Bugzilla – Bug 521549
OpenOffice.org consistently causes problems (missing dependencies) in FACTORY
Last modified: 2011-10-28 12:47:40 UTC
This has happened a number of time the last weeks, today's incarnation is: Computing upgrade... 5 Problems: Problem: OpenOffice_org-3.1.0.98.1-1.1.i586 requires OpenOffice_org-libs-extern = 3.1.0.98.1, but this requirement cannot be provided Problem: OpenOffice_org-libs-core-3.1.0.98.1-1.3.i586 requires OpenOffice_org-libs-gui = 3.1.0.98.1, but this requirement cannot be provided Problem: OpenOffice_org-mailmerge-3.1.0.98.1-1.3.i586 requires OpenOffice_org-pyuno = 3.1.0.98.1, but this requirement cannot be provided Problem: OpenOffice_org-l10n-extras-3.1.0.98.1-1.3.noarch requires OpenOffice_org-ure = 3.1.0.98.1, but this requirement cannot be provided Problem: kde4-marble-4.2.90-2.4.i586 requires kde4-marble-data = 4.2.90, but this requirement cannot be provided Problem: OpenOffice_org-3.1.0.98.1-1.1.i586 requires OpenOffice_org-libs-extern = 3.1.0.98.1, but this requirement cannot be provided deleted providers: OpenOffice_org-libs-extern-3.1.0.98.1-1.5.i586 Solution 1: Following actions will be done: deinstallation of OpenOffice_org-3.1.0.98.1-1.1.i586 deinstallation of OpenOffice_org-Quickstarter-1.0-711.34.i586 deinstallation of OpenOffice_org-base-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-calc-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-draw-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-gnome-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-impress-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-kde-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-math-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-writer-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-filters-optional-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-filters-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-components-3.1.0.98.1-1.3.i586 deinstallation of OpenOffice_org-base-extensions-3.1.0.98.1-1.1.i586 deinstallation of OpenOffice_org-calc-extensions-3.1.0.98.1-1.1.i586 deinstallation of OpenOffice_org-draw-extensions-3.1.0.98.1-1.1.i586 deinstallation of OpenOffice_org-impress-extensions-3.1.0.98.1-1.1.i586 deinstallation of OpenOffice_org-writer-extensions-3.1.0.98.1-1.1.i586 Solution 2: keep OpenOffice_org-libs-extern-3.1.0.98.1-1.5.i586 Solution 3: ignore some dependencies of OpenOffice_org
it happens on every version update of OOo when there is one build failure. And the right solution is to keep the old version intact - solution 2.
This is for Petr, right?
I think that the problem is combination of more things: 1. OOo build is split into more pieces and it takes long time. 2. I do my best to submit buildable packages but it takes many days before the submit request is accepted. Other submitted packages causes new build breakages. 3. Build Service is much slower than autobuild and it takes many days until the full build finishes. OOo is a leaf package and it is sometimes not built many days at all. 4. The build hosts are unstable and cause many false build errors. At least, I see this in the external Build Service. 5. It is not easy to check for build failures. I use osc -A https://api.suse.de prjresults "SUSE:Factory:Head" -n "OpenOffice_org" It shows "b = blocked" most of the time; It is hard to catch "F = failure" We are working on bigger changes in the split build that should reduce the build dependencies and might help to build OOo earlier. Unfortunately, it is quite complex task and we will use it for OOo-3.2 (after openSUSE-11.2 release). I think that we also need changes in the other areas: + faster check in + faster Build Service + more stable build hosts + more effective access to the build results
As the reporter, I think I have seen the situation improve significantly since filing the report (without knowing what, if anything) has been changed. From this perspective I'd be fine seeing this as RESOLVED FIXED.
Yup, I think that the Build Service improved a lot and OOo builds more reliably now. Closing as RESOLVED FIXED as suggested by Gerald in the comment #6.
.
*** Bug 480788 has been marked as a duplicate of this bug. ***