republishing on reconnect bug in the example

[Boris-Plato]

Describe the bug

I found this your example on how to implement a beautiful reconnect logic. I assume that pending channel that we use for retry failed delivery on the next session should be moved out of sessions for loop f.e. to the line 93. I made a test simulating error on publishing and the message was not republished to the new session with the current version of an example. When I moved pending channel out of the sessions loop it worked fine. I can provide some code examples if it is required.

Reproduction steps

1. simulate 1 time error on publishing and call
   pending <- body
   pub.Close()
   break Publish
2. send a message and check if it was republished and correctly processed with the subscriber

Expected behavior

message republished to the new reconnected session

Additional context

No response

[lukebakken]

Hello, thanks for using this library

If you have a fix for the example, please submit a pull request to fix the example.

[Boris-Plato]

Here we go
https://github.com/rabbitmq/amqp091-go/pull/250/files