Andreas Dilger wrote: >On Jan 11, 2002 19:59 -0500, Hubert Chan wrote: > >>Andrew> Perhaps file(1) should know about such packages, and call an >>Andrew> extracter, and reinvoke itself? >> >>Yes, so then we need to write another extension to file(1)... >> > >>file also doesn't seem to (currently) handle uncompressing files, and >>checking the type of the decompressed contents, so that's another >>extension that needs to be written. >> > >Well, "file -z" should do this already. > >>schedule.tex: LaTeX 2e document text >>schedule.tex: text/x-c++; charset=us-ascii >> >>I guess I can give this to them, since the file with no -i gave a >>correct answer. The discrepancy is strange, though, and perhaps it >>illustrates the hardness in keeping a correct set of databases? >> > >Well, part of the issue is that "file -i" uses a different (much smaller) >magic file, so it may be that the LaTeX magic is not in there yet. > >Granted, magic is not foolproof (as we all know), but it is mostly good >enough, and if it would (in addition, probably as an option) use the >file extension to help classify some files that would be best. > >Cheers, Andreas >-- >Andreas Dilger >http://sourceforge.net/projects/ext2resize/ >http://www-mddsp.enel.ucalgary.ca/People/adilger/ > > > The file command is the wrong answer, you don't determine the type in C by looking at the contents unless you are running gdb, and the filesystem is much the same.
Hans
