[Bug 213832] Review Request: arpwatch - Network monitoring tools for tracking IP addresses on a network
bugzilla at redhat.com
bugzilla at redhat.com
Thu Nov 9 19:48:53 UTC 2006
Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug report.
Summary: Review Request: arpwatch - Network monitoring tools for tracking IP addresses on a network
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=213832
------- Additional Comments From tmraz at redhat.com 2006-11-09 14:48 EST -------
(In reply to comment #11)
> > The snmpwalk non-requirement looks to me like a non issue because user running
> > the arpfetch command will get a message snmpwalk not found if it is not there.
>
> That's what I call broken. A user running a script installed in the default
> PATH by a package should not get any error. Or it should be documented
> prominently.
>
> > And I wouldn't say that the script is non functional, it just requires
> > installation of another package. This is really only issue of aesthetics and I'd
> > like to leave that on Miroslav to decide.
>
> It's not aesthetic, it's poor packaging. Packaged software should work
> out of the box, or have things that won't work out of the box
> documented.
OK, if README.fedora in %doc would satisfy you I think that Miroslav should add
it. Or arpfetch could be moved to %doc.
> > The release number should be probably just a single number (+ disttag) for FC
> devel.
>
> If it is a pre-release it should be named according to
>
http://fedoraproject.org/wiki/Packaging/NamingGuidelines#head-d97a3f40b6dd9d2288206ac9bd8f1bf9b791b22a
>
> If it is a post release version, it is right as is, as seen here:
>
http://fedoraproject.org/wiki/Packaging/NamingGuidelines#head-18aa467fc6925455e44be682fd336667a17e8933
It is post release, I didn't comment on the version but on the release number
(1.1) - two part number shouldn't be used in devel without a reason.
> > The scripts aren't one-to-one copy of the scriptlets from guidelines but I don't
> > think it is mandatory to have one-to-one copy if they work the same.
>
> I agree on the principle, but I'd like to have some explanations.
> Is it true that they work the same? Is it sure that the exit 0 is enough
> to avoid any failure? Some snippets on the wiki page have ||:, is it
> unusefull?
I've actually tested the scriptlets and even if /sbin/service return nonzero
exit code the chkconfig runs OK so the || : is unnecessary.
> > As the things above are only minor nits and comments and the package is OK
> > otherwise I think I can call it ACCEPTED.
>
> One of my questions hasn't been answered. It is certainly not a blocker,
> but I think it also deserves an explanation (it may even be that it is
> the other possibility, ie doing useradd only for the first install which
> is wrong):
>
> * why isn't the useradd only done for the first install?
>
I'll leave this one on Miroslav to answer.
--
Configure bugmail: https://bugzilla.redhat.com/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
More information about the Fedora-package-review
mailing list