Created attachment 915852 [details]
Comment

(This comment was longer than 65,535 characters and has been moved to an attachment by Red Hat Bugzilla).

Command:

gcc -c  -I/usr/lib/perl5/vendor_perl/PDL/Core  -I/usr/include/plplot -I/usr/include -O2 -g -pipe -Wall -Werror=format-security -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong --param=ssp-buffer-size=4 -grecord-gcc-switches  -m32 -march=i686 -mtune=atom -fasynchronous-unwind-tables -g   -DVERSION=\"0.67\" -DXS_VERSION=\"0.67\" -fPIC "-I/usr/lib/perl5/CORE"   PLplot.c

Was it necessary to cut'n'paste such a huge part of the output into this ticket?

Have you tried _adding_ the optflags without wiping out existing CFLAGS?

(In reply to Michael Schwendt from comment #3)
> Was it necessary to cut'n'paste such a huge part of the output into this
> ticket?
> 
> Have you tried _adding_ the optflags without wiping out existing CFLAGS?

Not yet, will be done later. I've marked NotReady.

NotReady is meaningless for normal bug reports.


Building with

  make %{?_smp_mflags} OPTIMIZE="%{optflags} %{__global_ldflags}"

would be one solution. Of course, adding guard to the spec file that verifies whether the OPTIMIZE variable is still supported would be a good idea.

The many warnings you get are only due to the -Wall option in %optflags.

Taking. This is a trivial packaging bug.

It's an arched packages using a Makefile.PL.

Such packages are supposed to be built
%{__perl} Makefile.PL INSTALLDIRS=vendor OPTIMIZE="${RPM_OPT_FLAGS}"

Passing __global_ldflags usually is not necessary.

However, and this worse: When having fixed this bug, an FTBFS occurs.

I am going to apply patches to fix both.

http://koji.fedoraproject.org/koji/taskinfo?taskID=7134571