Quoting Efraim Flashner (2017-01-09 05:14:29)
> On Sun, Jan 08, 2017 at 04:09:17PM -0500, Troy Sankey wrote:
> > I'm trying to fix the python-gpg package.  Latest build [0] was a
> > failure because gpgme.h claims gpgme was compiled with _FILE_OFFSET_BITS
> > = 64, implying the current build (python-gpg) doesn't define any
> > _FILE_OFFSET_BITS (it should also be set to 64, I think).
> > 
> > Relevant build log snippet:
> > 
> > ---- cut here / start ----
> > swigging gpgme.i to gpgme_wrap.c
> > swig -python -py3 -builtin -threads -outdir gpg -o gpgme_wrap.c gpgme.i
> > gpgme.h:104: Error: CPP #error "GPGME was compiled with _FILE_OFFSET_BITS = 64, please see the section "Largefile support (LFS)" in the GPGME
> > manual.". Use the -cpperraswarn option to continue swig processing.
> > error: command 'swig' failed with exit status 1
> > ---- cut here / end ----
> > 
> > Attached is a patch which forces _FILE_OFFSET_BITS = 64 and
> > _LARGEFILE_SOURCE = 1, as per the gpgme documentation [1].  This fixes
> > the build on my laptop (i686), but I'm quite sure this is a bad hack.
> > 
> 
> Since your laptop is actually 32-bit, does python-gpg build without the
> patch? If it does build without the patch, then it means that the
> 64-bit-ness of the build-machine on hydra is leaking into the build
> environment, and it should be set to the 32/64-bit-ness of the
> %target-system.

No, without the patch my 32-bit system produces the same error as
hydra's 32-bit build.

Troy