Dr Andrew C Aitchison wrote:
On Fri, 10 Oct 2008, Troy Dawson wrote:
Akemi Yagi wrote:
On Fri, Oct 10, 2008 at 12:23 PM, Troy Dawson <[EMAIL PROTECTED]> wrote:
Hi,
Well, it looks like I'm going to have to put out a version of yum with
the
"versionfix" plugin for SL5. In summary, rpm is picking newer and older
package names different than I expected.
According to rpm
<name>-1.el5.1 is newer than <name>-1.el5_x.1
which is backwards from what I was thought. So all the checks that I
did
when looking for poorly named rpms in SL5, ... well ... I need to
recheck
them.
Anyway, expect a fix next week. I'm sorry about this.
Troy
An EVR comparison tool may be handy when checking which is
newer/older. I use fedora-rpmvercmp in rpmdevtools (EPEL repo).
Akemi
I will give it a shot. Because already I have found exceptions to what I
just said above.
Are there any epoch's involved here ?
Bug 227860 meant that different tools compared "Epoch 0" with
"no epoch" differently.
No.
Thanks to the rpmvercmp I was able to track it down. And I completely do *not*
agree with the rpm people on this. They seem to think that a _ and . are the
same thing when doing compares.
Look at this
# fedora-rpmvercmp 0 6.6 2.el5_2 0 6.6 2.el5.7
0:6.6-2.el5.7 is newer
# fedora-rpmvercmp 0 6.6 2.el5_8 0 6.6 2.el5.7
0:6.6-2.el5_8 is newer
# fedora-rpmvercmp 0 6.6 2.el5_7 0 6.6 2.el5.7
These are Equal
So, I will be going through all the rpm's, yet again, for both SL4 and SL5 and
figure out what is bad, and what is good. I'll try to keep everyone posted
when I delete and add things to the repositories.
Troy
p.s. Thanks again Akemi. It would have taken me forever to track this down
without fedora-rpmvercmp
--
__________________________________________________
Troy Dawson [EMAIL PROTECTED] (630)840-6468
Fermilab ComputingDivision/LCSI/CSI DSS Group
__________________________________________________
