NB: this also raises theValueError for file names consisting entirely of slashes (which are considered directories, even if they actually contain data), which seems correct to me, but (unlike getting rid of theIndexErrors) could potentially break some existing code (though all tests are green).

>>>importzipfile>>>fh=zipfile.ZipFile("x.zip")>>>fh.infolist()[<ZipInfofilename='///'filemode='drwxrwxr-x'external_attr=0x10>]>>>fh.extract("///","some_dir")ValueError:Emptyfilename.

Note thatunzip fails to extract file names consisting entirely of slashes as well though:

$ unzip x.zipArchive:  x.zipwarning:  stripped absolute path spec from ///mapname:  conversion of  failed