Description of problem: Looks like this spec file works. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Created attachment 138350 [details] spec
How exactly did you get around this issue: http://projects.scipy.org/pipermail/scipy-dev/2006-August/006070.html I'm seeing the same thing with a 1.0 release build...
*** This bug has been marked as a duplicate of 212147 ***
(In reply to comment #2) > How exactly did you get around this issue: > > http://projects.scipy.org/pipermail/scipy-dev/2006-August/006070.html > > I'm seeing the same thing with a 1.0 release build... I don't recall, and I'm busy now - please try my spec and see if that answers the question
Nope, same thing happens with both my spec and your spec.
Created attachment 139408 [details] spec Try this. IIRC, it seems that the original separation of build/install is broken. If build goes ahead and does install, then the problem isn't there.
Hrm, okay, that does work. Looks like specifically, the --skip-build option is busted. Omitting that and adding in all the assorted flags from %build under %install works around this and feels less dirty than doing the install in %build. I'll go ahead and push this into rawhide for now. -- spec snippet -- %build ATLAS=%{_libdir} FFTW=%{_libdir} BLAS=%{_libdir} \ LAPACK=%{_libdir} CFLAGS="$RPM_OPT_FLAGS" \ %{__python} setup.py build %install rm -rf $RPM_BUILD_ROOT #%{__python} setup.py install -O1 --skip-build --root $RPM_BUILD_ROOT # skip-build currently broken, this works around it for now ATLAS=%{_libdir} FFTW=%{_libdir} BLAS=%{_libdir} \ LAPACK=%{_libdir} CFLAGS="$RPM_OPT_FLAGS" \ %{__python} setup.py install --root $RPM_BUILD_ROOT