[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: "checking for cholmod_start in -lcholmod... no"
From: |
Electronnix |
Subject: |
Re: "checking for cholmod_start in -lcholmod... no" |
Date: |
Wed, 14 May 2014 20:27:20 -0700 (PDT) |
Hello Mike,
Thank you for answering.
In order to avoid cluttering this thread, I posted the full octave
dependency building and configure log as well as the bash script I used for
this task, on another website.
Here are the links:
http://www.cobrasov.com/CoBra%20Project/SuiteSparse_compiler
(this is the bash script)
http://www.cobrasov.com/CoBra%20Project/Octave_Dep_Build+Configure.log
(this is the log resulted from the script)
Basically this script compiles suitesparse and then creates 2 directories:
/usr/lib/suitesparse
/usr/include/suitesparse
The .a libraries are copied to /usr/lib/suitesparse and then converted to
.so libraries.
The headers are copied to /usr/include/suitesparse.
I had to create symlinks for libgfortran.a and libgfortranbegin.a in
/usr/lib pointing to these libraries which in my system are located in
/usr/lib/gcc/i486-linux-gnu/4.1/
Then the script builds GraphicsMagick in /opt/GraphicsMagick-1.3.7,
and FLTK, GLPK and QRUPDATE in /usr/local.
At the end is the octave configure command.
As I mentioned previously, I do have cholmod libraries in my system which
were installed the standard way from Debian packages. The command "ldconfig
-p | grep cholmod" gives the following result:
=========================
libcholmod.so.3.1.0 (libc6) => /usr/lib/libcholmod.so.3.1.0
libcholmod.so.1 (libc6) => /usr/lib/libcholmod.so.1
=========================
but I expect octave to only look at and see the cholmod libraries that were
built from the suitesparse 3.4.0 compiled by my script. Am I too optimistic?
Anyway, I tried even deleting these libs from /usr/lib but the octave
configure gives the same message about cholmod_start not being found.
My script stops at the configure command because I tried to run make and it
runs fine, but the building process generated a huge amount of files and
when it reached about 700 MB it plugged up my ramdrive so I stopped it.
Anyway, the cholmod_start remains the only mystery in this affair.
--
View this message in context:
http://octave.1599824.n4.nabble.com/checking-for-cholmod-start-in-lcholmod-no-tp4663980p4664083.html
Sent from the Octave - General mailing list archive at Nabble.com.
- "checking for cholmod_start in -lcholmod... no", Electronnix, 2014/05/10
- Re: "checking for cholmod_start in -lcholmod... no", Tatsuro MATSUOKA, 2014/05/10
- Re: "checking for cholmod_start in -lcholmod... no", Tatsuro MATSUOKA, 2014/05/10
- Re: "checking for cholmod_start in -lcholmod... no", Cristian Savu, 2014/05/13
- Re: "checking for cholmod_start in -lcholmod... no", Mike Miller, 2014/05/13
- Re: "checking for cholmod_start in -lcholmod... no",
Electronnix <=
- Re: "checking for cholmod_start in -lcholmod... no", Mike Miller, 2014/05/15
- Re: "checking for cholmod_start in -lcholmod... no", Electronnix, 2014/05/17
- Re: "checking for cholmod_start in -lcholmod... no", Electronnix, 2014/05/17
- Re: "checking for cholmod_start in -lcholmod... no", Sergei Steshenko, 2014/05/17
- Re: "checking for cholmod_start in -lcholmod... no", Mike Miller, 2014/05/17
- Re: "checking for cholmod_start in -lcholmod... no", Electronnix, 2014/05/14