[ZODB-Dev] Startup errors handled poorly

[Greg Ward]

Hi all --

using the zeod script distributed by Andrew from SourceForge (a modified
version of ZEO's start.py), I am experiencing some *intensely* annoying
bugs when zeod has problems starting up.  Mainly, this happens when it
can't write to its file (eg. I've just created, but forgotten to chgrp
and chmod g+w it), or if the daemon is already running.  Since this is a
development environment and I am only human, these both happen fairly
often.

The two problems, in order of increasing annoyance, are:
  * the error traceback happens in a child process post-forking,
    so it's scrambled with my shell prompt or anything else
    that might come on stdout after the original zeod process
    (the forking parent) exits

  * the error traceback is repeated ad infinitum -- in a forked
    child, so I can't just "Ctrl-C" it, or "kill %1" it -- I
    have to open another terminal window, "ps ... | grep" for it,
    and *then* kill it.  Arghh!!

Has anyone else experienced these problems, either with Andrew's zeod or
with DC's start.py?  Anyone have any clues how to fix them?  Or should I
do my bit of community service and spend a morning drilling in?

        Greg