

#

Thank you and let us know if you have any more questions!

sleek spoke Oct 19, 2022, 9:46 PM

#

do you realize you're yelling at a bot?

#Repeating pattern in encryption