John Hunter wrote:
"Gary" == Gary <pajer@...97...> writes:
> I changed axes.py to remove 'min' from the 'from numerix
> import ...' line.
> Now it chokes on the very next line, from numerix import max
> as nxmax
Let's be systematic. From the python shell, what happens when you do
>>> import Numeric, MLab
>>> Numeric.__version__
>>> from MLab import min, max
>>> import matplotlib
>>> matplotlib.__version__
>>> from matplotlib.numerix import min, max
C:\Documents and Settings\Gary>python
Python 2.3.4 (#53, May 25 2004, 21:17:02) [MSC v.1200 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import Numeric, MLab
>>> Numeric.__version__
'23.6'
>>> from MLab import min, max
>>> import matplotlib
>>> matplotlib.__version__
'0.65'
>>> from matplotlib.numerix import min, max
Traceback (most recent call last):
File "<stdin>", line 1, in ?
ImportError: cannot import name min
This will help narrow down whether it's a matplotlib or Numeric
problem. Check the reported version numbers to make sure the numeric
upgrade went as expected.
Do you have only one python installed on your system?
No. IBM (this is a thinkpad) has installed a python in c:\ibmtools\python22
I'm guessing they use it for some kind of maintenance.
Up to now, it's existance has been almost entirely benign, except for a warning I get when starting ipython. There's an issue somewhere with the case of the file fcntl.pyc (vs FCNTL.pyc) and/or the pythoncaseok environment variable. I swatted it down a couple of times, but it keeps coming back. I realize that things aren't exactly as they should be, but nothing has broken ... yet.
IBM has set pythonpath for me:
C:\Documents and Settings\Gary>echo %pythonpath%
C:\IBMTOOLS\utils\support;C:\IBMTOOLS\utils\logger
However, if I un-set %pythonpath% I still can't import pylab. (No change in behavior)
%path% references c:\ibmtools\python22, but only after all the listings for c:\python23. AFAICT no other environment variable references the IBM python.
> Am I the only one suffereing from this? Isn't this impossible?
My win32 setup appears to be working fine. It's clearly not
impossible <wink>. Can we get some positive confirmation from other
win32 users that 0.65 is working for them?
JDH
Thanks for your help on this. Since I appear to be an isolated case, shall we take this off-group?