Jamie

I think so long as it's opt-in then in worst case someone can revert back to case-sensitive routing if necessary. I wonder if adding additional options such as regex based...

I've faced the same issue with MsTest. The above step of adding ``` netstandard2.0 ```` worked for me to fix this. I'm surprised more people aren't faced with this issue...

Huh. Never ever occured to me that .NET Standard can't be used for tests! Cheers!

Another advantage of this is that `jnr-unixsocket` currently pulls in an LGPL licensed dependency `jnr-posix`: ``` [INFO] +- com.datadoghq:java-dogstatsd-client:jar:2.8:compile [INFO] | \- com.github.jnr:jnr-unixsocket:jar:0.18:compile [INFO] | +- com.github.jnr:jnr-ffi:jar:2.1.4:compile [INFO] | |...

You are correct, messages in TCP are not guaranteed to remain in order in Hazel, that is intended behaviour though. It's because Hazel starts the next receive operation before invoking...

I changed the title just so it's a bit more reflective of the actual issue, I hope you don't mind. Jamie

Hi Rich, Sorry about the delay in reviewing this, my internet's been intermittent the last week! The changes all look good but my only concern is that some people may...

Hard to say if this is actually a bug or if it's good behaviour but I'd probably say it should return the actual port as you expect. I don't have...

It's not too bad, but it still indicates a problem somewhere that could always trip up lower client counts in the wrong circumstances!

Profiler won't help here really, just need to find the set of locks causing the deadlock (assuming it's a deadlock)