Re: [21.5] make hexl-find-file use find-file-literally

Samuel Bronson writes: Since .tar, .gz, and .tgz are all there already, I don't have much hope that it will help. ;-) IMO that's a broken idea. Coding-system detection applies only to text/* media. Files where magic number can be used to determine the coding are by definition binary. Eg, consider XML. I don't think it is useful to try to match "\000u\000t\000f\000-\0001\0006" and its swabbed version, when reading it as text will win nearly 100% of the time. Consider an RFC 822 message, where the header is in ASCII and the body can be in an arbitrary encoding (including non-ASCII compatible encoding) via Content-Transfer-Encoding and Content-Type. This is a binary format, no matter how much the mail you see in Kansas looks like text/*. For sure, it would be a bad idea to apply the MIME charset detected by matching "^Content-Type:\(.\|\n\s-+)*charset" in an mbox file containing multiple messages! And the same thing is true of tar files. There is a facility called `format-alist' which matches regexps against a prefix of the file, but I'm not sure at what stage this happens. It might be useful here, but it depends on what decides to do autodetection of coding systems. It might be jka-compr; it might be the process code; it might be a find-file format handler. In the first two cases, it won't work, format-alist is only used for files, I think. In the latter, you have to be very careful or you'll end up giving somebody else the same problem you are experiencing. That's why I say this is hard. _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://lists.xemacs.org/mailman/listinfo/xemacs-beta