Le 05/02/2011 12:01, Dr. David Kirkby a écrit :
It would help if you opened tickets for each particular problem, but
have them linked to one ticket. For example:
http://trac.sagemath.org/sage_trac/ticket/9026
has the issues which are preventing 64-bit builds on Solaris.
I did open other tickets for the different problems, but on the various
upstream systems, and gave the links.
Reading through the ticket you reference, it is hard to see the wood
from the trees.
Search for the "make ptestlong" comments, that's where the latest
problems are.
I would point out one thing worth noting. Your ticket says "-m32" is an
x86 specific option. That is not true. It works on SPARC too and many
other platforms.
>
I'm also aware that on HP-UX (with at least the PA-RISC CPU), one needs
to build a different gcc for 32-bit and 64-bit executables. So whilst
gcc can build both 32-bit and 64-bit binaries, you need two copies of
gcc. They will not accept the -m32 or -m64 options.
This is something I learned later on ; it's in fact a platform-specific
flag, which happens to be common for several platforms. In any case, it
shouldn't be added unconditionally to the flags, but only when needed
and applicable.
Is ARM running big endian or little endian? If I understand correctly,
it may be possible to use either mode on those CPUs.
It can do both -- but not at the same time. The debian/ubuntu packages
are named "ARMel" though, with "el" standing for "endian little" (pun
intended, probably).
It builds, so I guess those questions are less important now.
Snark
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
