Nextflow script showing how to repeat the execution of a process

feedback.nf

#!/bin/env nextflow

Channel 
  .from('abc')
  .map { tuple(0,it) }
  .set { foo }

Channel.create().set { feedback }

index = 0 
loop = feedback.take(3).map { tuple(index++,it) }

process bar {
  input: 
  set val(index), file(x) from foo .mix(loop) 
  output: 
  file q 
  
  script:
  if( index == 0 ) 
  """
  rev $x > q
  """
  
  else if( index >1 )
  """
  rev $x > q
  """
}

process baz {
  input: 
  file q from q 
  output: 
  file z into feedback
  file z into result
  
  """
  cat $q > z 
  echo 'hello' >> z 
  """ 
}

result.last().println { it.text }

The tricky part is line 11. The input for process bar is given by a channel that emits both the content of foo and feedback channels. The first provides the initial input while the second brings the output produced by the baz process. The take operator takes the first 3 items thus stopping the iteration after that.

This is a very useful example! Would this look similar in DSL2? I guess it could be used as a work around the fact that each process can only be called once?