When executing a Pipeline script with a long-running loop that uses nested closures, a StackOverflowError occurs and the job aborts abruptly.

Reproduction steps:

1. Create a Pipeline job with this as the script:

   1. def finished = false
      def someMap = ['foo':'bar', 'baz':'spam']
      def someOtherMap = ['foo':'baz', 'baz':'bar']
      node {
        def incr = 0
        while (true) {
          someMap.each{ foo, bar ->
            sh "echo ${foo} ${bar} ${incr}"
            incr++
            someOtherMap.each{ baz, spam ->
              sh "echo ${baz} ${spam} ${incr}"
            }
          }
        }
      }

1. Run the job
2. Wait for about an hour. For me, the loop executes nearly 4000 times before the error. Maybe easier to reproduce on a smaller stack.

Attached is the (not-very-interesting) resultant stack trace.