#9872: PolyBoRi incorrectly reports a GNU linker is used with gcc and produces
libraries with text relocations.
------------------------+---------------------------------------------------
Reporter: drkirkby | Owner: tbd
Type: defect | Status: new
Priority: major | Milestone: sage-4.6
Component: packages | Keywords:
Author: | Upstream: N/A
Reviewer: | Merged:
Work_issues: |
------------------------+---------------------------------------------------
Changes (by drkirkby):
* milestone: => sage-4.6
Comment:
Replying to [comment:1 AlexanderDreyer]:
> Is the !OpenSolaris environment somewhere available on sage.math?
>
> My best,
> Alexander
No, there are no virtual machines on sage.math with !OpenSolaris.
However, whilst not !OpenSolaris on x86, the issue with text relocations
is seen on Solaris 10 on SPARC too - i.e. t2.math
{{{
kir...@t2:64 ~/t2/32$ elfdump -d
./sage-4.5.3.alpha0/local/lib/libpolybori-0.6.4.so | grep TEXTREL
[19] TEXTREL 0
[27] FLAGS 0x4 [ TEXTREL ]
kir...@t2:64 ~/t2/32$
}}}
On !OpenSolaris, the Sun linker does actually take all the GNU options, so
getting the linker wrong is not a problem, so I would not waste too much
time over that.
The method chosen to get the linker is not ideal. It is based on something
I wrote, but I think my method is flawed. It assumes the first linker in
the path is the one used by gcc, but there is no reason for that to be so,
as gcc has the linker path hard-coded. I really no not know how best to
find the linker used by gcc. As a long term solution, it would probably be
worth looking at how autoconf does this.
If you set up your environment on t2.math as detailed in /etc/motd, then
!PolyBoRi will chose the right linker, but the text relocation problem
will be seen. This does not cause a problem with 32-bit builds, but it
does with 64-bit builds.
Since Sage is stable on 32-bit versions of Solaris 10 and !OpenSolaris, we
are looking to get it working 64-bit. The text relocation problem then
comes more serious. There's a discussion of this issue on this blog
http://blogs.sun.com/rie/entry/my_relocations_don_t_fit
What I found with ECL is that the Sun compiler will in fact produce a
binary that does not have the issue, but that does not help us with Sage,
as we must use gcc to compile a lot of Sage.
Dave
--
Ticket URL: <http://trac.sagemath.org/sage_trac/ticket/9872#comment:2>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica,
and MATLAB
--
You received this message because you are subscribed to the Google Groups
"sage-trac" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/sage-trac?hl=en.
