Weird Python startup behavior between different drives on PowerPC platform

jwahlmann at gmail.com jwahlmann at gmail.com
Mon Jul 21 17:51:10 CEST 2008


Bump.  Anyone have any ideas on this?  My next step is to either link
together a static version of the compiler or create a debug version.

Thanks,
Jon

On Jul 18, 11:43 am, jwahlm... at gmail.com wrote:
> I'm experiencing some strange behavior when starting up python on a
> Debian-based PowerPC platform.  Normally, I operate from this platform
> with a root file system on an IDE flash drive (/dev/hda1).  However,
> I'm trying to get my system to run with root on a mechanical SATA
> drive (/dev/sda1).  Both are installed on a PowerPC board via a PMC
> daughter board.
>
> When running off the flash drive, the Python interpreter loads and
> runs just fine.  However, when running from SATA, the interpreter
> seems to have problems with importing things like site, os, etc.  I've
> played around with PYTHONHOME to no effect.  I even went as far as
> setting PYTHONHOME to some off the wall location (so no stdlibs will
> load) and invoking Python as:
>
> host$ PYTHONHOME=/tmp python -d -v -S
> # installing zipimport hook
> import zipimport # builtin
> # installed zipimport hook
> Python 2.4.4 (#2, Apr  5 2007, 19:01:44)
> [GCC 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)] on linux2>>> i=0
>
>   File "<stdin>", line 1
>     i=0
>      ^
> SyntaxError: invalid syntax
>
>
>
> You'll note that doing something as simple as setting "i=0" results in
> a syntax error.  If I run the same thing on the IDE flash drive, it
> works:
>
> host$ PYTHONHOME=/tmp python -d -v -S
> # installing zipimport hook
> import zipimport # builtin
> # installed zipimport hook
> Python 2.4.4 (#2, Apr  5 2007, 19:01:44)
> [GCC 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)] on linux2
>
> >>> i=0
>
> Any ideas?!  What could possibly cause a difference of operation
> between running from a IDE flash drive (/dev/hda1) vs. SATA (/dev/
> sda1)?  Other than toggling hda1 vs. sda1 in /etc/fstab, the
> filesystems are being built identically.  Could it be that shared
> libraries aren't being loaded correctly or fast enough off the SATA
> drive vs. flash?  Shooting in the dark here...
>
> Thanks,
> Jon Wahlmann




More information about the Python-list mailing list