Correct me if I am wrong, but shouldn't such build fail due to unsatisfied dependencies? And if the build fails, the package maintainer will probably notice that. Is the purpose of this proposed test to detect such builds that failed and its maintainer didn't notice it, or is there another purpose? I'm not a package maintainer and I don't have much experience building packages and using Koji, so further explanation is helpful. Thanks.
There are two type of dependencies mandatory and optional. If mandatory dependence missing then build will failed. But if optional dependence missing build will be complete but some of functionality will be missing.
For example see this okular build log, where BuildRequires qca2-devel was unintentionally missed: http://kojipkgs.fedoraproject.org/packages/okular/4.6.95/2.fc16/data/logs/i6...
There is cmake output: ----------------------------------------------------------------------------- -- The following OPTIONAL packages could NOT be located on your system. -- Consider installing them to enable more features from this software. ----------------------------------------------------------------------------- * QCA (2.0.0 or higher) http://delta.affinix.com/qca/ Qt Cryptographic Architecture (QCA) Support for encrypted OpenDocument Text documents in Okular.
But this build was complete.
This was wixed in next build by adding needed BuildRequires http://kojipkgs.fedoraproject.org/packages/okular/4.6.95/3.fc16/data/logs/i6...
So such missing optional dependencies needs attention.