I tried setting up Oxygen for VFAT, and it works - and required very little updates. Dachstein should be simple, too. Main things to do and to remember are:
1. Include VFAT support with MSDOS support in the kernel when compiling. There is a space cost, but this may be worth it. 2. Change configuration files, lrp.cfg, whatever to use vfat instead of msdos: PKGPATH=/dev/fd0u1680:vfat ...or whatever... 3. Make sure boot time scripts don't assume msdos. In most cases, just changing msdos mounts to vfat mounts should be fine. 4. If you have it, make sure /etc/fstab uses vfat instead of msdos. Underlying all this is the (valid) assumption that if vfat fails the kernel will use the msdos format. Of course, if you now start creating: nc-1.2.3-beta7.lrp files (or whatnot) then anybody using Eiger / EigerStein / LRP / etc. etc. etc. will be unable to use these new files. Of course, Oxygen has already done some of this by switching to glibc 2.1 (sigh). I think, if we switch to long filenames, it will be a benefit. Perhaps the description files could contain a line: ShortName: nc ...in fact, this is probably a good idea... A problem with long names is that this "name" propogates into myriad other places: /var/lib/lrpkg/<name>.* (including .list, .version, .help, .desc, etc....) Contents of /var/lib/lrpkg/<name>.list Contents of /var/lib/lrpkg/packages Perhaps we can stick to a format of: name-ver-rel.lrp ...or anything where the name does NOT have a '-' in it; then the "shortname" would be everything up to the first '-' ... What do you all think, now that a VFAT-based system is proven and functional? _______________________________________________ Leaf-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-devel
