| Summary: | Conflict in libgamin packages | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | Doug Laidlaw <laidlaws> |
| Component: | RPM Packages | Assignee: | Anssi Hannula <anssi.hannula> |
| Status: | RESOLVED WORKSFORME | QA Contact: | |
| Severity: | normal | ||
| Priority: | Normal | CC: | anssi.hannula, fundawang, stormi-mageia |
| Version: | 2 | Keywords: | Triaged |
| Target Milestone: | --- | ||
| Hardware: | i586 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Source RPM: | gamin | CVE: | |
| Status comment: | |||
|
Description
Doug Laidlaw
2012-12-25 02:23:09 CET
This is still there after a fresh install of Beta1. have you any urpmi logs ? Source RPM:
(none) =>
libgamin
Manuel Hiebel
2012-12-25 14:27:08 CET
Source RPM:
libgamin =>
gamin No rest even for Christmas? I should have a log for the system upgrade. Will boot into it tomorrow (00.40 local here.) >No rest even for Christmas?
well thanks to christmas I have time to do some triaging again :)
Nothing about urpmi in the journal and no /var/log/urpmi. I haven't become familiar with systemd. The error message given is: "1 installation transactions failed There was a problem during the installation: libgamin-1_0 < 1.0-9 is obsoleted by (installed) libgamin1_0-0.1.10-9.mga3.i586" But what I can't understand is that until I uncheck libgamin, nothing gets installed. Once I do, all the other updates can proceed. (In reply to comment #6) > But what I can't understand is that until I uncheck libgamin, nothing gets > installed. Once I do, all the other updates can proceed. The message about the CD-ROM not being mounted, appears in the box along with the message in Comment 5. Once I mount the CD-ROM manually, I can see the progress bar, and everything LOOKS as though the other packages are being installed, but they are not. They install on the second attempt, after I have excluded libgamin. Perhaps I am wrong in concluding that libgamin is holding up the others? As I was typing the above, further updates came down and libgamin was still listed. I unchecked libgamin first, and the others installed normally. But the CD-ROM was already mounted, and may not have been called for, even. and rpm -qa libgamin give what ? Nothing at all, i.e. no version installed. But rpmdrake says that libgamin1_0 is installed, and libgamin-1_0 needs to be installed (blue arrow.) libgamin-1_0 is said to be from Mga2. It sounds as though it shouldn't be asked for at all. Note that the message was that libgamin is obsoleted, not quite the same as a conflict. I think that a remake of the rpm database is in order? don't know, funda, anssi, ideas ? CC:
(none) =>
anssi.hannula, fundawang There should not be any MGA2 repositories set up if one is installing MGA3 packages. This will cause issues as urpmi will try to install the MGA2 libgamin-1_0 package instead o MGA3 libfam0 package. However, I do not understand from the report how any MGA2 repositories would be set up. Maybe I am missing something, since comment #1 says there was a fresh install of MGA3, while comment #3 talks about upgrade. If it was indeed an upgrade, it would be an installer bug if not all MGA2 repositories were disabled during upgrade. I first experienced this running Alpha 3. I got the error we are talking about. Then I did an "upgrade" type installation from the Beta1 DVD. When that got to installing packages, it gave me the same error message and looped back to the beginning. Then I did a clean reinstall. The error is still there. I have just rebuilt my local RPM database. Nothing at all has changed since comment #10. The installed version seems to be the correct one. The only version on the Cauldron mirror is libgamin1_0-0.1.10-9.mga3.i586.rpm, the one I have installed. I can't see any Mga2 repos in /var/lib/urpmi, nor can I see any in /var/cache/urpmi/mirrors.cache. *** I got the command wrong. I did rpm -qa | libgamin That had no output. I have just done rpm -qa | grep libgamin That shows libgamin1_0-0.1.10-9.mga3 Sorry about that. So the only issue left is, where is the other version coming from? The RPM isn't sitting on my HD in /var/cache/rpm anywhere. libgamin-1_0-0.1.10-8.mga2.i586.rpm is on the installation CD. That is where it is coming from. Why is it still being seen as an upgrade? Search me! Unless (a long shot and more of an example) libgamin-1 is seen as newer than libgamin1? I have noticed other conflict situations on Bugzilla that boiled down to something similar. OK, it seems the libgamin change is very recent (one week). This will work properly with the next installation CDs when the old libgamin is no longer in the CDs.
Here's a graph showing what's happening:
------------------------------------------------------------
Some very old distribution version
Package libfam0
- has libfam.so.0
|
| Upgrade (since libgamin-1_0
| path says it upgrades
| any libfam0)
---------------------------+--------------------------------
Mageia 2 |
V
Package libgamin-1_0 <----+ Upgrade path to
- has libgamin-1.so.0 | the wrong
- has libfam.so.0 | direction
/ | since libgamin1_0
/ Upgrade | says it upgrades
/ path | any libfam0
/ |
---------------+-------------------------+------------------
Mageia 3 | |
V |
Package libgamin1_0 Package libfam0
- has libgamin-1.so.0 - has libfam.so.0
(no upgrade path from
Mageia 2, but will be
automatically installed
if something needs
libfam.so.0)
------------------------------------------------------------
If you have libgamin1_0 installed, and you have libgamin-1_0 in the repositories/CD, and something needs libfam.so.0, urpmi will see both libgamin-1_0 and libfam0 would provide it, but it will think libgamin-1_0 is newer since it says so, and will therefore try to install libgamin-1_0, which is not possible since libgamin1_0 (a newer version of libgamin-1_0) is already installed.
One way to avoid this issue completely would be to rename libfam0 to e.g. libgamin-fam0, at least temporarily (though maybe it even actually makes a better name for it...), so that the Mageia 2 would no longer be considered newer than the Mageia 3 libfam. Funda, WDYT?
Glad that it all makes sense now. I suspect that the RPM list for Beta1 was frozen a few days too early to avoid this mix-up. Murphy's Law. My thought from the beginning was to add the troublesome one to the blacklist, but that may not work? Adding libgamin-1_0 into skip.list may indeed work, but I'm not 100% sure. I'm in favour of fixing it mga2. Just push versions into obsoletes of libfam0. Version:
Cauldron =>
2 You know a lot better than I do. At the moment I am still unchecking it in the Updates. I would like not to have to do this, but it will disappear with the next set of media. (In reply to comment #18) > I'm in favour of fixing it mga2. Just push versions into obsoletes of libfam0. I don't see how that would fix anything. The current libgamin package would still stay in mga2 core/release and it would still obsolete all libfam0. (In reply to comment #20) > I don't see how that would fix anything. The current libgamin package would > still stay in mga2 core/release and it would still obsolete all libfam0. Drop those obsoletes in core/updates. Then we'll be all happy :) Anssi, Funda, what was or should be the follow up to this bug? Keywords:
(none) =>
Triaged This message is a reminder that Mageia 2 is nearing its end of life. Approximately one month from now Mageia will stop maintaining and issuing updates for Mageia 2. At that time this bug will be closed as WONTFIX (EOL) if it remains open with a Mageia 'version' of '2'. Package Maintainer: If you wish for this bug to remain open because you plan to fix it in a currently maintained version, simply change the 'version' to a later Mageia version prior to Mageia 2's end of life. Bug Reporter: Thank you for reporting this issue and we are sorry that we may not be able to fix it before Mageia 2 is end of life. If you would still like to see this bug fixed and are able to reproduce it against a later version of Mageia, you are encouraged to click on "Version" and change it against that version of Mageia. Although we aim to fix as many bugs as possible during every release's lifetime, sometimes those efforts are overtaken by events. Often a more recent Mageia release includes newer upstream software that fixes bugs or makes them obsolete. -- The Mageia Bugsquad This bug is history as far as I anm concerned. I thought that it was solved, then the discussion between Anssi and Funda started. This is my bug. Should I mark it as fixed? It was a temporary conflict between packages. looks we can Status:
NEW =>
RESOLVED Good. The packaging discussion is over my head. Good to see your profile, Manuel. I wish I could do more, but (a) I don't have the training; and (b) my health makes me very unreliable. |