A Hash and Unhash

[tooltiperror]

So, I want to write a Hashing and Unhashing function. Basic idea:

• [ljass]function Hash[/ljass] takes a salt and a string, and creates a new string. takes strings 50 chars or shorter, and returns a 50 char or shorter string.
• [ljass]function Unhash[/ljass] takes the salt and a string, and only only returns the original string if the salt is correct.
• The salt is more like a password than an average 'salt'.
• Write a JASS version and a C# version so I can hash things and unhash them safely without deprotection meddling.

Any tips? Are there any good algorithms for this kind of things?

(also, my charset is as follows: "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890")

 

[GetTriggerUnit-]

Does the Hash as to return the same length as the value given?

 

[Xienoph]

From what you've explained, it's an encryption, not a hash (you cannot unhash. In fact, if you can, then that's a bad hash). Also, as you've pointed out, the salt is not an average salt. It's an encryption key. So let's call your functions "encrypt" and "decrypt".

One basic way to do this would be to use the key as a seed for a pseudo random number generator (PNG). Use the PNG to generate a sequence of characters with the same length as the input string. Final step is to XOR the randomly generated string and the input string.

For example, if the input (plain text) string is "trigger", generate the random sequence "qjfmamj". Next, look up the binary encoding of every letter. t is 0x74, and q is 0x71. If you XOR them, you'll get 0x05, or ENQ (Enquiry ... whatever that is). Do that for every pair of character, and you'll get the encrypted string. To decrypt, do exactly the same thing, but using the encrypted string instead of the plain text string.

You don't have to use ASCII for your charset. Just make sure that the size is a power of 2.

One big catch though ... make sure that you always change the key whenever you're encrypting a different plain text (guess why ... +rep if you figure it out).

PNG algorithms are generally very complex. But here's a simple one I found in Wikipedia.

 

[tooltiperror]

My only problem so far is that my compiler doesn't display [LJASS]0x41 xor 0x42[/LJASS] (or other any other xor'd characters I've tried)

#include <iostream>

using namespace std;

int main(int argc, char * argv[])
{
	char A = 0x41;
	char B = 0x42;
	char X = A xor B;
	cout << "X = " << X << "\n"; // "X =  "
	return EXIT_SUCCESS;
}

(I use ANSI Style, shoot me)

 

[GetTriggerUnit-]

char X = A xor B; <----

I used that and it worked like a charm.

#include <iostream>
using namespace std;

int main()
{
	char A = 65;
	char B = 66;
	short i = A xor B;
	cout << i << endl;
	cin.get();
	return 0;
}

It returned 3 with the above setup and it returned 0 with the A = 0 and B = 0 setup. I used MingW compiler that comes with the latest codeblocks (10.05)

 

[Xienoph]

Well, it doesn't appear since the code 3 in ASCII is a non printable character. The result of the xor should be treated as a binary data. If you want to print it out, you can print out in hexadecimal form (sprintf("%x", A xor B))

#include <iostream>
using namespace std;

int main()
{
	char A = 65;
	char B = 66;
	short i = A xor B;
	cout << i << endl;
	cin.get();
	return 0;
}

That works too. Just remember to add a space / other delimiters between each character, or pad each character, so you can tell how many digits each character has.