[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Help compiling CVS 2.9.3

From: Quentin Spencer
Subject: Re: Help compiling CVS 2.9.3
Date: Fri, 28 Oct 2005 12:47:43 -0500
User-agent: Mozilla Thunderbird 1.0.7-1.1.fc4 (X11/20050929)

Andy Adler wrote:

I'm trying to get the CVS to compile with all the sparse matrix
features on a Xandros (debian) box. I'm having trouble with
cholmod and metis.

Here is my story, and where I got stuck.

First, I needed to patch ./
   address@hidden:~/octave$ cvs diff ./
   RCS file: /cvs/octave/,v
   retrieving revision 1.491
   diff -r1.491

Next, I did the following (from my .bash_history)
   # Get normal libs
   apt-get install libreadline5-dev
   apt-get install gperf flex bison dejagnu

   # Get ufsparse - not available for debian
   alien -i ufsparse-0.9-3.fc5.i386.rpm
   alien -i ufsparse-devel-0.9-3.fc5.i386.rpm

   # Get metis
   cd ~
   tar xfz metis-4.0.tar.gz
   cd metis-4.0
   vi Lib/rename.h  # NEED TO EDIT __log2 to __int_log2 TO AVOID CONFLICT
   cp libmetis.a /usr/local/lib/
   cp */*.h /usr/local/include/metis/
   chmod a+r /usr/local/include/metis/* /usr/local/lib/libmetis.a
   vi /usr/local/include/metis/metis.h  # NEED TO EDIT "metis/" INTO 4 LAST 

My trouble now is the following (from config.log):

   configure:14265: checking for cholmod_start in -lcholmod
   configure:14303: gcc -o conftest -g -O2   conftest.c -lcholmod -lamd -lcolamd 
-lccolamd -lblas  -L/usr/lib/gcc-lib/i486-linux/3.3.5 
-L/usr/lib/gcc-lib/i486-linux/3.3.5/../../.. -lz -lfrtbegin -lg2c-lm -lgcc_s -lmetis 
-lz -lm  >&5
 In function `__i686.get_pc_thunk.bx':
   : multiple definition of `__i686.get_pc_thunk.bx'
 first defined here

Any ideas?

OK, I tried configuring current CVS on my Fedora box with ufsparse installed and no metis, and it passed all of the ufsparse-related tests, including this one. It's not obvious to me what the source of your problem here is, but I did notice something I hadn't thought of before. The Fedora Extras development RPMS are build on gcc 4 (and a particularly bleeding-edge version of it), and it appears you're using 3.3.5. That could probably cause plenty of other problems even if it isn't causing this one. I'd suggest building ufsparse locally and then see what results you get.


Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:
How to fund new projects:
Subscription information:

reply via email to

[Prev in Thread] Current Thread [Next in Thread]