[Bug 613001] Review Request: Heimdal - Alternative Kerberos implementation

[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=613001

--- Comment #1 from Andy Cobaugh <phalenor at gmail.com> 2010-09-16 16:27:49 EDT ---
I've been meaning to comment on this bug for a while now. Just getting back to
it.

I recently took the specfile that you came up with and modified it a bit to
hopefully bring it more inline with Fedora packaging guidelines, but more
importantly to turn it into something I was happy with distributing to our
machines here. On that note, I'm sure there are more things that can be done to
bring it up to par before inclusion in Fedora.

First, the SRPM used to build the packages we're currently distributing to our
clients:
ftp://ftp.bx.psu.edu/software/yum/centos/5/source/heimdal-1.3.3-3.bx.src.rpm

Offhand, a couple of things:
1) Heimdal is up to 1.4.1rc1, with a 1.4.1 release very soon according to Love.
2) Use %configure instead of ./configure
3) Change the way BuildRoot is defined, something like %(mktemp -ud
%{_tmppath}/%{name}-%{version}-%{release}-XXXXXX)
4) I'm not sure the Provides statements in the sub-packages are standard, or
even do anything? Someone else may comment on those.
5) I disabled make check entirely, as it was failing for me on Centos 5.5, and
I haven't had a chance to figure out why.
6) I redefine things like _bindir and _sbindir, among others, so that if we
should ever someday make heimdal the default, or if someone wants something
other than /usr/heimdal, changing that is as simple as changing %define hprefix
%{_prefix}/heimdal at the top. Handling the coexistance of Heimdal with MIT krb
is tricky either way, and I'm not sure if there's a pre-defined method of
dealing with this in Fedora. I don't think using _prefix in %files is the right
way to handle that (or rather, redefining _prefix).
7) I moved libexec to sbin. That's just a personal preference though.
8) The scriptlets (preun, postun, etc) should check $1 and modify their
behavior accordingly. 

Side note: Maybe we could talk to the MIT folks at some point and see if they
might be interested in using alternatives to handle the userland stuff like
kinit, klist, etc?

I'd be happy to discuss this further. I will probably be dropping my 1.4.1
specfile somewhere as soon as it's released.

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