On Jul 29, 12:47 pm, Dustin <[email protected]> wrote: > I'm looking at this. I've never heard of an issue here, and I'm not > entirely sure how to test for it proactively at the moment, but you > seem to have done quite a good job of investigating it. Thanks a lot.
One really dumb way that I was able to reproduce the error every time (in testing for the possible crash cause) was to update_event(c, EV_WRITE | EV_PERSIST) in drive_machine() at the conn_waiting case (before the EV_READ). The current event flags in the connection data type are set to EV_READ | EV_PERSIST, so this call gets past the _redundant request_ check in update_event(). Thanks for looking at this and I am looking forward to hearing your assessment. thom
