Re: Two improvements to etcupdate(8)

[Edgar Fuß <ef%math.uni-bonn.de@localhost>]

> After some reflexion your case is different, etcupdate(8) is more
> oriented towards an admin that updates the system rather than the dev
> that updates etc/ because the OS evolved.
I don't get the difference between the two.

But in fact, my case is different: I have a (self-written) management 
system that takes care of most of the configuration files.
Now on the one hand, after an OS update, for technical reasons, for some of 
those (syslog.conf, newsyslog.conf, ...), it needs the pristine OS files 
installed before runnig while for most, it doesn't.
On the other hand, it doesn't manage some ``dangerous'' files (master.passwd, 
group, rc), I take care of those manually and want to make sure I don't 
accidently overwrite them on an interactive etcupdate run.

> However, what you propose is from my PoV dangerous, you could completely
> wreck the conf if you happen to play the wrong "autoanswer" file.
Yes, of course. However, my autoanswer file consists of a few ``m'' entries 
for the ``dangerous'' files (so I don't overwrite them by accident), very 
few ``i'' entries (for the special case mentioned above) and mostly ``d'' 
entries for the bulk of files taken care of by the management system.