From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.8.1.5) Gecko/20070718 Fedora/2.0.0.5-1.fc7 Firefox/2.0.0.5 Description of problem: # yum install ctapi-cyberjack Loading "allowdowngrade" plugin Loading "skip-broken" plugin Loading "fastestmirror" plugin Setting up Install Process Parsing package install arguments Loading mirror speeds from cached hostfile Resolving Dependencies --> Running transaction check ---> Package ctapi-cyberjack.i386 0:3.0.3-1.fc7 set to be updated Importing additional filelist information filelists.sqlite.bz2 100% |=========================| 68 kB 00:00 filelists.sqlite.bz2 100% |=========================| 7.6 MB 00:00 filelists.sqlite.bz2 100% |=========================| 3.0 MB 00:00 ---> Package ctapi-cyberjack.x86_64 0:3.0.3-1.fc7 set to be updated --> Processing Dependency: /usr/lib64/ctapi for package: ctapi-cyberjack --> Processing Dependency: libstdc++.so.6(GLIBCXX_3.4) for package: ctapi-cyberjack --> Processing Dependency: group(ctapiusers) for package: ctapi-cyberjack --> Processing Dependency: libstdc++.so.6 for package: ctapi-cyberjack --> Processing Dependency: libgcc_s.so.1 for package: ctapi-cyberjack --> Processing Dependency: /usr/lib/ctapi for package: ctapi-cyberjack --> Processing Dependency: libgcc_s.so.1(GCC_3.0) for package: ctapi-cyberjack --> Processing Dependency: libstdc++.so.6(CXXABI_1.3) for package: ctapi-cyberjack --> Processing Dependency: libusb-0.1.so.4 for package: ctapi-cyberjack --> Restarting Dependency Resolution with new changes. --> Running transaction check ---> Package ctapi-common.x86_64 0:1.1-2.fc7 set to be updated ---> Package libusb.i386 0:0.1.12-7.fc7 set to be updated ---> Package libstdc++.i386 0:4.1.2-12 set to be updated ---> Package libgcc.i386 0:4.1.2-12 set to be updated ---> Package ctapi-cyberjack.x86_64 0:3.0.3-1.fc7 set to be updated ---> Package ctapi-common.i386 0:1.1-2.fc7 set to be updated ---> Package ctapi-cyberjack.i386 0:3.0.3-1.fc7 set to be updated --> Processing Dependency: group(ctapiusers) for package: ctapi-cyberjack --> Finished Dependency Resolution Error: Missing Dependency: group(ctapiusers) is needed by package ctapi-cyberjack Version-Release number of selected component (if applicable): ctapi-cyberjack.i386 0:3.0.3-1.fc7 How reproducible: Always Steps to Reproduce: 1. yum install ctapi-cyberjack Actual Results: --> Processing Dependency: group(ctapiusers) for package: ctapi-cyberjack --> Finished Dependency Resolution Error: Missing Dependency: group(ctapiusers) is needed by package ctapi-cyberjack Expected Results: Something would provide the group Additional info:
Does this installation have any fedora-usermgmt package installed?
The bug is in the ctai-common package.
So are you going to fix it or do I have to file a bug about it? ctapi-common installs just fine, it is ctapi-cyberjack that fails to install.
This is not a ctapi-common bug - as I told Frank in PM last weekend, ctapi-cyberjack should just remove the be group(ctapiusers) stuff and do "Requires: ctapi-common >= 1.1-2".
ctapi-common no longer provides group(ctapiusers) afaict nothing NEEDS to require group(ctapiusers) kill the dep in ctapi-cyberjack, submit it for rebuild move along. This is like a 10 second bug fix.
No ctapi-common must be fixed. The provides group entry is missing, so I can't do anything.
Frank, Could you explain to me why you think ctapi-common should be changed instead of ctapi-cyberjack? Thanks -sv
Yes, because to add an provides group .. is much cleaner than simply modify the depend version. For example when in an later update of the common part the group was removed by an mistake, then it will only comes out very late.(At the first time when somebody try to access the reader). The requere group will prevents of missing the group on the system.
So how do we resolve this? You point the finger at the -common package it then gets pointed back... The tree is broken and it needs to be resolved, please figure it out.
Frank you say this bug is in ctapi-common, please quit closing it as NOTABUG. Fix one of your packages I don't care if it is ctapi-common or ctapi-cyberjack. You are a maintainer of both of these packages, fix one of them. If you and your co-maintainer of ctapi-common can not come to an agreement I am of the opinion that these packages be removed from the tree. Users will end up paying the price.
Reply to Comment #10: Yes, I'd agree to that, if a bug is keeping a package from being even installed and the maintainers cannot or will not fix it, sounds like ground for removal.
I have fix it at the ctapi-common package, but my fix was removed. So I can't do anything more.
Ville, You and Frank need to solve this b/t the two of you. -sv
Seth, others, I do not intend to do anything about this - from my POV it's not a ctapi-common bug. Here's what has happened: I asked Frank to take over ctapi-common completely, he didn't want more than co-maintainership and I thought it was better than nothing and would result in less things for me to worry about (but later I've learned it's been the opposite). Then, Frank pushed a broken ctapi-common update without discussing the largish and questionable changes with me, I posted how I intended to fix it in Bugzilla. Meanwhile the broken update slipped through to F7 updates-testing and devel even though I had asked rel-eng to hold it, then I pushed the fix ASAP after learning what had happened. The fixed update stayed in updates-testing for almost a week, and I pinged Frank during that time that ctapi-cyberjack will need adjustments or it will break, he probably received broken dep reports also from buildsys before the breakage hit the non-testing updates tree, and he told me I should just go ahead and fix it. I said if he wants the Provides, I thought it was his job to take care of either in ctapi-common or ctapi-cyberjack and he's free to take care of it but not reintroduce the earlier breakage. I never heard back, and looks like he did nothing except pushed ctapi-cyberjack to non-testing updates in a state that had broken deps both with and without the ctapi-common update. Frank can still go ahead and either adjust ctapi-cyberjack's dependencies or add Provides: group(ctapiusers) to ctapi-common. As said, I do not intend to do anything about this, I don't think it's a ctapi-common bug, it's not a standard practice in Fedora at the moment to use these Provides/Requires: group(foo) things, and I don't use or maintain ctapi-cyberjack. So, ctapi-common is still up for grabs for a new main or co-maintainer, for both Fedora and EPEL as far as I'm concerned. The current co-maintainership arrangement has not worked out well at all, and I'm now even more ready to let others worry about this stuff than I was before, and will ask for removal of myself from its maintainers tomorrow.
Ville, thanks for the explanation and fair enough. Frank, it sounds like the ball is in your court for both of these packages. If neither of you wants to fix it then please send an orphan message to maintainers. this little sideshow has gone on for long enough, I think.
(In reply to comment #14) Thank you Ville
No the ball is not me it is at Ville!!!! But I can add it to the common package again.
The Fedora 7 tree is still broken because of these packages. This bug DOES exist.
Latest builds in koji appear to be OK, but they have not been pushed afaik. Who gets this show on the road?
The last package contain the fix.