"Stephen J. Turnbull" <stephen(a)xemacs.org&gt; writes: There is a transparent way, however. Note that utf-8 is an encoding scheme that can, even within 4-byte values, encode more than just legal utf-8. Pick a 256-byte code page from there (either beyond the 2^21-something threshold, or, saving one byte but being more obfuscate, in the Unicode pages reserved for utf-16 surrogates and thus left free). Now this is our XEmacs-internal code page. _Any_ bytes that are not part of valid codes in a particular encoding (and this _includes_ non-minimal code sequences in utf-8 and utf-16) are encoded using this XEmacs-internal encoding into "bad byte of value xxx" and are displayed as \xxx octal escapes byte by byte. When writing out, such "byte" code points get encoded back into single bytes. Garbage in, garbage out. Meticulously preserved garbage. Note that XEmacs-internal utf-8 in this case is not necessarily legal utf-8 because of those surrogate byte-code patterns not corresponding to legal utf-8. The advantage is that a _legal_ utf-8 file is represented as itself XEmacs-internally. And that even random byte streams can get read and rewritten without modification. Is it stellar? Dealing with garbage never is stellar. But it certainly is better than ignoring the problem or calling it somebody else's. -- David Kastrup, Kriemhildstr. 15, 44793 Bochum _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta