Problem with directory recursion!

Robert Dailey rcdailey at
Fri Aug 17 20:19:13 CEST 2007

Here's part of the output that's incorrect:


Notice that the file BYOS_C.entity.xml is showing up as a file in the
directory W_BoomerEdge. This file does not exist in this folder, but yet in
a different folder. The output is consistent with the 'print' statements you
will see in the function I posted earlier.

On 8/17/07, Robert Dailey <rcdailey at> wrote:
> Hi,
> I've created a function that is used to recurse a directory tree in
> Windows XP using os.walk(). For the most part it works, however in some
> instances the data is incorrect and I'm getting invalid sub-directory paths.
> Here's the function:
> def __doSearch( root_dir, sub_path, restype, ext ):
>     print sub_path
>     # Searches the specified directory and generates a
>     # list of files to preload.
>     complete_path = osp.normpath( osp.join( root_dir, sub_path ) )
>     for root, dirs, files in os.walk( complete_path ):
>         # Record the list of file hash ID's
>         for file in files:
>             split = __resType( file )
>             if split[1] == restype:
> #                print sub_path
>                 print "\t", file
>                 __appendResource( ext, sub_path, split[0] )
>         # Remove .svn subdirectories; we don't walk these.
>         if ".svn" in dirs:
>             dirs.remove( ".svn" )
>         # Recurse child directories
>         for dir in dirs:
>             __doSearch( root_dir, osp.join( sub_path, dir ), restype, ext
> )
> Does anyone see anything immediately suspicious about the code? I'm
> assuming that in Python, every time a function is called recursively it gets
> its own copy of local variables. For some reason the sub_path variable isn't
> consistent depending on where I use it. For example, if you notice the print
> call at the very start of the function, it will output something like
> "models/ships". However, after passing it into __appendResource(), it
> appears to be just "models".
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <>

More information about the Python-list mailing list