--------------Original Comment History----------------------------
Comment from = ScottKurz on 2015-11-02 15:04:11 +0000
It may be helpful for a reader/writer to know that the close()/open() is being driven via retry with rollback vs. "normal" open/close.
Perhaps it might also help to know that you're in "one at a time" mode though I haven't seen that really myself.
Was thinking that adding a flag to the step context ( possibly an enum of multiple states) could be helpful.
Would be interested to hear others' thoughts.
Comment from = ScottKurz on 2015-11-04 15:48:53 +0000
A similar idea (though probably deserving of its own issue), is allowing the processor and writer to know that the reader has read its last item ( has returned 'null')
Originally opened as bug 7375 by ScottKurz
--------------Original Comment History---------------------------- Comment from = ScottKurz on 2015-11-02 15:04:11 +0000
It may be helpful for a reader/writer to know that the close()/open() is being driven via retry with rollback vs. "normal" open/close.
Perhaps it might also help to know that you're in "one at a time" mode though I haven't seen that really myself.
Was thinking that adding a flag to the step context ( possibly an enum of multiple states) could be helpful.
Would be interested to hear others' thoughts.
Comment from = ScottKurz on 2015-11-04 15:48:53 +0000
A similar idea (though probably deserving of its own issue), is allowing the processor and writer to know that the reader has read its last item ( has returned 'null')