A non-zero exit code is returned for a missing package when --skip-broken used. The available package is not installed. Reproducible: Always Steps to Reproduce: dnf install -y --skip-broken tree forest echo $? Actual Results: Failed to resolve the transaction: No match for argument: forest # echo $? 1 Expected Results: tree package installed, missing forest ignored and 0 returned dnf5-5.0.14-1.fc39.x86_64
The --skip-broken is in DNF5 split between --skip-broken (to skip packages uninstallable due to depsolv problems) and --skip-unavailable (to skip packages unavailable in repositories). The change is documented here: https://github.com/rpm-software-management/dnf5/blob/main/doc/changes.rst#strict-configuration-option-deprecation (this describes the "strict" configuration option, but --skip-broken was an alias for it).
So in your case try `dnf5 install -y --skip-unavailable tree forest`. I discovered a gap in the dnf5 install command man pages - these options are not documented - upstream issue https://github.com/rpm-software-management/dnf5/issues/644.
Thanks for the info. Just wonder, doesn't the "old" --skip-broken behaviour cover quite a frequent use case? This seems like a backward-incompatible change which will affect many users, I'd say.
Why dnf is changing behavior of --skip-broken, rather then preserving it? This change broke our test infrastructure on many place and we see huge risks it will affect Fedora/CentOS Stream/RHEL users and customers. changes.rst mention that strict config option was a problem that it did two things - but we see it as benefit do these together. Is there any issue introducing --skip-uninstallable and --skip-unavailable while keeping --skip-broken ?
Also with introducing this incompatilibity, customers which have mutiple versions of RHEL would need to start maintaining RHEL release specific scripts/config
# dnf5 --skip-broken upgrade Unknown argument "--skip-broken" for command "dnf5". Add "--help" for more information about the arguments. # rpm -qa | grep dnf5 libdnf5-5.0.15-2.fc39.x86_64 libdnf5-cli-5.0.15-2.fc39.x86_64 dnf5-5.0.15-2.fc39.x86_64 dnf5-plugins-5.0.15-2.fc39.x86_64
The issue with the position of the argument is being discussed in related bugzilla there: https://bugzilla.redhat.com/show_bug.cgi?id=2216055.
Can you please provide ETA for resolution of this regression? It causes stability issues in our CI/Testing pipelines.
This bug appears to have been reported against 'rawhide' during the Fedora Linux 39 development cycle. Changing version to 39.