Christophe de Vienne
Christophe de Vienne
This is a very good point
A global MarshalOptions, named JsonMarshalOptions, would be better so we can change any option, not only UseProtoNames. About the use_proto_names option it should not be a protoc flag because it...
Better indeed thanks. Passing the options to the base functions is not very elegant but avoiding this without having global variables requires a little more refactoring so we will keep...
- Setting 'UseProtoNames' changes the protocol, so it should be in the .proto files - We want a same message to always be encoded the same way, so it should...
Hello @zeeraw, You can use the package subject parameters to do that: https://github.com/nats-rpc/nrpc/blob/053d2d13751c00d8da297657f5550e14a1b1c55e/examples/alloptions/alloptions.proto#L8
@dev-mull could you give my PR a try? I think it covers your need.
@mdevan I would like your opinion on this before coding Thanks
Client side, I agree: the user should not have to handle the pause/resume. Server side, I think propagate the back-pressure should be the result of the 'send' function not returning...
The timeout is a good idea: very simple API, and a lot can be done without adding a complex mechanism. Will do that. I should be able to work on...
I was reading the go-nats sources and discovered that it already have a reception buffer, and we can get access its level at any time. It should be a good...