Determining package set for build

Mathieu Bridon bochecha at
Sat Sep 22 08:31:05 UTC 2012

On Saturday, September 22, 2012 03:45 PM, Richard W.M. Jones wrote:
> On Fri, Sep 21, 2012 at 08:21:29PM -0600, Orion Poplawski wrote:
>> I'm trying to track down a problem with scipy not building on F18
>> due to a test crashing.  However it builds fine in rawhide, and on
>> F18 using my local mock builder (which runs F17).  I'd like to
>> compare the package list for the buildroots for each build to look
>> for differences.  I can get this with a fair amount of difficulty
>> from root.log, but it seems like it would be nice if the output of
>> rpm -qa could be added to the mock logs and provided in koji.  Does
>> this seem reasonable?  Or perhaps another rpm -qa in root.log after
>> installing the buildrequires?  In the meantime, if anyone has any
>> root.log parsers to generate this, I'd be grateful.
> I guess the potential amount of information that could be logged by
> mock and would be useful to someone sometime is enormous!

And yet, this information is already available.

This is the latest failed build of scipy in F18:

In this page, there is a link to the buildroot used to attempt building 
the package:

And on that second page, there is a link to the RPMs installed in the 
buildroot (« Component RPMs »):


More information about the devel mailing list