[omniORB] RPM spec: python 2.0 dependency?

Bastiaan Bakker bastiaan.bakker at lifeline.nl
Tue Jul 8 16:46:31 BST 2003


On Tue, 2003-07-08 at 16:23, Thomas Lockhart wrote:
> >
> >
> >The RPM spec in contrib/RPMs specifies a build requirement of Python
> >2.0. On RedHat 7.3 this causes a build failure, since that platform has
> >Python 1.5.2.
> >If I recall correctly omniORB would be perfectly happy with 1.5.2. Maybe
> >the '2.0' is a Mandrake(?) specific artifact? 
> >If so, can it be fixed?
> >
> I'm not certain if "contrib/RPMs" is the same as the RPMs Sander and I 
> are building and posting. The spec file was developed from one done by 
> others for 3.x omniORB (so there is nothing Mandrake-specific about its 
> origins). I haven't run with pre-2.0 Python in a long time, and without 
> testing do not know whether in fact there is no problem with doing so.
> 

The specfile in contrib is your one with a change by Duncan. The
changelog in the specfile says:

%changelog
* Tue Jun 10 2003 Duncan Grisby <dgrisby at apasphere.com> 4.0.2pre1
- Fix some text, bump version number, add init script, minor tweaks.
 
* Wed Feb 12 2003 Thomas Lockhart <lockhart at fourpalms.org> 4.0.0
- Rename catior to catior.omni to avoid name conflict with TAO
 
Since you're building for Mandrake I assumed the 2.0 requirement had
something to do with it. But apparently you didn't put in the 2.0
version requirement...

> If you (or someone) has a chance to do at least moderate testing then 
> report back and we can update the spec file to relax the version 
> requirement.
> 

According to omniORB docs 1.5.2 is sufficient. At least for the C++
stuff I haven't had any problem with building and using omniORB on a
RedHat 7.3 platform (= Python 1.5.2) after relaxing the Python version
requirement in the specfile.


Cheers,

Bastiaan

> Or, update your copy of the spec file then rebuild and send 'em to me 
> and we'll post the built RPMs so others in the same rev-locked boat can 
> try them out.

> 
>                              - Tom
> 




More information about the omniORB-list mailing list