On 07/ 5/10 09:22 PM, Georg S. Weber wrote:
On 5 Jul., 18:03, "Dr. David Kirkby"<david.kir...@onetel.net> wrote:
On 07/ 5/10 11:14 AM, Fran ois Bissey wrote:
Hi,
In making a patch for ticket #9097 I simplified the
logic of the building sage_clib on OSX and 64 bit platforms.
It works for Dave on solaris 64 bits. But it would be nice
if we didn't have any surprise from the OSX side.
So if a friendly OSX tester could try the patch at
http://trac.sagemath.org/sage_trac/ticket/9097
that would be very nice.
Francois
There's no need, I have tested this on OS X - I have stuck evidence on the trac
ticket.
If we can get this along with #9399 merged, the Sage library will build on
OpenSolaris. In fact all of Sage apart from R and Maxima then builds, though it
does not run properly - giving a SIGSEGV, which means its trying to read or
write memory it does not have permission to do.
Dave
From "just reading", I would say that the interesting case left, is to
test this on a MacIntel with a 64-bit capable CPU (e.g. Core2Duo, but
not CoreDuo), OS X 10.5 / Leopard (which defaults to build as a 32-bit
OS), and "SAGE64" set to "yes" (i.e. wanting Sage to build in 64-bit
mode nevertheless). I don't have access to such a system
configuration, though.
Cheers,
Georg
No, me neither.
But I have built it on OS X 10.6, setting SAGE64=yes. That of course adds the
-m64 flag, which does not do change the binaries on a system which builds only
64-bit. However, it proves beyond reasonable doubt that the flags that would be
issued on the system you describe, would work. The compile line shows the -m64,
as it should do.
It was to avoid these complications that my original patch had
if env['PLATFORM'] != "darwin" and os.environ['SAGE64']=="yes":
as that basically means the code would only change on Solaris system. (Linux
users will not be using SAGE64 anyway, and that line excludes OS X users). Of
course I could have made it Solaris specific, but I'd not want to stop a port to
HP-UX or similar.
But I can't see anything wrong with François Bissey's code. It looks right, and
I'm unable to break it on any system I have access to. I agree with François it
looks cleaner than mine, though it does unfortunately make testing a bit more
difficult. But I'm as confident as I can be that it is ok.
Dave
--
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