When users want to diagnose a task that failed on Swarming, it's not intuitive
how to try to reproduce the problem.
A good solution would be to have a way to reproduce a task locally directly
from the task itself. It's not hard, the idea is to add:
swarming.py reproduce -t <taskid>
It would do:
- Fetch the command from the TaskRequest.
- Run the commands locally.
It's fairly trivial to implement, given the right endpoint to retrieve the data
(see issue 118).
This saves from having to explain how run_isolated works at all.
Original issue reported on code.google.com by maruel@chromium.org on 3 Sep 2014 at 3:09
Original issue reported on code.google.com by
maruel@chromium.org
on 3 Sep 2014 at 3:09