rpms/igraph/devel igraph.spec,1.10,1.11
Mamoru Tasaka
mtasaka at ioa.s.u-tokyo.ac.jp
Sun Feb 17 16:22:27 UTC 2008
Neal Becker wrote, at 02/18/2008 01:10 AM +9:00:
> Jesse Keating wrote:
>
>> On Sun, 17 Feb 2008 10:53:02 -0500
>> Neal Becker <ndbecker2 at gmail.com> wrote:
>>
>>> I'm confused.
>>> Without adding Provides: %{name}-%{version}, I have:
>>>
>>> rpm -q --provides -p
>>> ~/RPM/RPMS/x86_64/igraph-devel-0.5-8.fc8.x86_64.rpm igraph-devel =
>>> 0.5-8.fc8
>>>
>>> igraph-devel and python-igraph have the same version, but different
>>> release, so don't I need this provides in igraph-devel and requires in
>>> python-igraph, so the requirement is:
>>> igraph-devel = 0.5
>>> without release number?
>> I'm having a real hard time following this.
>>
>> Can you start from the beginning, what you think the problem is?
>>
>
> We have a new upstream for igraph, version 0.5.
>
> I'm trying to build igraph-0.5 and python-igraph-0.5.
>
> igraph-0.5-8.fc8.src.rpm builds
> igraph-0.5-8.fc8.x86_64.rpm
> igraph-devel-0.5-8.fc8.x86_64.rpm
>
> We have python-igraph-0.5-4.fc8.src.rpm.
>
> If I don't put any provides in igraph-devel, it says:
> rpm -q --provides -p ~/RPM/RPMS/x86_64/igraph-devel-0.5-8.fc8.x86_64.rpm
> igraph-devel = 0.5-8.fc8
>
> And IIRC, if I don't put any requires in python-igraph, it will say:
> igraph-devel = 0.5-4.fc8
>
> Which don't match. So I thought the correct thing to do was put
> provides/requires that strip off the release/dist and just say:
> igraph-devel = 0.5
In short, you don't have to write "Provides: igraph-devel = 0.5" in your
spec file. rpmbuild surely judges that igraph-devel-0.5-4.fc8 provides
"igraph-devel = 0.5".
And anyway "Provides: %{name}-%{version}" is wrong as this is expanded as
"Provides: igraph-0.5".
A example is
http://cvs.fedoraproject.org/viewcvs/*checkout*/devel/mecab/mecab.spec
http://cvs.fedoraproject.org/viewcvs/*checkout*/devel/ruby-mecab/ruby-mecab.spec
http://koji.fedoraproject.org/koji/getfile?taskID=393045&name=root.log
Regards,
Mamoru
More information about the fedora-devel-list
mailing list