[sage-devel] Re: Building SAGE on NexentaOS?
On May 29, 8:50 am, didier deshommes [EMAIL PROTECTED] wrote:
Hi there,
NexentaOS (aka GNU/Solaris) is an open-source unix with the SunOS
kernel and GNU tools on top of it. Package management is done with
Debian's apt-get. I am investigating using NexentaOS as a file/media
server on a newish P4 (mainly because of zfs) and thought I would try
to build SAGE on it for fun. So far, the following packages compile
(some with small modifications):
bzip2-1.0.4
conway_polynomials-0.2
cremona_mini-0.1
dir-0.1
extcode-2.5.3
freetype-2.1.10
gd-2.0.33.p4
gdmodule-0.56.p4
givaro-3.2.6
gmp-4.2.1.p6
graphs-0.1
gsl-1.9
iml-1.0.1.p6
ipython-0.8.0.p1
libpng-1.2.16.p3
linbox-20070522 (has to compile w/o --enable-optimization)
matplotlib-0.90.0.p2 (requires python modification in pyconfig.h)
mercurial-0.9.3.p1
mpfi-1.3.4-rc3.p4
mpfr-2.2.1
mwrank-20061123
ntl-5.4.1.p3 (passes erroneous -Kpic flag to gcc instead of -
fPIC)
numpy-20070307
openssl-0.9.8d.p1
pari-2.3.1.cvs-20070305
pexpect-2.0.p1
prereq-0.2
pycrypto-2.0.1
python-2.5.1.p2
readline-5.2
sage_c_lib-2.5.3
sage_scripts-2.5.3
sagex-20070504
sqlite-3.3.17.p1
termcap-1.3.1
zlib-1.2.3.p2
However with Singular (more specifically libSingular), I'm hitting a
wall . Here's the output I get:
{{{
/usr/lib/gcc/i386-pc-solaris2.11/3.4.6/crt1.o: In function
`_start':crt1.s:(.text+0x59): undefined reference to `main'
/export/home/dfdeshom/custom/sage-2.5.3/local/lib/libreadline.so:
undefined reference to `tgetnum'
/export/home/dfdeshom/custom/sage-2.5.3/local/lib/libreadline.so:
undefined reference to `tgoto'
/export/home/dfdeshom/custom/sage-2.5.3/local/lib/libreadline.so:
undefined reference to `tgetflag'
/export/home/dfdeshom/custom/sage-2.5.3/local/lib/libreadline.so:
undefined reference to `tputs'
/export/home/dfdeshom/custom/sage-2.5.3/local/lib/libreadline.so:
undefined reference to `tgetent'
/export/home/dfdeshom/custom/sage-2.5.3/local/lib/libreadline.so:
undefined reference to `tgetstr'
collect2: ld returned 1 exit status
make[1]: *** [libsingular] Error 1
}}}
Any ideas? It looks like building SAGE on Nexenta should be possible.
Hello,
as far as I can tell these are symbols from the termcap library. Could
you post the exact command of the compiler/linker before the failure?
There have been several efforts to get SAGE to build and run on
proper Solaris, so I might be worthwhile that you post some patches.
Cheers,
Michael
--~--~-~--~~~---~--~~
To post to this group, send email to sage-devel@googlegroups.com
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://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/
-~--~~~~--~~--~--~---
[sage-devel] Re: Building SAGE on NexentaOS?
Hello, as far as I can tell these are symbols from the termcap library. Sorry, I should have written curses/ncurses library - too early for me. Could you post the exact command of the compiler/linker before the failure? There have been several efforts to get SAGE to build and run on proper Solaris, so I might be worthwhile that you post some patches. Cheers, Michael Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com 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://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: Building SAGE on NexentaOS?
The problem doesn't seem to be with libtermcap or curses. It's a flag
problem. When compiling libsingular, the --shared flag has to be
passed to gcc, instead of the default -export-dynamic. For that, the
Makefile in Singular/Makefile has to have a secial section for
nexentaos as follows:
{{{
ifeq ($(SINGUNAME),ix86-nexentaos)
SO_SUFFIX = so
LIBSINGULAR_FLAGS = --shared
LIBSINGULAR_LIBS = -lsingfac -lsingcf -lntl -lreadline -lgmp -lomalloc_ndebug
endif
}}}
And for singular to find ix86-nexentaos, singuname.sh has to have the following:
{{{
elif (echo $uname_a | $egrep SunOS $devnull)
then
# NexentaOS ###
if (echo $uname_a | $egrep NexentaOS $devnull)
then
#echo
echo ${prefix}-nexentaos
#echo
#exit 0
else
echo ix86-SunOS
#exit 0
fi
exit 0
}}}
Problem is, I don't know how to tell the make file Singular/Makefile
that --shared has to be passed to gcc for nexentaos. I could just
stick the first fragment showed above, but Singular/Makefile is
generated dynamically. How do I stick this piece of code in
Singular/Makefile dynamically?
didier
On 5/29/07, mabshoff [EMAIL PROTECTED] wrote:
Hello,
as far as I can tell these are symbols from the termcap library.
Sorry, I should have written curses/ncurses library - too early for
me.
Could
you post the exact command of the compiler/linker before the failure?
There have been several efforts to get SAGE to build and run on
proper Solaris, so I might be worthwhile that you post some patches.
Cheers,
Michael
Cheers,
Michael
--~--~-~--~~~---~--~~
To post to this group, send email to sage-devel@googlegroups.com
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://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/
-~--~~~~--~~--~--~---
[sage-devel] Re: Building SAGE on NexentaOS?
On 5/29/07, Martin Albrecht [EMAIL PROTECTED] wrote:
Hi, just stick it into Makefile.in. This is the source for the Makefile.
That did it and Singular built succesfully!
On 5/29/07, William Stein [EMAIL PROTECTED] wrote:
Wow, Didier, many thanks for helping with the Solaris port
of SAGE! I've been wanting to finish this forever, and got
stuck -- for example on linbox -- your help is *greatly* appreciated.
Some of my notes related to this are here:
http://sage.math.washington.edu/home/was/lj/todo.txt
Again, thanks!
I consider most of it to be pure luck (e.g. the linbox problem). I'm
seeing how far I can push it :)
Other packages that built:
extcode-2.5.3
freetype-2.1.10
gap-4.4.9
gd-2.0.33.p4
quaddouble-2.2.p6 (with some modifications. Solaris has a known bug
with isinfinity(double))
sage-2.5.3 (with some modifications to real_rqdf.p*)
sage_c_lib-2.5.3
sage_scripts-2.5.3
sagex-20070504
sqlite-3.3.17.p1
zodb3-3.6.0.p1
I'm now having trouble building clisp.
The errrors I get all look like this
{{{
:/export/home/dfdeshom/custom/sage-2.5.3/spkg/standard/clisp-2.41.p2/src/syscalls/calls.c:2151:
undefined reference to `uint64_to_I'
:/export/home/dfdeshom/custom/sage-2.5.3/spkg/standard/clisp-2.41.p2/src/syscalls/calls.c:2161:
undefined reference to `uint64_to_I'
}}}
I googled around and found that Joe Weening had already run into this
problem and a solution
(http://osdir.com/ml/mathematics.sage.general/2006-07/msg00011.html):
2. clisp-2.38/src/makemake.in sets BINARY_DISTRIB=1 for Solaris. This
causes an incorrect definition of uint64_to_I (and perhaps other
problems). Commenting out BINARY_DISTRIB=1 in makemake.in gave me
a working clisp on 32-bit x86 Solaris. I'm still having problems
with clisp on 32-bit and 64-bit SPARC Solaris.
His solution does not seem to work for me. Does anyone has an idea of
what to try? I wish I could just apt-get it but Nexenta doesn't seem
to have it (it's still an alpha release).
didier
--~--~-~--~~~---~--~~
To post to this group, send email to sage-devel@googlegroups.com
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://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/
-~--~~~~--~~--~--~---
