cckellogg commented on a change in pull request #82: Added a message id tracker
for acking messages that are batched.
URL: https://github.com/apache/pulsar-client-go/pull/82#discussion_r342384600
##########
File path: pulsar/impl_message.go
##########
@@ -18,30 +18,31 @@
package pulsar
import (
+ "math/big"
+ "strings"
+ "sync"
"time"
- "github.com/apache/pulsar-client-go/pkg/pb"
"github.com/golang/protobuf/proto"
-)
-func earliestMessageID() MessageID {
- return newMessageID(-1, -1, -1, -1)
-}
+ "github.com/apache/pulsar-client-go/pkg/pb"
+)
type messageID struct {
ledgerID int64
entryID int64
batchIdx int
partitionIdx int
+
+ tracker *ackTracker
Review comment:
To me it looks like this design will cause us always have to check for two
different message ids types and then handle them in a different way. Is that
correct?
```
switch v := i.(type) {
case *messageID:
// handle non batch
case *msgTracker:
// handle batch
```
@wolfstudy how do you see your suggested design working when the client
sends back a MessageID interface that we need to ack?
The general goal of this patch was to make ack checking in the code simple
whether the message was batched or not
```
func (c *consumer) Ack(msgID MessageID) error {
if msgID.(*messageID).ack() {
// message id can be acked send to server
}
}
```
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
