[IPython-user] win32 installer woes...

[Fernando Perez]

Fernando Perez wrote:
> Hi all,
> 
> after Viktor's excellent win32 installer work, we have something which 
> _almost_ works.  Except I have a really nasty problem.  I have just posted a 
> message to c.l.py  asking for help, which I repost here in case anyone can 
> help or has a good solution.

OK, I got a solution.  It turns out the original os.chdir() instruction was 
completely unnecessary for the script to actually run.  So I simply removed 
it, and things work now fine regardless of where they run from.  In case 
anyone is interested in testing this fixed win32 installer, it is here:

http://amath.colorado.edu/faculty/fperez/tmp/ipython-0.6.9-1.win32.exe

Feedback on how this release works for Windows users would be greatly appreciated.

The old one (with the segfault if run from a different drive) is still at:
http://amath.colorado.edu/faculty/fperez/tmp/ipython-0.6.9.win32.exe

I still think this is a nasty bug somewhere in python's binary installer, but 
since this has cost me yet another Sunday dealing with the endemic brokenness 
of Windows, I won't pursue it further [1].  The fix I have works for me.  If 
someone cares to follow up with the upstream win32/distutils crowd, it would 
probably save others similar headaches in the future.

Cheers,

f

[1] There is a simple reason why I hate the Win32 platform with such a 
passion: _every_ time I've tried to get ipython to behave under windows in a 
rational manner, the simplest, most trivial things have always cost me an 
entire weekend.  And at the end of the day, without fail, I end up just having 
to build kludgy solutions to sidestep every one of the million ways in which 
Windows manages to break every computing standard known to man, without 
mentioning simple common sense.  And yes, I'm bitter :)