[Christoph Biedl] > Well, file has an option > > | -z, --uncompress > | Try to look inside compressed files. > > but it doesn't give much help since > > | $ file -z ./data/examples/himno_de_riego.rg > | ./data/examples/himno_de_riego.rg: XML 1.0 document, UTF-8 Unicode text, > with very long lines (gzip compressed data, was "himno_de_riego.rg.xml", last > modified: Wed Feb 23 00:22:42 2011, from Unix) > > So I'm considering to modify file(1) to pick the string that follows > the first "doctype". Does that sound like a solution for you?
Yes. It would make it possible to recognize some more file types, at definitely the Rosegarden format. Might be a good idea to recognize the toplevel XML tag (here <rosegarden-data>) ro recognize even more formats. But requiring -z for this to work is not a good idea, as most tools using file to figure out the file format will use file without the -z option. Perhaps some formats should be uncompress for a peek automatically? Might even make it possible to recognize tar.gz files as such. :) -- Happy hacking Petter Reinholdtsen
