Function computeDatasetDigest

src/debug.cpp:295–329

Compute the dataset digest. Since keys, sets elements, hashes elements * are not ordered, we use a trick: every aggregate digest is the xor * of the digests of their elements. This way the order will not change * the result. For list instead we use a feedback entering the output digest * as input in order to ensure that a different ordered list will result in * a different digest. */

 * as input in order to ensure that a different ordered list will result in
 * a different digest. */
void computeDatasetDigest(unsigned char *final) {
    int j;
    uint32_t aux;

    memset(final,0,20); /* Start with a clean result */

    for (j = 0; j < cserver.dbnum; j++) {
        redisDb *db = g_pserver->db[j];

        if (db->size() == 0) continue;

        /* hash the DB id, so the same dataset moved in a different
         * DB will lead to a different digest */
        aux = htonl(j);
        mixDigest(final,&aux,sizeof(aux));

        /* Iterate this DB writing every entry */
        db->iterate_threadsafe([final](const char *key, robj_roptr o)->bool {
            unsigned char digest[20];
            robj *keyobj;

            memset(digest,0,20); /* This key-val digest */
            keyobj = createStringObject(key,sdslen(key));

            mixDigest(digest,key,sdslen(key));

            xorObjectDigest(digest,o);

            /* We can finally xor the key-val digest to the final digest */
            xorDigest(final,digest,20);
            decrRefCount(keyobj);
            return true;
        });
    }
}

#ifdef USE_JEMALLOC
void mallctl_int(client *c, robj **argv, int argc) {