Red Hat Bugzilla – Bug 111092
Dependency failure when provider has epoch and requirer has not
Last modified: 2007-11-30 17:10:34 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4.1)
Description of problem:
When a package provides an explicit epoch, and another package
requires the provider's package version with epoch unspecified, the
dependency check fails.
According to /usr/share/doc/rpm-4.2.1/dependencies, an unspecified
epoch in a Requires line should match any provided epoch.
I will submit spec files for three test packages which demonstrate
the dependency treatment of a package which provides an epoch:
which provides provides-epoch-37:1.4.1-17
where 37 is the epoch
which requires provides-epoch = 37:1.4.1
which requires provides-epoch = 1.4.1
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Copy provides-epoch.spec, requires-epoch.spec and
requires-noepoch.spec into .../rpm/SPECS/.
2. Create packages:
rpmbuild -ba provides-epoch.spec
rpmbuild -ba requires-epoch.spec
rpmbuild -ba requires-noepoch.spec
3. Install packages (as root):
rpm -iv provides-epoch-*.i386.rpm
rpm -iv requires-epoch-*.i386.rpm
rpm -iv requires-noepoch-*.i386.rpm
Actual Results: The first and second packages are installed.
Installation of the third package fails with:
error: Failed dependencies:
provides-epoch = 1.4.1 is needed by requires-noepoch-1.4.2_01-2.cern
Expected Results: All three packages should have been installed.
Additional info: None.
Created attachment 96223 [details]
Created attachment 96224 [details]
Created attachment 96225 [details]
Not specifying an epoch is broken behaviour. RHL 9 complained about
it, and I believe Fedora treats no epoch as epoch=0, so I believe this
is not a bug.
Here is the relevant paragraph in the RPM documentation file
"The epoch (if present) is a monotonically inceasing integer, neither
the version or the release can contain the '-' hyphen character, and
the dependency parser does not permit white space within a definition.
Unspecified epoch and releases are assumed to be zero, and are
interpreted as "providing all" or "requiring any" value."
It is the second sentence that makes me believe than an unspecified
epoch in a Requires line should match any provided epoch. I feel that
the behaviour described in this sentence is logical. Do you want to
change the Requires line in every dependent package, at the moment
when you are forced to introduce, or to change, an epoch in a