bdeggleston commented on code in PR #7:
URL: https://github.com/apache/cassandra-accord/pull/7#discussion_r1005874805


##########
accord-core/src/main/java/accord/local/Command.java:
##########
@@ -402,155 +518,250 @@ private void onChangeInternal(PartialCommand command)
             case Executed:
             case Applied:
             case Invalidated:
-                if (isUnableToApply())
-                {
-                    updatePredecessor(command);
-                    if (isWaitingOnCommit())
-                    {
-                        removeWaitingOnCommit(command);
-                    }
-                }
-                else
-                {
-                    command.removeListener(this);
-                }
-                maybeExecute(false);
+                updatePredecessor(command, false);
+                maybeExecute(safeStore, progressShard(safeStore), false, true);
                 break;
         }
     }
 
-    @Override
-    public void onChange(Command command)
-    {
-        onChangeInternal(command);
-    }
-
-    protected void postApply()
+    protected void postApply(SafeCommandStore safeStore)
     {
         logger.trace("{} applied, setting status to Applied and notifying 
listeners", txnId());
-        status(Applied);
-        notifyListeners();
+        setStatus(Applied);
+        notifyListeners(safeStore);
     }
 
-    private static Function<CommandStore, Void> callPostApply(TxnId txnId)
+    private static Function<SafeCommandStore, Void> callPostApply(TxnId txnId)
     {
-        return commandStore -> {
-            commandStore.command(txnId).postApply();
+        return safeStore -> {
+            safeStore.command(txnId).postApply(safeStore);
             return null;
         };
     }
 
-    protected Future<Void> apply()
+    protected Future<Void> apply(SafeCommandStore safeStore)
     {
         // important: we can't include a reference to *this* in the lambda, 
since the C* implementation may evict
         // the command instance from memory between now and the write 
completing (and post apply being called)
-        return writes().apply(commandStore()).flatMap(unused ->
-            commandStore().process(this, callPostApply(txnId()))
+        CommandStore unsafeStore = safeStore.commandStore();
+        return writes().apply(safeStore).flatMap(unused ->
+            unsafeStore.submit(this, callPostApply(txnId()))
         );
     }
 
-    public Future<Data> read(Keys scope)
+    public Future<Data> read(SafeCommandStore safeStore)
     {
-        return txn().read(this);
+        return partialTxn().read(safeStore, this);
     }
 
-    private Future<Void> maybeExecute(boolean notifyListenersOnNoop)
+    // TODO: maybe split into maybeExecute and maybeApply?
+    private boolean maybeExecute(SafeCommandStore safeStore, ProgressShard 
shard, boolean alwaysNotifyListeners, boolean notifyWaitingOn)
     {
         if (logger.isTraceEnabled())
-            logger.trace("{}: Maybe executing with status {}. Will notify 
listeners on noop: {}", txnId(), status(), notifyListenersOnNoop);
+            logger.trace("{}: Maybe executing with status {}. Will notify 
listeners on noop: {}", txnId(), status(), alwaysNotifyListeners);
 
         if (status() != Committed && status() != Executed)
         {
-            if (notifyListenersOnNoop) notifyListeners();
-            return Writes.SUCCESS;
+            if (alwaysNotifyListeners)
+                notifyListeners(safeStore);
+            return false;
         }
 
-        if (isUnableToApply())
+        if (isUnableToExecute())
         {
-            BlockedBy blockedBy = blockedBy();
-            if (blockedBy != null)
-            {
-                logger.trace("{}: not executing, blocked on {}", txnId(), 
blockedBy.command.txnId());
-                commandStore().progressLog().waiting(blockedBy.command, 
blockedBy.someKeys);
-                if (notifyListenersOnNoop) notifyListeners();
-                return Writes.SUCCESS;
-            }
-            assert !isWaitingOnApply();
+            if (alwaysNotifyListeners)
+                notifyListeners(safeStore);
+
+            if (notifyWaitingOn)
+                new NotifyWaitingOn(this).accept(safeStore);
+            return false;
         }
 
         switch (status())
         {
             case Committed:
                 // TODO: maintain distinct ReadyToRead and ReadyToWrite states
-                status(ReadyToExecute);
+                setStatus(ReadyToExecute);
                 logger.trace("{}: set to ReadyToExecute", txnId());
-                boolean isProgressShard = progressKey() != null && 
handles(txnId().epoch, progressKey());
-                commandStore().progressLog().readyToExecute(this, 
isProgressShard, isProgressShard && progressKey().equals(homeKey()));
-                notifyListeners();
+                safeStore.progressLog().readyToExecute(this, shard);
+                notifyListeners(safeStore);
                 break;
+
             case Executed:
-                logger.trace("{}: applying", txnId());
-                if (notifyListenersOnNoop) notifyListeners();
-                return apply();
+                if (executeRanges(safeStore, 
executeAt()).intersects(writes().keys, 
safeStore.commandStore()::hashIntersects))
+                {
+                    logger.trace("{}: applying", txnId());
+                    apply(safeStore);

Review Comment:
   I think we need to return the future/chain back to the caller here, and wait 
on it's result before responding to an Apply message. As implemented here, the 
apply is started, but we respond to the coordinator / progressLog with 
`ApplyOutcome.Success` before it has actually completed



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to