[Python-Dev] A wart which should have been repaired in 3.0? (original) (raw)

skip at pobox.com skip at pobox.com
Mon Dec 29 00:14:00 CET 2008

"skip" == skip <skip at pobox.com> writes:

Martin> I don't think any change is necessary. os.path.commonprefix
Martin> works just fine on path components:

skip> Ummm...

>>> os.path.commonprefix(["/export/home", "/etc/passwd"])
'/e'

skip> I suppose that's correct given the defined behavior of the
skip> function, but it certainly doesn't seem to be very path-like to
skip> me.

I should also point out that most people will not have the foresight to use it the way Martin demonstrated. Documentation or not, I'll be a fair fraction of all usage assumes the return value represents a valid path.

Martin> Of course, using it that way would require a library function
Martin> that reliably splits a path into components; I think one would
Martin> have to do abspath on arbitrary inputs.

Kinda what I think os.path.split ought to do. Should I tackle that next? ;-)

Skip

More information about the Python-Dev mailing list