+1 on replacing with native client . Agree with Matteo on retaining the CGO client code for sometime though.
Regards, Shivji Kumar Jha http://www.shivjijha.com/ +91 8884075512 On Mon, Apr 13, 2020 at 7:49 PM xiaolong ran <ranxiaolong...@gmail.com> wrote: > Hello all: > > > Pulsar has provided [pulsar-client-go]( > https://github.com/apache/pulsar/tree/master/pulsar-client-go < > https://github.com/apache/pulsar/tree/master/pulsar-client-go>) > based on the CPP. However, CGO has many problems, which causes performance > to degrade > easily, makes it difficult to debug, and makes Pulsar hard to support the > Go community better. > Consequently, we rewrite the [Go Client]( > https://github.com/apache/pulsar-client-go < > https://github.com/apache/pulsar-client-go>) based on native > Go to replace the CGO client and allow Pulsar to better support the > applications and integrations > of the Go community. > > Currently, pulsar-client-go 0.1.0 is released with multiple features > (producer, consumer, reader, and so on). > Except for schema, pulsar-client-go 0.1.0 can entirely replace CGO > client, so may I suggest replacing > CGO client with pulsar-client-go 0.1.0 in the next major Pulsar release. > > If agree, we need to do the following tasks: > Update the Go client doc on the Pulsar website > Remove the related contents of CGo client and add the related contents of > CPP client in release validation > Delete the CGo client code from Pulsar repo > Celebrate after finishing this project > > -- > Thanks > Xiaolong Ran
