terry terry - 1 year ago 57
Java Question

prevent conflict with spymemcached

I'm trying to use spymemcached 2.6 with synchronization on update and I found the two following way to use it:

  • using CASMutation defining a CASMutator which is quite an invasive way to implement it, let see an example:

    public List<Item> addAnItem(final Item newItem) throws Exception {

    // This is how we modify a list when we find one in the cache.
    CASMutation<List<Item>> mutation = new CASMutation<List<Item>>() {

    // This is only invoked when a value actually exists.
    public List<Item> getNewValue(List<Item> current) {
    // Not strictly necessary if you specify the storage as
    // LinkedList (our initial value isn't), but I like to keep
    // things functional anyway, so I'm going to copy this list
    // first.
    LinkedList<Item> ll = new LinkedList<Item>(current);

    // If the list is already "full", pop one off the end.
    if(ll.size() > 10) {
    // Add mine first.

    return ll;

    // The initial value -- only used when there's no list stored under
    // the key.
    List<Item> initialValue=Collections.singletonList(newItem);

    // The mutator who'll do all the low-level stuff.
    CASMutator<List<Item>> mutator = new CASMutator<List<Item>>(client, transcoder);

    // This returns whatever value was successfully stored within the
    // cache -- either the initial list as above, or a mutated existing
    // one
    return mutator.cas("myKey", initialValue, 0, mutation);

or by using

cas(String key, long casId, Object value)

after have done:

gets(String key, Transcoder<T> tc)

The second one is really more simple, and I understand why I would use CASMutation...
I would be really pleased to get some feedback about the use of this couchbase client.

Answer Source

CASMutator/CASMutation captures the best practices and workflow so that the right thing is done for you.

Your counter-example looks more simple since you aren't saying what you'd actually do with those methods. The sample you posted above is showing a List being pulled out of memcached, a new item added to it, conditionally removing some items from it, and then putting it back. At least half of that text you posted you'd still need to write.

If you don't use CASMutator, you'll end up reinvent it, and that's not all that simple. This is what it does for you today:

public T cas(final String key, final T initial, int initialExp,
        final CASMutation<T> m) throws Exception {
    T rv=initial;

    boolean done=false;
    for(int i=0; !done && i<max; i++) {
        CASValue<T> casval=client.gets(key, transcoder);
        T current=null;
        // If there were a CAS value, check to see if it's compatible.
        if(casval != null) {
            T tmp = casval.getValue();
        // If we have anything mutate and CAS, else add.
        if(current != null) {
            // Declaring this impossible since the only way current can
            // be non-null is if casval was set.
            assert casval != null : "casval was null with a current value";

            // There are three possibilities here:
            //  1) It worked and we're done.
            //  2) It collided and we need to reload and try again.
            //  3) It disappeared between our fetch and our cas.
            // We're ignoring #3 because it's *extremely* unlikely and the
            // behavior will be fine in this code -- we'll do another gets
            // and follow it up with either an add or another cas depending
            // on whether it exists the next time.
            if(client.cas(key, casval.getCas(), rv, transcoder)
                    == CASResponse.OK) {
        } else {
            // No value found, try an add.
            if(initial == null) {
                done = true;
                rv = null;
            } else if(client.add(key, initialExp, initial, transcoder).get()) {
    if(!done) {
        throw new RuntimeException("Couldn't get a CAS in " + max
            + " attempts");

    return rv;