From kde-core-devel Fri Nov 16 13:29:36 2001 From: Rolf Magnus Date: Fri, 16 Nov 2001 13:29:36 +0000 To: kde-core-devel Subject: Re: RFC: kio_mount X-MARC-Message: https://marc.info/?l=kde-core-devel&m=100591728209320 On Friday 16 November 2001 13:20, Rob Kaper wrote: > The more I use KIO, the more I think it does not belong in KDE. It shou= ld > be available in a much broader context. It should live on top of the > kernel, as system service. I'm thinking the same. The features that kio provides aren't kde specific= =2E > It's a bit unlikely that will ever happen, but a nice compromise would = be > to have the ability to mount KIO slaves. This would give every applicat= ion > on the system access to any content behind KIO slaves. For example, you > could forego NFS alltogether and kio_mount a SFTP account. I also looked at this, but I don't know enough about kernel internals. My= idea=20 was similar, but a bit different. In my "vision" there is a kio filesyste= m=20 that you can just mount (e.g. to /mnt/kio) and that has a subdirectory fo= r=20 every kioslave. The kernel would create a device inode through which it=20 communicates with a kiofs daemon that creates the slaves. > Martijn Klingens informed me about the following page > > http://lwn.net/2001/1115/a/fuse.php3 > > which is about user-space mounting. Requires kernel 2.4 to compile, but > doesn't require a kernel recompilication to work. I found something called userfs, which is very old and not ported to 2.4.= It=20 also makes it possible to mount user space filesystems. > Ideas? Flames? Praise? I like it.