Bug 441066

Summary: Conflict resolution does not work
Product: [openSUSE] openSUSE 11.1 Reporter: Richard Biener <rguenther>
Component: YaST2Assignee: Stefan Schubert <schubi>
Status: RESOLVED DUPLICATE QA Contact: Jiri Srain <jsrain>
Severity: Critical    
Priority: P2 - High CC: kkaempf, locilka, schubi
Version: Factory   
Target Milestone: ---   
Hardware: Other   
OS: Other   
Whiteboard:
Found By: Development Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: yast logs
solver testcase

Description Richard Biener 2008-11-03 14:04:21 UTC
If you update from Beta3 to Beta4 on x86_64 you will run into the problem that
the Beta4 tree has mismatched gcc43-java and gcc43 packages.  Manual
intervention on update is required and as Conflict Resolution a

"replacement of gcc43-java-4.3.3_20081002-1.x86_64 with gcc43-java-4.3.3_20081002-1.x86_64"

is presented (note the same package name is printed twice - an error?).
The replacement version is what is currently installed on the system
(from Beta3).

Selecting this alternative doesn't have any effect though, the same
dialog comes up again.  Possibly a satsolver interface problem?

The alternative suggestion 'deinstallation of gcc-java' makes progress but
ends up with the above problem unsolved and unsolvable.
Comment 1 Lukas Ocilka 2008-11-03 14:06:00 UTC
Duncan: duplicate?
Comment 2 Richard Biener 2008-11-03 14:07:02 UTC
Created attachment 249377 [details]
yast logs
Comment 3 Stefan Schubert 2008-11-03 14:31:54 UTC
yes

*** This bug has been marked as a duplicate of bug 439134 ***
Comment 4 Richard Biener 2008-11-03 14:52:17 UTC
That duplicate is about the bogus printing not about the non-existant effect
of enabling that conflict resolution.
Comment 5 Stefan Schubert 2008-11-03 16:00:28 UTC
I have still the opinion that it is the same problem. But perhaps I am wrong.
Please attach a solvertestcase.
Comment 6 Richard Biener 2008-11-03 16:28:48 UTC
Created attachment 249414 [details]
solver testcase
Comment 7 Richard Biener 2008-11-04 09:35:06 UTC
Foo.
Comment 8 Klaus Kämpf 2008-11-06 07:29:22 UTC
raising prio (see #441765)
Comment 9 Stefan Schubert 2008-11-06 09:25:13 UTC
The testcase does not represent the described problem.

>!> Solution :
>!> install aaa_base-11.1-9999.x86_64[openSUSE-11.1-Updates]
>!> install dbus-1-32bit-1.2.4-2.x86_64[openSUSE-11.1-Updates]
>!> install dbus-1-devel-1.2.4-2.x86_64[openSUSE-11.1-Updates]
>!> install dbus-1-x11-1.2.4-2.x86_64[openSUSE-11.1-Updates]
>!> install libgcj43-jar-4.3.3_20081022-1.2.x86_64[openSUSE-dvd 11.1-0]
>!> install openSUSE-release-11.1-1.34.x86_64[openSUSE-dvd 11.1-0]
>!> install procps-3.2.7-144.x86_64[openSUSE-11.1-Updates]
>!> install update-test-interactive-0-2.0.noarch[openSUSE-11.1-Updates]
>!> install update-test-optional-0-2.0.noarch[openSUSE-11.1-Updates]
>!> install update-test-reboot-needed-0-2.0.noarch[openSUSE-11.1-Updates]
>!> install update-test-security-0-2.0.noarch[openSUSE-11.1-Updates]
>!> install update-test-trival-0-2.0.noarch[openSUSE-11.1-Updates]
>!> install product:openSUSE-11.1.x86_64[openSUSE-dvd 11.1-0]
>!> update aaa_base-11.1-10000.1.x86_64
>!> update dbus-1-32bit-1.2.4-2.2.x86_64
>!> update dbus-1-devel-1.2.4-2.2.x86_64
>!> update dbus-1-x11-1.2.4-2.3.x86_64
>!> delete gcc43-java-4.3.3_20081002-1.1.x86_64
>!> delete libgcj43-devel-4.3.3_20081002-1.1.x86_64
>!> update libgcj43-jar-4.3.3_20081002-1.1.x86_64
>!> update procps-3.2.7-147.2.x86_64
>!> update update-test-interactive-99.99-99.99.noarch
>!> update update-test-optional-0-1.3.noarch
>!> update update-test-reboot-needed-99.99-99.99.noarch
>!> update update-test-security-99.99-99.99.noarch
>!> update update-test-trival-99.99-99.99.noarch
>!> delete product:openSUSE-11.1-0.noarch

comment #1 shows the problem of bug 439134

So, I cannot do very much more here.
Feel free to reopen the bug if you can provide a testcase which shows your problem.

*** This bug has been marked as a duplicate of bug 439134 ***
Comment 10 Richard Biener 2008-11-06 09:33:04 UTC
Well, I have no clue how to "verify" the testcase itself.  But the solver
testcase was generated with zypper --debug-solver dup which showed the problem.
Well - a problem similar to the described one - it's hard to "reconstruct"
the original situation after the update.

If it happens to be fixed with the current libzypp version fine, but the testcase
definitely should fail with the Beta4 libzypp.  How to test that I have no
clue.  Where's documentation for all this?
Comment 11 Stefan Schubert 2008-11-06 09:54:36 UTC
Well, you have had already done the correct things but I fear at the wrong time
( the update has been already done ). So it is really "hard to "reconstruct"
the original situation after the update."
But I think the problem is fixed. :-)