On Mon, Oct 25, 2004 at 04:56:16PM -0400, Jan Harkes wrote:
> > after startup was complete. This allows scripts to start venus and do 
> > something with the coda file system directly afterwards.
> > 
> > Is it not true that something similar is needed for the coda server?
> 
> Yes, I just hadn't gotten to doing that yet. Basically I was always

Hello Jan,

I have to issue a warning, daemonization is a sensitive thing.
I had to disable the current related code in Venus on Darwin as it misbehaves
otherwise.

AFAIK the only more-or-less portable daemonization is double fork
plus possibly some black magic...

Imho "&" semantics together with a sensible test for start completion
is more portable and quite reliable.
It adds also the liberty to have a timeout on start (which I am using)
and/or run some unrelated scripts in parallel.

My 2c,
--
Ivan