python-paver requires python-devel for most things to work. so it should be a requires. (see http://koji.fedoraproject.org/koji/getfile?taskID=913072&name=build.log)
A couple thoughts: * python-devel shouldn't be required by paver an more than glibc-devel should be required by make. * This is actually a distutils requirement rather than a paver requirement. James, this came up before during Jeremy's maintenance of python WRT setuptools:: https://bugzilla.redhat.com/show_bug.cgi?id=251645#c18 The basic problem is that distutils uses the Makefile and those pyconfig*.h files as data sources. So certain distutils and setuptools (and now paver) commands fail when those files are not present. Do you think that moving those files to the python package would be okay?
It won't be the end of the world if python itself includes those 3 .h files, my biggist worries are: 1. This is the wrong fix, and the tools should be fixed. 2. This would be pretty unique, cpufrequtils is the only "normal" package I can find that puts .h files into /usr/include ... and that looks like a packaging bug.
Ian, sorry, I've decided I'm not going to have paver depend on python-devel for the reasons I gave in c#1. We can see what upstream python has to say, though. If they're willing to change where the data being read from the Makefile and pyconfig.h is located then we're all set. If not we'll have to see. I've opened a bug upstream: http://bugs.python.org/issue4359
This bug appears to have been reported against 'rawhide' during the Fedora 10 development cycle. Changing version to '10'. More information and reason for this action is here: http://fedoraproject.org/wiki/BugZappers/HouseKeeping
I'm going to close this as WONTFIX so I can get it out of my bugs list. Thanks for reporting upstream, Toshio.