bump version to 1.3.4 (rabbitmq#63)

Co-authored-by: Daniele Palaia <[email protected]>

bump buildVersion for release (rabbitmq#61)

Co-authored-by: Daniele Palaia <[email protected]>

bump versioning to 1.3.2 (rabbitmq#58)

* bump versioning to 1.3.1

* fix script / gofmt fixes

Co-authored-by: Daniele Palaia <[email protected]>

These ones were the ones testing Open scenarios. The issue is that Op… (

rabbitmq#57)

* These ones were the ones testing Open scenarios. The issue is that Open and Close, rwc.Open and rwc.Close can at the same time write on:

c.allocator = newAllocator(1, c.Config.ChannelMax)
connection.go line 444 and
connection.go line 849

while shutdown is protected by the structure mutex m, OpenComplete() is not causing the race.

While it's not clear if the library should protect this eventuality, the tests are testing the Open function, so I think the close can be put in the main thread avoiding the race and not affecting the test validity

* adding t.Cleanup

Co-authored-by: Daniele Palaia <[email protected]>

Merge pull request rabbitmq#33 from benmoss/deferred-confirmations

Expose a method to enable out-of-order Publisher Confirms

Merge pull request rabbitmq#25 from benmoss/fix-confirms-deadlock

Fix deadlock between publishing and receiving confirms

Merge pull request rabbitmq#13 from pabigot/issue/20210830a

connection: clarify documented behavior of NotifyClose