Hi all,
On Windows, it's very common to have a string of long directories in the
pathname for files, like "C:\Documen ts and Settings\My Long User Name\My
Documents\My Long Subdirectory Name\...". For a wxPython application I'm
working on, this has actually caused me to run into what appears to be
Python's pathname length limit for opening files. (247 chars on Win) Yes,
I can hear people saying "yipes!" but this stuff does happen sometimes on
Windows. :-)
My first inclination was to use win32api.GetSho rtPathName(mypa th), which
worked fine until I had an unicode pathname with non-ascii characters in
it. Those give me the 'oridnal not in
range' errors performing an ascii encode, meaning that GetShortPathNam e
doesn't handle unicode objects. The problem is that I'm working with
Unicode filenames that contain characters that are not only non-ascii
characters, but characters not in the current locale's character set as
well. So I can't just 'down convert' from a unicode object to a string
object in the current character set without corrupting the filename.
Looking at the pyWin32 sources, it does look like only the ASCII version
of this function exists, which suggests that for now this route is a
dead-end.
The only other solution I could think of is to call
os.chdir(long_p athname) and open the file using a filename relative to
long_pathname instead of an absolute path. But I was wondering if there
was another solution, preferably one that doesn't require me to muck with
the current directory, and I was also wondering if there was a simple way
to get one or both of the above limitations removed. :-)
Thanks in advance for any help,
Kevin
On Windows, it's very common to have a string of long directories in the
pathname for files, like "C:\Documen ts and Settings\My Long User Name\My
Documents\My Long Subdirectory Name\...". For a wxPython application I'm
working on, this has actually caused me to run into what appears to be
Python's pathname length limit for opening files. (247 chars on Win) Yes,
I can hear people saying "yipes!" but this stuff does happen sometimes on
Windows. :-)
My first inclination was to use win32api.GetSho rtPathName(mypa th), which
worked fine until I had an unicode pathname with non-ascii characters in
it. Those give me the 'oridnal not in
range' errors performing an ascii encode, meaning that GetShortPathNam e
doesn't handle unicode objects. The problem is that I'm working with
Unicode filenames that contain characters that are not only non-ascii
characters, but characters not in the current locale's character set as
well. So I can't just 'down convert' from a unicode object to a string
object in the current character set without corrupting the filename.
Looking at the pyWin32 sources, it does look like only the ASCII version
of this function exists, which suggests that for now this route is a
dead-end.
The only other solution I could think of is to call
os.chdir(long_p athname) and open the file using a filename relative to
long_pathname instead of an absolute path. But I was wondering if there
was another solution, preferably one that doesn't require me to muck with
the current directory, and I was also wondering if there was a simple way
to get one or both of the above limitations removed. :-)
Thanks in advance for any help,
Kevin
Comment