hash function - int to char string + vice versa

Hi everyone,

So I am using a C model that I inherited from someone else (no documentation unfortunately). There's a C function (well its developed using the OPNET simulator so it's a sort of a C/OPNET hybrid) that takes a value convert its to a hash value (the dht_get_node_id function) and then that hash can be converted to an int value with the sha_get_int_id function.

Code:

---

// A bit more semantically-friendly name for a SHA-1 hash.
typedef unsigned char* DhtT_Key_Bytes

DhtT_Key_Bytes
        dht_get_node_id(int val)

{
      ShaT_Ctx                        context;
      DhtT_Key_Bytes        key = OPC_NIL;

    FIN (dht_get_node_id(IpT_Address ip_addr_ptr));
                       

            sha_init(&context);
 
            // allocate room for the SHA-1 Hash

            key = (DhtT_Key_Bytes) op_prg_mem_alloc(20);
       
            // update() keeps a "running total" of the SHA-1, but since we're only hashing 32 bits, we only use it once

            sha_update(&context, (unsigned char*)&val, sizeof(int));

            // store the final SHA-1 as the DHT ID
            sha_final(key, &context);
 
                        printf ("\nThe value of my node ID is: ");
                        sha_print_char(key);

---

Code:

---

int sha_get_int_id(unsigned char digest[ShaC_Bytes])
{
        int k, runningTotal;
        char        temp;
       
        runningTotal = 0;
        temp = '2';
               
        for (k = 0; k < ShaC_Bytes; k++)
        {
                if (digest[k] == '\a')
                {
                        runningTotal += 7;
                        //printf("%c", temp);                       
                }
                else if(digest[k] != '\a')
                {
                        runningTotal += digest[k];
                        //printf("%c", digest[k]);       
                }
               
                if (!((k + 1) % 4))
                        printf(" ");
        }
       
        return runningTotal;       
}

---

So for example, from debugging, if I pass the value 5 into the function I get a hash ID of 221 'Y' and if I pass this to sha_get_int_id it gives me an integer ID of 3144.

My problem is that I need a way of converting an integer back into the appropriate hash ID e.g. converting 3144 back to 221 'Y'.

Can anyone give me any advice as to how I can go about this? I suppose what's confusing me is the whole hash/digest thing which I'm not sure I fully follow and whether it's even possible to convert back (because from my limited knowledge maybe that's the whole point that you can't convert back because this is used for encryption).

Many thanks in advance.

Quote:

---

Originally Posted by kerrymaid

Can anyone give me any advice as to how I can go about this? I suppose what's confusing me is the whole hash/digest thing which I'm not sure I fully follow and whether it's even possible to convert back (because from my limited knowledge maybe that's the whole point that you can't convert back because this is used for encryption).

---

No, you can't. However, this is not "because this is used for encryption" -- hashes can also be used to structure and sort data. The reason you cannot convert them back is because they are reductive (usually drastically so), meaning completely different messages could have the same hash. Eg, this produces a hash key from 0-12 for any char string:

Code:

---

#include <stdio.h>
#include <string.h>

int hash (char *str) {
        int r = 0, i, len = strlen(str);
        for (i=0;i<len;i++) r += str[i];
        return r%12;
}

int main(int argc, char *argv[]) {
        char eg[2][8] = { "word", "static" };
        printf("%s = %d\n%s = %d\n",eg[0],hash(eg[0]),eg[1],hash(eg[1]));
        return 0;
}

---

"word" and "static" have the same key.

The SHA1 hash is more complicated (to yield 2^64 possible keys, I think, as opposed to just 12). It's purpose is to ensure that a transmitted message has not been corrupted -- it IS NOT an encryption since the message itself cannot be produced only from the hash. In fact, it widely used for this purpose with unencytpred messages (eg, software packages). If the hash of the received message does not match the SHA1 it came with, the message/package was corrupted somewhere. If the corruption was intentional (but the corrupter could not rely on altering the hash associated with the package), you could attempt to make it match, however, this adds a significant complication (in fact, an impossible one unless you understand how the hash was done, and I would guess in the case of the SHA1 nearly impossible anyway, but I dunno the details of it).

It's used in cryptography for the same purpose (to help ensure the decoded message was not corrupted).

Hi MK27,

Thanks for the reply. I suspected as much (but wasn't sure exactly why) so thanks for the further explanation. I'll have to come up with a work around for what I need to do.

Cheers.