On Tue, 2008-02-19 at 22:55 -0800, Matthew Mueller wrote: > On Tue, Feb 19, 2008 at 09:31:41PM +0100, Alain Kalker wrote: > > When verifying .md5 or .cfv files which contain filenames containing > > chars with values > 127 (in ISO-8859), cfv reports the corresponding > > files as missing. > > I don't know if this problem is related to bug #406761, but that bug > > suggests a specific problem with the encoding of UTF-8 characters in > > .torrent files. If these problems are related, please feel free to merge > > the bugs. > > It's similar to that bug, but not the same cause. In this case, cfv > (1.x) has no understanding of encodings for text checksum files, and > since the files are encoded in iso-8859-? but your system is probably > using utf-8 for filenames, the bytes just aren't going to match up. For > the text-based formats you can work around this pretty easily by > manually converting the files: > "iconv -f iso-8859-1 -t utf-8 foo.md5 > foo.utf8.md5" > > Therefore, I say that this isn't really a bug at all. cfv 1.x certainly > can verify files with chars > 127 as long as the encoding matches the > one used by your system. > > If you want, you can consider it as a feature request for built-in > encoding handling. Actually, the cfv 2.x devel code already has > encoding knowledge so you could just do: > "cfv --encoding=iso-8859-1 -f foo.md5" > However I'm (slowly) doing some major refactoring on it and it's not > ready for release yet :(
Thanks for the very quick reply and helpful workaround! Yes, my system is setup to use UTF-8, and the checksum files I used use one of the iso-8859 encodings. Perhaps it would be a good idea to mention your workaround in README.Debian, so others faced with this problem don't have to check bug reports for the solution. Please consider the encoding issue a feature request, and good luck with your rewrite. Kind regards, Alain -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
