Fix panic when scale down partitions

[wolfstudy]

Signed-off-by: xiaolongran [email protected]

Motivation

When the program is running, if the business is forced to delete certain sub partitions, the following error message will be caused, that is, old_partitions is greater than new_partitions, it looks like it is doing scale down partitions, and the current code logic only deals with the scenario of scale up partitions , So if the user is forced to delete some sub partitions, the following error will be encountered:

level=info msg="[Changed number of partitions in topic]" new_partitions=1 old_partitions=20 topic="persistent://pulsar-xxxxxxx/xxxx/gxxxxxxxx"

panic: runtime error: index out of range [1] with length 1

goroutine 166288 [running]:
github.com/apache/pulsar-client-go/pulsar.(*producer).internalCreatePartitionsProducers(0xc0070aa6e0, 0x0, 0x0)
        github.com/apache/pulsar-client-go/pulsar/producer_impl.go:194 +0x785
github.com/apache/pulsar-client-go/pulsar.(*producer).runBackgroundPartitionDiscovery.func1(0xc004167cd0, 0xc00559f5c0, 0xc006af6dc0, 0xc0070aa6e0)
       github.com/apache/pulsar-client-go/pulsar/producer_impl.go:152 +0xce
created by github.com/apache/pulsar-client-go/pulsar.(*producer).runBackgroundPartitionDiscovery
       github.com/apache/pulsar-client-go/pulsar/producer_impl.go:144 +0xcd

Modifications

Increase the processing logic of scale down partition

[wolfstudy]

@merlimat @cckellogg @zymap PTAL thanks