Cody Piersall

@chaoflow Which transport were you using when you saw these races? I'm seeing the GC race issue I was talking about only with the `inproc` transport. I'm going to do...

Thanks for opening an issue here—a month ago! Sorry for the lack of an update. Things have been busy. The problem I believe is the hacky solution I came up...

After a little more investigation, it turns out that `inproc` was a red herring: in my `ipc` tests, the sockets were never actually establishing a connection, so I wasn't seeing...

In a surprising turn of events, I was able to write a patch for nng and make the callbacks behave where they can *never* be called in order of `NNG_PIPE_EV_REM_POST`...

Soon I'll make a PR in nng with a [branch I made](https://github.com/codypiersall/nng/tree/pipe-notify-race) that fixes this issue. I think I should be able to get that upstream. If my fix gets...

I'm hoping to open an issue in nng, but haven't had time to do it lately. I'm confident we'll come to a solution eventually, but not sure about the timeline.

Hi @dongdongzzz, That failing test I believe is due to the way that nng formats the sockaddr changing. I got the same results. I'm not convinced the branch I made...

So, I finally made a C reproducer that shows that there is a race condition in nng: https://github.com/nanomsg/nng/pull/961#issuecomment-534854985 That said, the race, as far as I can tell, can only...

If you can post more of the build log I *might* be able to help. I don't have a Mac so I can't test anything on my end. The chunk...

I think there was a PR that fixed this, if it's possible for you to retry from master I'd appreciate it. I don't have a Mac so can't actually test...