hello,
I downloaded 2.92, replaced the numpy spkg by the one provided by
joshua, undet SAGE_FORTRAN, and issued make, but I still have the same
issue :
building extension numpy.core.multiarray sources
Generating build/src.linux-i686-2.5/numpy/core/config.h
error: don't know how to compile Fortran
Ok. I'm going to have to trace through the numpy distutils to try to
track down where that is exactly generated to get a better idea why
its showing up,
that will take some time.
The binaries work for you right?
On Jan 6, 7:01 am, Johannct [EMAIL PROTECTED] wrote:
hello,
I downloaded 2.92,
I did not try the binaries yet, but I had built the version 2.8.1 from
sources, and this issue did not show up You might want to follow
the modifs in the distutils then.
best,
Johann
On Jan 6, 11:56 am, Joshua Kantor [EMAIL PROTECTED] wrote:
Ok. I'm going to have to trace through the numpy
We modify numpy so that it is supposed to use sage_fortran, which
wraps either the g95 fortran we include
or a fortran compiler the user specifies. The point of this is to
avoid problems with the user having
multiply incompatible fortran compilers (since there are a bunch)
Unfortunately
hi Josh,
unsetting the env vars did not change anything seemingly. So I am now
pointing to gfortran the way you described and I am rebuilding
everything (id est I just made a make distclean).
BTW, I build LPAACK and ATLAS on my own for scipy recently, and it was
absolutely critical to set my
On Jan 4, 6:18 pm, Johannct [EMAIL PROTECTED] wrote:
Hi Johann,
hi Josh,
unsetting the env vars did not change anything seemingly. So I am now
pointing to gfortran the way you described and I am rebuilding
everything (id est I just made a make distclean).
Good.
BTW, I build LPAACK and
ok, now it fails before numpy seemingly : it did not build all the
atlas librairies and bail out after complainng that it cannot cp
them... I relaunched a make, and it seems to build all the libs again,
which is strange...
Anyway :
cp
I rebuilt in place, and I just realized that I might have issued the
command make -j2 the first time, as I have a dual core machine. So I
am just remaking with this command just to see. If that fails I
will start from a fresh source again.
On Jan 4, 11:18 am, mabshoff [EMAIL PROTECTED]
On Jan 4, 2008 12:58 PM, mabshoff
[EMAIL PROTECTED] wrote:
On Jan 4, 9:52 pm, Johannct [EMAIL PROTECTED] wrote:
fresh rebuild from a vanilla source, LD_LIBRARY_PATH, PYTHONPATH
etc... unsetenved, and using make and *not* make -j2 and using
SAGE_FORTRAN set to gfortran, I still stumble
On Jan 4, 9:52 pm, Johannct [EMAIL PROTECTED] wrote:
fresh rebuild from a vanilla source, LD_LIBRARY_PATH, PYTHONPATH
etc... unsetenved, and using make and *not* make -j2 and using
SAGE_FORTRAN set to gfortran, I still stumble :
make[3]: `libcblas.so' is up to date.
make[3]: Leaving
ok, I am doing a little debugging : the numpy error message comes from
numpy/distutils/fcompiler/__init__.py in the new_fcompiler function.
This function seems to be called with plat=None and compiler=gfortran.
As plat=None, the first thing that this method function does is
plat=os.name I
fresh rebuild from a vanilla source, LD_LIBRARY_PATH, PYTHONPATH
etc... unsetenved, and using make and *not* make -j2 and using
SAGE_FORTRAN set to gfortran, I still stumble :
make[3]: `libcblas.so' is up to date.
make[3]: Leaving directory `/data1/sources/sage-2.9.1.1/spkg/build/
strange... I just tried in the cluster of machines at work, where I
have a different shell environment, and have access to RHEL3 and 4
machines, and they all result in os.name returning posix and
sys.platform returning linux2
I will try your spkg today if I have still time, or tomorrow.
best,
hi Josh, here it is :
Sage subshell$ pwd
/data1/sources/sage-2.9.1.1/spkg/build/numpy-20071020-1.0.3.1.p3/src
Sage subshell$ cd ../../../../
Sage subshell$ ls
COPYING.txt data example.sage install.log ipython local
makefile matplotlibrc README.txt sage sage-2.9.1.txt sage-python
spkg
Hello Johann,
On Jan 3, 9:30 am, Johannct [EMAIL PROTECTED] wrote:
[EMAIL PROTECTED] sage-2.9.1.1]$ env
SSH_AGENT_PID=3331
HOSTNAME=localhost.localdomain
SAGE_ATLAS=/usr/local/atlas
This is wrong/unneeded since we now build ATLAS. Did you set this
before or after the build of scipy
Hi Michael,
I put it afterwards, because of the warning message present in the
output. As I said, I did not think there was any relevance, but if on
top of it it is wrong then the warning should go! For reference I am
putting the part of the log at the end of this email.
As for sage_fortran.bin,
On Jan 3, 4:33 pm, Johannct [EMAIL PROTECTED] wrote:
Hi Johann,
Hi Michael,
I put it afterwards, because of the warning message present in the
output. As I said, I did not think there was any relevance, but if on
top of it it is wrong then the warning should go! For reference I am
I think I only installed ifort
On Jan 3, 9:33 am, mabshoff [EMAIL PROTECTED]
dortmund.de wrote:
On Jan 3, 4:33 pm, Johannct [EMAIL PROTECTED] wrote:
Hi Johann,
Hi Michael,
I put it afterwards, because of the warning message present in the
output. As I said, I did not think there
On Jan 3, 6:43 pm, Johannct [EMAIL PROTECTED] wrote:
Hello Johann,
I think I only installed ifort
In case you have admin rights on the box could you move it out of
$PATH and try again? That seems to be the best lead so far. I hope
Josh has an idea how to fix this if it really turns out
hi there,
ok I can't debug the distutils config, it is all arcane for me, but I
noticed that there is a sage-fortran in /data1/sources/sage-2.9.1.1/
local/bin/ , which is just a shell wrapper :
Sage subshell$ more /data1/sources/sage-2.9.1.1/local/bin/sage_fortran
#!/bin/sh
sage_fortran.bin
Hmm. That is very weird that it saw your system as posix and not
linux.
Out of curiosity what does uname -a output on your system.
On Jan 2, 5:34 pm, Johannct [EMAIL PROTECTED] wrote:
hi,
I downloaded sage-2-9-1-1 and I am running on the following system :
UNAME: Linux
21 matches
Mail list logo