Add occ todos / comments (#317)

## Describe your changes and provide context
This adds some comments with some useful code pointers for existing
logic and discussing future OCC work

## Testing performed to validate your change
NA

baseapp/abci.go

@@ -239,6 +239,8 @@ func (app *BaseApp) CheckTx(ctx context.Context, req *abci.RequestCheckTx) (*abc
 // Otherwise, the ResponseDeliverTx will contain releveant error information.
 // Regardless of tx execution outcome, the ResponseDeliverTx will contain relevant
 // gas execution context.
+// TODO: (occ) this is the function called from sei-chain to perform execution of a transaction.
+// We'd likely replace this with an execution task that is scheduled by the OCC scheduler
 func (app *BaseApp) DeliverTx(ctx sdk.Context, req abci.RequestDeliverTx) (res abci.ResponseDeliverTx) {
 	defer telemetry.MeasureSince(time.Now(), "abci", "deliver_tx")
 	defer func() {

baseapp/baseapp.go

@@ -821,6 +821,7 @@ func (app *BaseApp) getContextForTx(mode runTxMode, txBytes []byte) sdk.Context
 
 // cacheTxContext returns a new context based off of the provided context with
 // a branched multi-store.
+// TODO: (occ) This is an example of where we wrap the multistore with a cache multistore, and then return a modified context using that multistore
 func (app *BaseApp) cacheTxContext(ctx sdk.Context, txBytes []byte) (sdk.Context, sdk.CacheMultiStore) {
 	ms := ctx.MultiStore()
 	// TODO: https://github.com/cosmos/cosmos-sdk/issues/2824
@@ -974,6 +975,7 @@ func (app *BaseApp) runTx(ctx sdk.Context, mode runTxMode, txBytes []byte) (gInf
 			storeAccessOpEvents := msCache.GetEvents()
 			accessOps := ctx.TxMsgAccessOps()[acltypes.ANTE_MSG_INDEX]
 
+			// TODO: (occ) This is an example of where we do our current validation. Note that this validation operates on the declared dependencies for a TX / antehandler + the utilized dependencies, whereas the validation
 			missingAccessOps := ctx.MsgValidator().ValidateAccessOperations(accessOps, storeAccessOpEvents)
 			if len(missingAccessOps) != 0 {
 				for op := range missingAccessOps {
@@ -1118,6 +1120,8 @@ func (app *BaseApp) runMsgs(ctx sdk.Context, msgs []sdk.Msg, mode runTxMode) (*s
 		storeAccessOpEvents := msgMsCache.GetEvents()
 		accessOps := ctx.TxMsgAccessOps()[i]
 		missingAccessOps := ctx.MsgValidator().ValidateAccessOperations(accessOps, storeAccessOpEvents)
+		// TODO: (occ) This is where we are currently validating our per message dependencies,
+		// whereas validation will be done holistically based on the mvkv for OCC approach
 		if len(missingAccessOps) != 0 {
 			for op := range missingAccessOps {
 				ctx.Logger().Info((fmt.Sprintf("eventMsgName=%s Missing Access Operation:%s ", eventMsgName, op.String())))

store/cachekv/store.go

@@ -113,11 +113,13 @@ func (store *Store) Get(key []byte) (value []byte) {
 
 	cacheValue, ok := store.cache.Get(conv.UnsafeBytesToStr(key))
 	if !ok {
+		// TODO: (occ) This is an example of when we fall through when we dont have a cache hit. Similarly, for mvkv, we'll try to serve reads from a local cache thats transient to the TX, and if its NOT present, then we read through AND mark the access (along with the value that was read) for validation
 		value = store.parent.Get(key)
 		store.setCacheValue(key, value, false, false)
 	} else {
 		value = cacheValue.Value()
 	}
+	// TODO: (occ) This is an example of how we currently track accesses
 	store.eventManager.EmitResourceAccessReadEvent("get", store.storeKey, key, value)
 
 	return value
@@ -239,6 +241,8 @@ func (store *Store) iterator(start, end []byte, ascending bool) types.Iterator {
 	store.mtx.Lock()
 	defer store.mtx.Unlock()
 
+	// TODO: (occ) Note that for iterators, we'll need to have special handling (discussed in RFC) to ensure proper validation
+
 	var parent, cache types.Iterator
 
 	if ascending {