*Motivation*

Trying to fix #3975

When a reset of a cursor is performed with some timestamp on a non-durable
consumer the message finder will fail with null pointer exception due to
`cursor.getName()` being null.

*Modifications*

  - Add method overloading for `newNonDurableCursor()` with subscription name.
  - Fix method getNonDurableSubscription to call `newNonDurableCursor()` with
    proper subscription name
  - Add test to assert issue.

### Motivation

Log Producer/Consumer Id when broker logs "Producer/consumer is already connected" to help in debugging when client is keep failing to create producer and broker is keep logging  the same error.

* Fixed conversion of unix timestamp to golang time (#3659)

* fix cannot use size (type _Ctype_int) as type _Ctype_ulong
Signed-off-by: Nxiaolong.ran <ranxiaolong716@gmail.com>

*Motivation*

AlwaysCompatible flag doesn't work in 2.3.1 because the flag is using NeverSchemaValidator

*Modifications*

Change to use AlwaysSchemaValidator

* Fixed issue with Authorization header missing after client gets redirected

* Use AsyncHttpClient instead

* Fixed merge issue

* Convert headers after body to not break multipart encoding

* Fixed tests

* Added missing property file for AsyncHttpClient in the shaded jar

* Moved ahc.properties into org.asynchttpclient.config

* One value was missing from properties file

* Also add default file for non-shaded scenario

* Another missing AHC config key

* Print topic load exception

* Removed constraint on unit tests executor

* Exposing getSchemaVersion in the client by making it public.

* Implemented getSchemaVersion in TopicMessageImpl.java

* Changed the release version

* fix submit function via url

* cleaning up

* add test

* make method private

* add additional tests

* cleaning up

* improving tests

### Motivation

Fix comments of schema compatibility strategy.

### Modifications

Change comment of SchemaCompatibilityStrategy.BACKWARD and SchemaCompatibilityStrategy.FORWARD

1) Collect input topics from the function details spec
2) Catch all errors during source/sink close since its user code

* By default use `journalFormatVersionToWrite=5` for Bookie journal

* Also set fileInfoFormatVersionToWrite=0

issue#3838 : Allow incompatible schemas to co-exist on a topic, allow set-schema-autoupdate-strategy = none (#3840)

Fixes #3838 : The set-schema-autoupdate-strategy policy currently allows disabling schema update (--disabled) and one of --compatibility(FULL, BACKWARD, FORWARD). There is no way to currently allow multiple schemas for a topic.

It would be useful to allow --compatibility so the schemas can be added without checking with the previous schemas. This allows for extra flexibility for instance having multiple objects on a topic, supports use cases like a simple event bus etc.

This change added tests and can be verified as follows:

  - *Added test cases under SchemaUpdateStrategyTest*

Fixes #3211
Fixes #2786 

*Motivation*

When a schema is deleted, the schema is not removed directly.
You can still fetch the latest schema but its state is marked as `deleted`.

So when we apply schema compatibility check, we should ignore deleted schema.

*Modifications*

Ignore deleted schema when doing schema compatibility check

* Since partitioned topics have a -partition-<partitionid> affixed to the topic name,
when doing explicit acking, check for the case to determine the right topic name

* added unittests

* Upgrade jackson version to 2.9.8

* Upgrade commons-collections version to 1.10

* Added ability to specify consumer queue size for function input topics

* fix and update PR

*Motivation*

Fixes #3925

We have 3 places of defining schema type enums. We kept adding
new schema type in pulsar-common. However we don't update the schema type
in wire protocol and schema storage.

This causes `SchemaType.NONE` is stored in SchemaRegistry.
It fails debeizum connector on restarting.

*Modifications*

Make sure all 3 places have consistent schema type definitions.
Record the correct schema type.

Fixes #3831

* Try both regular classloader as well as nar class loader for validating source/sinks

* Fixed test

* Fix unittest

* Added more comments to the code

* rename variables

* Wait for the create to succeed before updating. Otherwise there might be some reamnant producers

* Cleanup logic in JavaInstanceRunnable close method

* Added comments

In the POM file for tiered-storage-jcloud, the dependency on
com.jamesmurty.utils/java-xmlbuilder was marked with test scope, but in
fact this dependency is needed during normal use. This commit changes
the POM file so that the java-xmlbuilder dependency is packaged in the
NAR file.

### Motivation
Fix: unloading namespace logging

Add reference count feature to keep track of reused instances of a consumer
instance, for more details please see commit ff4db8db.

*Modifications*

  - Add refCount instance variable on ConsumerImpl.
  - Use new safeDecrRefCount() on consumer close() in order to know whether
    effective close call should occur or not.
  - Increment reference count when a previous built consumer instance is being
    used by caller.

*Future considerations*

Thereafter when feature preventing duplicated consumer is made for
PartitionedConsumer, MultiTopicsConsumer and PatternMultiTopicsConsumer,
incrRefCount() member could be turned into a pure virtual method.

## Motivation

To compromise avro‘s bug for avro filed logical type
https://issues.apache.org/jira/browse/AVRO-1891

## Modifications
add some initialize class

## Verifying this change
Add logical type test in AvroSchemaTest

*Motivation*

Schema version is set on single message's header. But it isn't propagated to
the batch message's header. So when the consumer receives the message batch,
it doesn't contain the schema version.

*Modifications*

When intializing a message batch, propagate the schema version from single message to batch message header.

*Verify this change*

Verified with an integration test. Need to write some unit tests for it.

* Fix NPE in ElasticSearchSink.

* Init by an empty string

* Enable generation of swagger definitions for functions/sources/sinks

* Added more changes to make it backwards compatible

* Add logic to server functions rest api

* Deleted swagger file since its supposed to be generated.
Also added some missing python packages for building

* Added back deleted file

* Fix comments

* Seperated out sources and sinks

* More changes to collapse all rest apis into one dropdown

### Motivation

Fix bug of reader.hasMessageAvailable().

```java
while (reader.hasMessageAvailable()) {
        Message msg = reader.readNext();
    // Do something
}
```

If lastDequeuedMessage with a batchMessageId, reader.hasMessageAvailable() will return false after read first message of batchMessage, because compared message id is a MessageIdImpl(lastMessageIdInBroker) 

### Modifications

Add batching message check.

### Verifying this change

- [x] This change added tests and can be verified as follows:
- *Added UT tests for read batch message and non-batch message*