Improve code documentation

spec/helpers/tcp_gate_scheduler.rb

@@ -116,7 +116,7 @@ def read( transfer_until: )
  # Option `transfer_until` can be (higher to lower priority):
  # :eof => transfer until channel is closed
  # :wouldblock => transfer until no immediate data is available
- # IO object => transfer until IO is writeable
+ # IO object => transfer until IO is writable
  #
  # The method does nothing if a transfer is already pending, but might raise the transfer_until option, if the requested priority is higher than the pending transfer.
  def write( transfer_until: )
@@ -282,7 +282,7 @@ def io_wait(io, events, duration)
  conn.observed_fd = io.fileno
 
  if (events & IO::WRITABLE) > 0
- puts "write-trigger #{conn.write_fds} until #{io.fileno} writeable"
+ puts "write-trigger #{conn.write_fds} until #{io.fileno} writable"
  conn.write(transfer_until: io)
 
  if (events & IO::READABLE) > 0
@@ -292,10 +292,10 @@ def io_wait(io, events, duration)
  else
  if (events & IO::READABLE) > 0
  puts "write-trigger #{conn.write_fds} until wouldblock"
- # Many applications wait for writablility only in a would-block case.
- # Then we get no trigger although data was written to the observed IO.
- # After writing some data the caller usually waits for some answer to read.
- # We take this event as a trigger to transfer of all pending written data.
+ # libpq waits for writablility only in a would-block case and not before writing.
+ # Since our incoming IO probably doesn't block, we get no write-trigger although data was written to the observed IO.
+ # But after writing some data the caller usually waits for some answer to read.
+ # We take this event as a trigger to transfer all pending written data.
  conn.write(transfer_until: :wouldblock)
 
  puts "read-trigger #{conn.read_fds} single block"