Bug#40766: PROPOSED] Rewrite of "Configuration files" section
> On 18-Jul-99, 14:43 (CDT), Julian Gilbey <J.D.Gilbey@qmw.ac.uk> wrote:
> > > I'm against saying that "every conffile is a configuration file" simply
> > > because I don't want to lock out some other legitimate use of the
> > > conffile mechanism.
> >
> > The very nature of the conffile mechanism seems to be to protect the
> > sysadmin from losing locally configurations. Having a conffile which
> > does not have this purpose seems silly, at least at this point.
>
> Exactly, except that I would change the phrase "losing local
> configurations" to "losing local changes". If someone has package that
> needs to protect local changes to a file, the the conffile mechanism is
> a good way to do that, and I don't want people filing bugs against that
> package just because "file x in package y is not a configuration file,
> and the policy says all conffiles are configuration files."
Agreed.
> > If we allow for this possibility, then we should at least take care to
> > clarify what happens to conffiles in the later sections of the changed
> > text. (Can't remember which para off-hand....)
>
> Not sure what you mean by "...what happens to conffiles", but if you
> mean the details of how dpkg treats conffiles, then I don't want that in
> policy; there is a reference to the packaging manual already, and that
> should be sufficient.
True, hadn't thought of that ;)
One other question about the proposal. Is it necessary for multiple
packages which share a configuration file for one of them to specify
it as a conffile? Maybe it is a configuration file which by nature
cannot be a conffile?
[Please don't feel I'm trying to nitpick; I think you've done very
good, long overdue work on this section of policy. Thanks!]
Julian
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Julian Gilbey, Dept of Maths, QMW, Univ. of London. J.D.Gilbey@qmw.ac.uk
Debian GNU/Linux Developer, see http://www.debian.org/~jdg
Reply to: