The SysV version of "file" is lame compared to the version used by some or all of the BSDs (and Mac OS X). I'll call that the DZ version, after the author (Ian Darwin) and current maintainer (Christos Zoulas).
The DZ version has a somewhat more powerful magic file format and a heck of a lot more entries. I haven't compared the hard-coded heuristics, nor done a detailed comparison of the magic files, but I have noticed that the DZ version usually manages to identify files that the SysV version fails to identify. One problem I see is whether anything else either uses the existing magic file format, or parses the "file" command output. For the first of those, the existing file could still be kept around. For the second...aside from whether they should be doing that anyway, that's tough. (I seem to recall that the old NeWSprint software used something very much like "file" and it's magic file to automatically identify how to convert files into PostScript for printing.) The DZ version of "file" has one other advantage: it tries fairly hard to use in its descriptions the word "text" for files that are pretty much readable without special tools, and "executable" for files that have the execute bit set and/or look like something executable in its format. That lends a certain clarity (although alas also might encourage the output to be examined by another program). -- This message posted from opensolaris.org _______________________________________________ opensolaris-discuss mailing list [email protected]
