Re: dbootstrap now working, busybox needs serial con change.

[Eric Delaunay <delaunay@lix.polytechnique.fr>]

Matt Porter wrote:
> After a short delay to fight a virus, I squashed the last known (to me)
> dbootstrap bug.  I can now go through a full install successfully until
> reboot.
> 
> There is one small problem with busybox (a bug report for Erik here).  On
> a serial console system, busybox's log mechanism is not real friendly yet.
> It whenever the run() calls are used it logs the "pid ? exited." messages
> to ttyS0 which spams the running dbootstrap.  Since busybox already has
> primary and secondary ser console called out up front, it should log these
> kind of messages to the ttyS1 if it is available (yes, there are systems
> with only one serial port these days) or send them to /dev/null if only
> one serial port exists.  Alternatively, this could be a build option where
> in the case of dbootstrap we would junk the messages, other applications
> may want the primary serial console or VT console to be sent those
> messages.

Oops, it was introduced after the patch I posted.
Systems with more than one serial console are not so common I guess (I never
seen one btw).
We could either disable logging or better redirect it to a file.  Maybe logging
to /var/log/messages or /tmp/log ?  An advanced user could then see what
happens by reading this file.  Even adding an item in dbootstrap menu to show
the contents of this file would be best.

-- 
 Eric Delaunay                 | "La guerre justifie l'existence des militaires.
 delaunay@lix.polytechnique.fr | En les supprimant." Henri Jeanson (1900-1970)