Re: we need more variation for conffiles
>>"Atsuhito" == Atsuhito Kohda <kohda@pm.tokushima-u.ac.jp> writes:
Atsuhito> there are cases in which it is better to replace files by default.
But these cases are version dependent; a file that in version
3.2 should not override the users changes, may need to override them
in version 3.2.1
Atsuhito> For example, devpts.sh contained the bug but it is conffiles
Atsuhito> so if a user answered with only pressing return key then
Atsuhito> it is not replaced with the corrected devpts.sh.
But, until the typo, not replacing by default would have been
correct behaviour -- and the next version, that would be correct
behaviour again.
Atsuhito> So I think that conffiles should be, at least, two kind and
Atsuhito> - one is not replaced by default (current conffiles)
Atsuhito> - other is replaced by default (new kind of conffiles)
How do you know, as you start to package, which is going to be
which? You should never be writing the new kind of conffiles,
anyway.
And, if you do know there is something that you need to
override the users wishes, you should *NOT& make it a conffile, just
have it included as default in the *REAL* conffile.
Frankly, I think that having a shell script be a conffile is a
bad design decision (the shell script should source the real
conffile), and the devpts fiasco is is living example of the major
reason why I think so.
manoj
--
Today's thrilling story has been brought to you by Mushies, the great
new cereal that gets soggy even without milk or cream. Join us soon
for more spectacular adventure starring... Tippy, the Wonder Dog!
Bob & Ray
Manoj Srivastava <srivasta@debian.org> <http://www.debian.org/%7Esrivasta/>
1024R/C7261095 print CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E
1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
Reply to: