math/sage build error

[Montgomery-Smith, Stephen]

On 01/12/2014 02:26 PM, Greg Rivers wrote:
> On Sun, 12 Jan 2014, Montgomery-Smith, Stephen wrote:
> 
>> On 01/12/2014 01:19 PM, Greg Rivers wrote:
>>> math/sage does not compile successfully on
>>> FreeBSD 10.0-PRERELEASE #0 r260501: Fri Jan 10 11:39:36 CST 2014 amd64
>>>
>>
>> Other people have reported this error to me.  However I am unable to
>> reproduce it in RC5.  Do you know if you were using an earlier version
>> of FreeBSD-10?
>>
> 
> The host had never run any version of FreeBSD-10 before I updated it
> last Friday.  Prior to that, it had been tracking /stable/9, and was
> running 9.2-STABLE from ~September 29.
> 
> I updated to 10-STABLE from SVN /stable/10 sources using the standard
> procedure, including `make delete-old` and `make delete-old-libs`.  The
> update went perfectly and there were no errors.
> 
> I then deleted all ports and rebuilt them all from scratch.
> 
>> I do get an error building ecl with FreeBSD-10-RC5.  So there is still
>> work to go.
>>
> 
> Thank you very much for maintaining this port.  Let me know if I can
> provide any additional information that would help.
> 

Also, I only have FreeBSD-10 on a i386, so that might make a difference
as well.

The error message is:
/usr/local/bin/ld:
/usr/ports/math/sage/work/sage-5.13/local/lib/libgmp.a(sizeinbase.o):
relocation R_X86_64_32S against `__gmpn_bases' can not be used when
making a shared object; recompile with -fPIC

I don't see how this error could have crept into the build of sage
merely by using a different version of FreeBSD (because it builds just
fine with FreeBSD-8).  So either someone else can give me an idea, or it
will be a difficult problem to track down.