[Bug 226663] Merge Review: ypbind

[bugzilla at redhat.com]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.


https://bugzilla.redhat.com/show_bug.cgi?id=226663





--- Comment #9 from Vitezslav Crhonek <vcrhonek at redhat.com>  2008-11-21 05:56:39 EDT ---
To make this complete:

(In reply to comment #8)
> Thanks for working on this.  I checked out the current devel branch; it builds
> fine; rpmlint says:
> 
>   ypbind.src: W: strange-permission ypbind.init 0755
> I don't understand why rpmlint is complaining here.  This seems fine to me.

Seems fine to me too.

> 
>   ypbind.x86_64: W: incoherent-version-in-changelog 1.20.4-9
>    ['3:1.20.4-9.fc10', '3:1.20.4-9']
> Again, this seems quite OK.  I'm guessing that it is complaining about not
> seeing the epoch in the changelog version, but don't think we generally add it
> there.

I'll add epoch to changelog entries. I'm not sure if this must be, but it's
missing only last few commits, so let rpmlint be happy:)

> 
> So all of that looks bogus.  I assume that the OTHER_YPBIND_OPTS thing is
> something to be set in /etc/sysconfig/network instead of being edited into the
> initscript.

You're right, configuration should be done in /etc/sysconfig/network (and this
file is then included in init script).

> 
> The changes to the package look good; packaging-wise I have no complaints. 
> It's a bit odd seeing %{PACKAGE_VERSION} in the spec instead of %{version};
> I've never seen it before but it seems to work well enough.

Interesting, I missed it... I'll change it to %{version}, I have no knowledge
of %{PACKAGE_VERSION}. %{version} is common, let's use it.

> 
> Any idea why autoreconf is run?  There's been a bunch of discussion about
> whether this should ever be run in a package, and while I don't fully
> understand that discussion, I do think it would be good to ensure that the
> autoreconf call is really needed and to add comments to the spec as
> appropriate.  I note that rpmdiff shows only timestamp differences between a
> build that calls autoreconf and one that doesn't.

autoreconf is used to generate new configure script and Makefile when
configure.in/.ac, Makefile.in files are changed (patched). There's no patch
changing it at the moment (one patch is changing variable in "po" directory,
this is only one source of unsureness for me), so autoreconf call is IMHO not
necessary. autoconf and automake in BuildRequires is then not necessary too.
I'll discuss it rather yet, but it seems to be redundant.

Fedora relate info is here:
https://fedoraproject.org/wiki/PackagingDrafts/AutoConf

> 
> As for the initscript patch in comment #7, it seems correct on its face but
> it's a bit tough to read with only a non-context diff and I'm not really an
> expert with the whole LSB init comment block thing anyway.  Unfortunately I no
> longer have any vestige of my NIS infrastructure around so I can't test this at
> all.

OK, there's open bugzilla on init script issue, so I think there's no need to
resolve it here.

> 
> So really the only open issue I see is the autoreconf thing.

-- 
Configure bugmail: https://bugzilla.redhat.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.