Alex

Yes. I have already fixed the issue for our system using custom AddressResolver. IMO it would be better to fix it for everybody. There is no reason to have a...

I've taken a look at the .Net implementation for comparison. It appears that DefaultEndpointResolver which is the mirror class there does shuffle. /// The default value creates an instance of...

Hello Guys, Any update on this issue? We are using version 3.5.5. We have run into this when using large number of items with Mono timeout. Try the test code...

I understand that. It's just the patterns used with the primary channelMono and ResourceManagement channelMono are different in the Sender class. Primary channelMono supports ChannelPool and returns channels to the...

Actually ReentrantLock provides the memory barrier semantics, so #2 should be OK then.

If Channel(s) in the Java can not be used concurrently this should be really defined much better in the API documentation. In that case the issue moves to Reactor Rabbit...

I'll have to find out what the company policy is regarding contributing code to the open source projects. I think I need to get an explicit permission. Generally, even if...

Is there any hope for a progress on this issue? Something like AsyncContext.fail(cause)? I don't think Tomcat actually allows response.sendError on a committed response as mentioned above. The code looks...

Setting one bit in a 32 bit value sets them all, so it's not really possible to detect what bits were set and which were not consistently and "null" takes...