Re: the netbase/inetd conspiracy
Hi,
Well, rather belatedly, I am joining this thread. Here are my
opinions o this:
a) firstly, dpkg has always treated missing files as a user
configuration, and not installed it. If the upstream
conffile changes in the package, the user is asked.
i) ucf managed configuration files also have the same
behaviour (this is a minor point)
b) Programs ought to have sane defaults; this is also a
reflected in the policy dictum not to require environment
variables for functioning. The behaviour of the program in
absence of a configration file may or may not be the same
as that in presence of one. Indeed, this difference, if it
exists, can, and has, been exploited for some programs.
c) In the case in question, of the configuration file can be
resurrected in a semi automated fashion, the defaults used
are well known
d) If the defaults are not easily discernible, and the program
can't provide sane behaviour in the absence of a
configuration file, it should either:
i) Stash a sane configuration file in /usr/lib/blah
ii) Abort, perhaps with a diagnostic
e) telling people that user changes under /etc shall be
respected, as far as we can make it so, is a powerful
invariant, and one that users have come to depend on.
Given the above, it is my opinion that packages that do
resurrect a configuration file that a human has deliberately removed
are indeed broken, and need to be fixed.
manoj
--
Old programmers never die, they just branch to a new address.
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: