Heiko Voigt <[EMAIL PROTECTED]> writes: > As I read from the standard a POSIX system only has to support the portable > characterset.
This has been a point of issue in the Open Group. Read one way, for example, the file name "a/-b/c" is not portable and need not be supported. But many people in the Open Group would say that (aside from length limits) every nonempty string of nonnull bytes is a valid file name. > It says: "The encoded values associated with the members of the portable > character set are each represented in a single byte. Moreover, if the value is > stored in an object of C-language type char, it is guaranteed to be positive > (except the NUL, which is always zero)." That's the requirement for portable characters. The requirement for portable file names is much stricter. See <http://www.opengroup.org/onlinepubs/009695399/basedefs/xbd_chap04.html#tag_04_06>. However, as I mentioned, this is a messy area in the spec. It's not clear that a POSIX implementation can refuse to create a file named "abc~", for example, simply because it doesn't like the "~". Certainly many applications would break on a system that didn't allow such file names. As far as 'tar' goes, I would say that by default tar should create the files with the names taken from the tar image, and should back off to mangled names only on user request. _______________________________________________ Bug-tar mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-tar
