Phil Mayers wrote:

We're seeing very occasional segfaults in the detail listener. I've managed to catch one under gdb, and the backtrace shows it dying at detail.c:601:

if (feof(data->fp)) goto cleanup;

...because data->fp == NULL

I can't follow the control flow to see why this might happen, but I did gather a backtrace etc.

<shrug> The fix is easy. Check if it's NULL. Pushed to git. Alan DeKok.