Fix receive issue of the client connection

[mostafa]

This PR partially fulfills the requirements and issues raised in #32, #209 and #219. I tried three solutions, and one partially worked:

• The first solution is to try to read from the connection again inside the forever loop. The initial message, aka. startup message, didn't get through, as there is no metric to know when to run the second round of read. But if I check the size of the message, which in my tests was 105 bytes from an ErrorResponse message, and then run the second round of read, it'll process everything well. The issue with that is quite obvious, as the message size is definitely not a reliable metric.
• The second solution is to try to set a read deadline on the connection to a very small value, like 100ms, and then try the second call to read. This works, but only on small messages. To process large messages, I needed to increase the read deadline, which caused significant delays in message receive time. The problem with this is that I don't know the size of the message in advance to set and reset the deadline, which practically makes this a 🐔 and 🥚 problem.
• The last solution is to use channels, goroutines and select to wait on reads asynchronously, which didn't work as expected. That is, it doesn't necessarily solve the problem, it just complicates it.

I was finally able to solve this using select with 1ms of wait.

• The PostgreSQL listen/notify works as expected, which closes Handle bi-directional communication between server and client #32:
  
• The errors are now properly propagated, which partially closes Handle SQL cancel command and errors properly #209, yet the cancel command works as expected, but the prompt is not released:
  
  
• The connection close problem is still standing, so I'll leave it for future.