Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread David Cournapeau
Mike Landis wrote: > Cygwin is present, so not just the dumbed down Windows CMD available. You should not use cygwin: if you use cygwin, it will build numpy against the cygwin python, or worse, will be very confused, because you will mix cygwin compilers and mingw compilers. Unless you want to bu

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread josef . pktd
On Sat, Jan 3, 2009 at 1:07 AM, Mike Landis wrote: > Cygwin is present, so not just the dumbed down Windows CMD available. > > I ran the numpy-1.2.1 superpak. Verified that it installed (cause you don't > get near as much output as you do from a shell prompt) by running: > >" > > python -c 'i

Re: [Numpy-discussion] installing numpy with Python2.5 already present

2009-01-02 Thread josef . pktd
On Sat, Jan 3, 2009 at 12:24 AM, David Cournapeau wrote: > On Sat, Jan 3, 2009 at 2:11 PM, Mike Landis wrote: >> This time I included both ATLAS and MKL in the site.cfg file and got >> a little further... D:\temp\numpy-1.2.1\site.cfg now looks like: >> ... >> ValueError: invalid version number '

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread Mike Landis
Cygwin is present, so not just the dumbed down Windows CMD available. I ran the numpy-1.2.1 superpak.  Verified that it installed (cause you don't get near as much output as you do from a shell prompt) by running:    "python -c 'import numpy; print numpy.__version__ '" and got the numpy version n

Re: [Numpy-discussion] installing numpy with Python2.5 already present

2009-01-02 Thread David Cournapeau
On Sat, Jan 3, 2009 at 2:11 PM, Mike Landis wrote: > This time I included both ATLAS and MKL in the site.cfg file and got > a little further... D:\temp\numpy-1.2.1\site.cfg now looks like: > ... > ValueError: invalid version number '2.18.50.20080625' This is a bug in python. Really, you should u

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread David Cournapeau
On Sat, Jan 3, 2009 at 12:10 PM, Mike Landis wrote: > Maybe the reason I'm having trouble is that I'm trying to get it > working on Windows, when almost everyone else is running on Linux? It is true that most developers use some sort of unix (linux, mac os X), but we definitely try to make sure i

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread David Cournapeau
On Sat, Jan 3, 2009 at 2:07 PM, Mike Landis wrote: > Does it sort of non-destructively overlay the 2.5 that I'm already running? > It only installs numpy into your existing python installation - it will not overwrite anything else in your python installation (e.g. everything else should work as b

[Numpy-discussion] installing numpy with Python2.5 already present

2009-01-02 Thread Mike Landis
This time I included both ATLAS and MKL in the site.cfg file and got a little further... D:\temp\numpy-1.2.1\site.cfg now looks like: [atlas] library_dirs = d:\Docs\ATLAS\build\lib atlas_libs = lapack, f77blas, cblas, atlas [mkl] include_dirs = D:\Programs\Intel\MKL\10.1.0.018\include library_d

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread Mike Landis
Does it sort of non-destructively overlay the 2.5 that I'm already running? At 11:45 PM 1/2/2009, you wrote: >There is a superpack for the python2.5 at the same page. Again a >binary .exe file that should make the installing a fair bit easier. > >Cheers > Tommy > >On Jan 2, 2009, at 11:26 PM,

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread Tommy Grav
There is a superpack for the python2.5 at the same page. Again a binary .exe file that should make the installing a fair bit easier. Cheers Tommy On Jan 2, 2009, at 11:26 PM, Mike Landis wrote: > Have to use Pyton 2.5 because I'm also using web2py. Python 2.5 and > a bunch of packages th

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread Mike Landis
Have to use Pyton 2.5 because I'm also using web2py. Python 2.5 and a bunch of packages that depend on it are already installed. At 11:05 PM 1/2/2009, you wrote: > > > >Is there any reason why you can not use the numpy-1.2.1-win32- >superpack-python2.4.exe >from the >http://sourceforge.net/proj

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread Tommy Grav
> Is there any reason why you can not use the numpy-1.2.1-win32- superpack-python2.4.exe from the http://sourceforge.net/project/showfiles.php?group_id=1369&package_id=175103 download page? I think that is what Mr. Kern meant by using the binaries. This will install already built code into the

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread Mike Landis
Maybe the reason I'm having trouble is that I'm trying to get it working on Windows, when almost everyone else is running on Linux? I have cygwin with f77, g++, make, ... installed, but it's definitely not a Linux machine. I'm working from the windows install documentation page. Maybe there

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread Mike Landis
David, You say "you should really use the distributed binaries" but I don't know how to accomplish even that from the install directions.  The Installing_SciPy page doesn't say what to do when you get errors or which warnings you can live with, so I don't know how to troubleshoot the install. If

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread Mike Landis
So, on page < http://www.scipy.org/Installing_SciPy/Windows#head-c7b42e8bc6116eb8108df13a49ab1ef6a883 >, what are the unspoken instructions that everyone else seems to know about (except me) that should be implemented between defining .numpy-site.cfg and running python setup.py config in the

[Numpy-discussion] Re : Alternative to record array

2009-01-02 Thread Jean-Baptiste Rudant
Thank you for everything, it works fine ant it is very helpful. Regards, Jean-Baptiste Rudant De : Francesc Alted À : Discussion of Numerical Python Envoyé le : Mardi, 30 Décembre 2008, 16h34mn 27s Objet : Re: [Numpy-discussion] Alternative to record array A

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread Matthieu Brucher
2009/1/2 Mike Landis : > Some of the install instructions are kind of ambiguous. > > When a library name ends in .a or .dll, it's obvious what it is, but > 'library' is sometimes used generically without indicating whether > you're talking about static or dynamic, e.g. how does numpy/scipy > link t

Re: [Numpy-discussion] building numpy/scipy

2009-01-02 Thread David Cournapeau
On Sat, Jan 3, 2009 at 3:29 AM, Mike Landis wrote: > Some of the install instructions are kind of ambiguous. > > When a library name ends in .a or .dll, it's obvious what it is, but > 'library' is sometimes used generically without indicating whether > you're talking about static or dynamic, e.g.

[Numpy-discussion] building numpy/scipy

2009-01-02 Thread Mike Landis
Some of the install instructions are kind of ambiguous. When a library name ends in .a or .dll, it's obvious what it is, but 'library' is sometimes used generically without indicating whether you're talking about static or dynamic, e.g. how does numpy/scipy link to MKL? Is it statically or dyn