Mário Freitas

Please notice too that I had to remove the locking around `Channel.call` for `Channel.Confirm` because I applied the lock inside `Channel.call` while calling `Channel.send`. Therefore this PR also applies the...

I am sorry, it took me a while to figure it out there was a deadlock in the use of pipes with go1.1. The last commit should work around it.

> Note that this PR also seems to include #197. Yes. I left a comment regarding that. Please check: https://github.com/streadway/amqp/pull/196#issuecomment-202242870

Just committed one of the changes you asked and the rebase against master.

Please note that travis is failing because it is detecting a data race in connection.go that was not fixed by #210 against which I handled the merged conflict by using...

@DmitriyMV OK, so I ended up removing all #210 code which was just a subset of the problems I addressed originally with this PR. The "illusiveness" of the bugs in...

@DmitriyMV I thought the reusing of the channel could be a good explanation for what is happening, but when you look at `TestRepeatedChannelExceptionWithPublishAndMaxProcsIssue46`, the function only uses 1 connection, creates...

This issue is 1year old, but anyway I will post a comment about it. I was having the same problem some minutes ago, and I noticed this happens because of...

Ohhh.... that is not what I expected. Do you still have the full trace of the exception?

Standalone would be possible if we make a driver for MacOSX that is able to follow a few set of commands (send these mouse click/drag events to this window, send...