I have several Solaris fixed I'd like to get integrated. Most are based 
around the fact that the Sun linker takes different options to the GNU 
linker, and some packages make assumptions about the linker being the 
GNU one.



If you try to build Sage on t2, you will find it will fail in several 
places. Fixes for all those I know about can be found here.

http://sage.math.washington.edu/home/kirkby/Solaris-fixes/

Most are pretty easy to understand, as they only involve a few lines.

1) Nice simple polybori patch

http://sagetrac.org/sage_trac/ticket/6437

This is *only* applied on Solaris if the liner is the Sun linker. The 
patch tests if the linker is the Sun linker by seeing if

$ ld --version

generates an error or not. That is perfectly acceptable to the GNU 
linker, but not for the Sun linker.

After the test, the right linker flags are selected (-soname for the GNU 
linker, -h for the Sun linker).

2) Pari patch

http://sagetrac.org/sage_trac/ticket/6445

This one is quite simple to understand, though the initial explanation 
does not convey this, as I did not originally know the problem.

The patch does two things.

i) Add '-fPIC' as a compiler flag when trying to build a shared library. 
It does this by direct editing of a file with 'sed'

ii) Removes a previous unsuccessful attempt to do the same by adding 
-fPIC to CFLAGS, as that does not get propagated properly to the file 
that actually needs it.


3)  zn_poly problem.

http://sagetrac.org/sage_trac/ticket/6443

This was an easy fix - I just substituted -soname for -h in 
src/makemakefile.py. I did this with a sed script in spkg-install


Dave




--~--~---------~--~----~------------~-------~--~----~
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-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---

Reply via email to