October 2011 - openSUSE Packaging - openSUSE Mailing Lists

[opensuse-packaging] Obsoletes < vs Obsoletes <=
by Jan Engelhardt 10 Oct '11

10 Oct '11

Hi, in discussions etc. people say one should use < in the obsoletes/provides relation: Obsoletes: foo < %version Provides: foo = %version However this does not reliably "work" when %version did not change across the change (e.g. changing EV from foo-1.0 towards libfoo-1.0), the old foo-1.0 will stick around because it is not obsoleted by libfoo-1.0. So "foo < %version" is actually bogus and should be either <=, or actually foo < %version-%release. <= makes technology unhappy ("package obsoletes itself"), so that leaves us with the %version-%release approach. Does baselibs.conf support obsoletes "foo < <version>-<release>" actually? -- To unsubscribe, e-mail: opensuse-packaging+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-packaging+help(a)opensuse.org

4 6

[opensuse-packaging] A couple fedora questions
by Greg Freemyer 07 Oct '11

07 Oct '11

All, I've got a normal package and a KMP I want to also build for Fedora. (The only 2 packages in home:next4:ext4-snapshots). The normal package is failing because I'm using the "Recommends" tag in my opensuse specfile. (I have it wrapped in %if suse currently). Is there a Fedora equivalent? The KMP is failing because I have %kernel_module_package in that specfile. Again, is there a Fedora equivalent? Thanks Greg -- To unsubscribe, e-mail: opensuse-packaging+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-packaging+owner(a)opensuse.org

2 1

[opensuse-packaging] RPM scriptlets -e argument confusion
by Johannes Meixner 07 Oct '11

07 Oct '11

Hello, http://en.opensuse.org/openSUSE:Packaging_scriptlet_snippets reads ----------------------------------------------------------------------- ... rpm in its default configuration does not at the moment execute shell scriptlets with the -e argument ... ----------------------------------------------------------------------- Therefore this test scriptlet should work: ----------------------------------------------------------------------- %pre # Test if RPM executes shell scriptlets without the -e argument # (i.e. do not exit if a command exits with a non-zero status). # Without the -e argument it would not exit at the next command... cat doesnotexist 1>/dev/null 2>/tmp/testRPMpre.stderr # ...and therefore it would run this command: echo "RPM pre scriptlet runs without -e" 1>/tmp/testRPMpre.stdout ----------------------------------------------------------------------- But a package with such a scriptlet fails during build with: ----------------------------------------------------------------------- running 06-check-installtest ... testing for pre/postinstall scripts that are not idempotent preinstall script of <package>.rpm failed ----------------------------------------------------------------------- The reason is that the post-build-checks RPM in the build system contains a usr/lib/build/checks/06-check-installtest script which runs each RPM scriptlet with explicit -e argument setting via ----------------------------------------------------------------------- SCRIPT_ARG=-e ... chroot $BUILD_ROOT sh $SCRIPT_ARG /.build_patchrpmcheck_scr ... ----------------------------------------------------------------------- where /.build_patchrpmcheck_scr is a copy of the RPM scriptlet. With "#!BuildIgnore: post-build-checks" I enforced to skip all post build checks to get the package built. And now - surprise! - surprise! - I can install the built package without an error using plain "rpm" which shows that rpm does not use the -e argument - at least not the rpm on my workstation. I wonder why then the post build checks use the -e argument? I would expect that the post build checks emulate what rpm does (i.e. what happens on the user's system who installs it) and not exaggerate. >From my point of view using the -e argument for the post build checks does not provide better RPM scriplets. It only enforces packagers to add dumb "|| true" to all statements: ----------------------------------------------------------------------- %pre do_crap || true do_more_crap || true mess_up_everything || true ----------------------------------------------------------------------- Kind Regards Johannes Meixner -- SUSE LINUX Products GmbH -- Maxfeldstrasse 5 -- 90409 Nuernberg -- Germany HRB 16746 (AG Nuernberg) GF: Jeff Hawn, Jennifer Guild, Felix Imendoerffer -- To unsubscribe, e-mail: opensuse-packaging+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-packaging+owner(a)opensuse.org

2 3

[opensuse-packaging] [reminder] openSUSE 12.1 RC 1 is next week
by openSUSE Roadmap Reminder 05 Oct '11

05 Oct '11

Hello all, openSUSE 12.1 RC 1 will be released next week, on Thursday 13 October 2011. Please submit your packages before Friday afternoon (UTC time) to make sure they will get included in this release. Leaf packages submitted during the week-end might also get accepted on Monday. For a quick overview of the openSUSE Roadmap, please see: http://en.opensuse.org/Roadmap Thanks, -- openSUSE Roadmap Reminder -- To unsubscribe, e-mail: opensuse-packaging+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-packaging+help(a)opensuse.org

2 1

[opensuse-packaging] another KMP questions
by Greg Freemyer 03 Oct '11

03 Oct '11

All, I just test installed the first KMP I've been part of building: The project is: home:next4:ext4-snapshots, and the package is: ext4dev-snapshots in that project. It installs via zypper in ext4dev-kmp-desktop. When I look in /lib/modules, it gets installed to: /lib/modules/3.1.0-rc7-3-desktop/ But the kernel is 3.1.0-rc7-5-desktop and the other modules are in /lib/modules/3.1.0-rc7-5-desktop/. (note -3 vs. -5 above). Is that okay, or do I need to do something to get the kernel and KMP version number to be exactly the same? Thanks Greg -- Greg Freemyer Head of EDD Tape Extraction and Processing team Litigation Triage Solutions Specialist http://www.linkedin.com/in/gregfreemyer CNN/TruTV Aired Forensic Imaging Demo - http://insession.blogs.cnn.com/2010/03/23/how-computer-evidence-gets-retrie… The Norcross Group The Intersection of Evidence & Technology http://www.norcrossgroup.com -- To unsubscribe, e-mail: opensuse-packaging+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-packaging+help(a)opensuse.org

1 1

[opensuse-packaging] Need help with a kernel module - depends on libc.so.6 ??
by Greg Freemyer 03 Oct '11

03 Oct '11

All, I'm trying to get a KMP to work against current factory / kernel:HEAD (3.1-rc7) The project is: home:next4:ext4-snapshots And the package is: ext4dev-snapshots in that project. It builds fine, but when I go to install it I get: ======================== #### YaST2 conflicts list - generated 2011-10-03 13:39:52 #### nothing provides libc.so.6(GLIBC_2.14)(64bit) needed by ext4dev_snapshots-1.0.13-15.1.x86_64 [ ] do not install ext4dev_snapshots-1.0.13-15.1.x86_64 [ ] break ext4dev_snapshots by ignoring some of its dependencies ======================== I'm running tumbleweed, but I thought a KMP would be distro neutral. Thanks Greg -- To unsubscribe, e-mail: opensuse-packaging+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-packaging+help(a)opensuse.org

2 2

[opensuse-packaging] build-compare problem?
by Stefan Seyfried 01 Oct '11

01 Oct '11

Hi all, I'm trying to get the list of unnecessary republishings in my repos down, even in the "fun" repos. One problem I was not able to solve is: seife@susi:~/buildservice/arduino/avr-libc> osc bl openSUSE_Factory x86_64|tail -25 +++ /tmp/tmp.ilF7FI2UUs 2011-09-29 12:41:38.176000001 +0000 @@ -110,9 +110,9 @@ 000006d0 00 00 00 00 00 00 00 00 44 00 dd 00 00 00 00 00 |........D.......| 000006e0 00 00 00 00 44 00 df 00 00 00 00 00 00 00 00 00 |....D...........| 000006f0 44 00 24 01 00 00 00 00 00 00 00 00 44 00 25 01 |D.$.........D.%.| -00000700 00 00 00 00 00 2f 74 6d 70 2f 63 63 56 6a 77 49 |...../tmp/ccVjwI| -00000710 79 38 2e 73 00 2f 74 6d 70 2f 63 63 56 6a 77 49 |y8.s./tmp/ccVjwI| -00000720 79 38 2e 73 00 76 6f 69 64 3a 74 31 3d 31 00 5f |y8.s.void:t1=1._| +00000700 00 00 00 00 00 2f 74 6d 70 2f 63 63 6a 46 54 74 |...../tmp/ccjFTt| +00000710 5a 43 2e 73 00 2f 74 6d 70 2f 63 63 6a 46 54 74 |ZC.s./tmp/ccjFTt| +00000720 5a 43 2e 73 00 76 6f 69 64 3a 74 31 3d 31 00 5f |ZC.s.void:t1=1._| 00000730 5f 76 65 63 74 6f 72 73 3a 46 31 00 2e 2e 2f 2e |_vectors:F1.../.| 00000740 2e 2f 2e 2e 2f 2e 2e 2f 63 72 74 31 2f 67 63 72 |./../../crt1/gcr| 00000750 74 31 2e 53 00 5f 5f 62 61 64 5f 69 6e 74 65 72 |t1.S.__bad_inter| ... saving built packages /usr/src/packages/RPMS/noarch/avr-libc-1.7.1-3.9.noarch.rpm /usr/src/packages/SRPMS/avr-libc-1.7.1-3.9.src.rpm build34 finished "build avr-libc.spec" at Thu Sep 29 12:41:38 UTC 2011. [ 269.320282] SysRq : Power Off [ 271.127880] Power down. build: extracting built packages... avr-libc-1.7.1-3.9.noarch.rpm avr-libc-1.7.1-3.9.src.rpm This is in home:seife:arduino, my private "develproject" for CT:avr, the package avr-libc is linked from CrossToolchain:avr. It looks like the problem is some temporary filename from compilation. Anything I can do about this? Or can this be whitelisted in build-compare? Best regards, seife -- Stefan Seyfried "Dispatch war rocket Ajax to bring back his body!" -- To unsubscribe, e-mail: opensuse-packaging+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-packaging+help(a)opensuse.org

2 2

[opensuse-packaging] Re: [opensuse-factory] regression in (GNOME) factory regarding gtk2 development
by Boris Manojlovic 01 Oct '11

01 Oct '11

Hi Dominique and lists, On Sat, Oct 1, 2011 at 11:31 AM, Dimstar / Dominique Leuenberger <dimstar(a)opensuse.org> wrote: > On Sat, 2011-10-01 at 10:22 +0200, Boris Manojlovic wrote: >> Hi, >> >> gtk2-devel package does not pull "shared-mime-info" (more specificaly >> it should be pulled by gdk-pixbuf-devel) >> required by gdk-pixbuf-devel or specifically gdk-pixbuf-csource >> command from same package GDK uses mime-info for controlling >> which "loader" will be used for loading images and if package is >> missing it will fail to call correct shared library, >> as can not determine given image type. Installing shared-mime-info is >> enough to make loaders work as expected but should be fixed on higger >> level. >> >> Boris >> >> ------------------- error snip --------------- >> gdk-pixbuf-csource --extern --struct --raw --name=startbanner_pixdata >> rsrc/build.bmp > rsrc/editor_banner.c >> failed to load "rsrc/build.bmp": Couldn't recognize the image file >> format for file 'rsrc/build.bmp' >> ------------------- error snip --------------- > > Hi Boris, > > Thank you very much for your report and the analysis (and also the > submit request to implement your suggested solution). > > Unfortunately, I'm afraid, there must have been something else wrong, as > shared-mime-info is pulled in by a bunch of other dependencies. > > Explicitly, for gdk-pixbuf-devel, the chain is like this: > > shared-mime-info => libgio-2_0-0 => libgdk_pixbuf-2_0-0 => > gdk-pixbuf-devel. > > The dependency on libgio-2_0-0 to shared-mime-info was added back in > September 2010, based on bnc#555606. > > Can you please verify on your system if this dependency chain is the > same for your system and possibly also with "zypper verify" if your > system is in a consistent state? > yes it is consistent. You are correct 100% on my live factory it is really pulled by libgio and a lot of other packages. Unfortunately it is on OBS not on live system that i have a problem because of missing shared-mime-info package, so it was a bit wrong assessment from my side to begin with. But my maybe not perfect fix would fix issues on OBS. Maybe question is what changed from few days ago and now we do not pull shared-mime-info into gtk2-devel chain any more? Anyway I'm forwarding this to opensuse-packaging as this is more appropriate list for this exact issue. So recap is Factory installation is ok for development, but OBS fails for specified reason. > For now I put your SR on review with a link to this mail thread, until > it's all sorted out. > > Best regards, > Dominique / DimStar > > -- To unsubscribe, e-mail: opensuse-packaging+unsubscribe(a)opensuse.org For additional commands, e-mail: opensuse-packaging+help(a)opensuse.org

1 0