Hello, An "alpha version" of a read-write tar filesystem is available at [1]. It can be used to "mount" almost any tar file, even compressed tar files. If you think typing "tar czvf file.tar.gz all my files" is too old-fashioned, you can even use it to create new tar files:
$ settrans -ca new /hurd/tarfs -cz newfile.tar.gz $ cp -r all my files new/ $ syncfs new Well, doing so is not really more efficient, but it's different at least. ;) In theory, it could also be used to compile a program without having to actually unpack it: $ settrans -ca chbouib /hurd/tarfs -j --volatile chbouib-0.1.tar.bz2 $ cd chbouib $ ./configure && make && cd .. && settrans -g tmp However, since gcc uses mmap () and since tarfs does not implement netfs_S_io_map, that doesn't work. It uses a gzip and a bzip2 stores, ie. libstore modules that can be used to transparently read and write gzip/bzip2 files, which were implemented using zlib and libbz2. A very basic test suite (shell script) is included in the tarball. However, tarfs *really* needs more testing. I would also be glad to get some feedback about the code itself (caching, for instance, could probably be implemented in a better way). I think it would be nice to make tarfs part of the GNU tar package which is probably the best place for it. On non-Hurd based systems, the 'configure' script could simply disable its compilation. Thanks, Ludovic. [1] http://people.type-z.org/ludo/hurd/ _______________________________________________ Help-hurd mailing list [EMAIL PROTECTED] http://mail.gnu.org/mailman/listinfo/help-hurd
