Referring to <http://lists.gnu.org/archive/html/bug-gnu-utils/2011-07/msg00000.html>:
An obvious problem with the patch is that it considers a file name to be a byte sequence. But different users may work in different locales, with different encodings. If a Chinese user with file names in GB18030 encoding sends a file to a user whose file names are UTF-8 encoded, or vice versa, the file name needs to be converted. The usual approach for such cases is to use UTF-8 as a "pivot" encoding. For example, in 'pax' [1] file names are transferred in UTF-8 encoding. But actually, what's the point of the patch? The most frequently used archive programs for interchange are probably 'tar'/'pax', 'zip', and '7-zip'. - 'pax' has support for Unicode file names [1]; the biggest problem is that the 'pax' format is the default one for GNU 'tar'. - 'zip' has support for Unicode file names [2][3]. - '7-zip' supports Unicode file names as well [4]. Users who really want to transfer files with non-ASCII names can use one of these three archive formats and send an uuencoded archive. Bruno [1] http://pubs.opengroup.org/onlinepubs/9699919799/utilities/pax.html [2] http://www.info-zip.org/UnZip.html [3] http://info.michael-simons.eu/2010/01/05/create-zip-archives-containing-unicode-filenames-with-java/ [4] http://www.7-zip.org/7z.html -- In memoriam Yuri Shchekochikhin <http://en.wikipedia.org/wiki/Yuri_Shchekochikhin>
