From: Csaba H. <csa...@cr...> - 2005-12-12 02:47:32
|
On Sun, Dec 11, 2005 at 09:42:53PM +0100, Miklos Szeredi wrote: > fuse_file_info is not persistent, only the 'fh' member. There is > space for a flag in 'fuse_read_in' however, so it's solvable. I see. > Real question is: how useful is this? Reading the ml archives, my perception was that it's sort of a feature request, that's why I brought it up. > Do you have specific ideas in mind? Well, from the abstract-logical POV, FUSE is a tool for giving an OS interface to tree-like structures. If you look at tree-like structures as is, a restriction that only leaf nodes can have content is rather underjustified... > Do the OS's supporting this have any interesting use, or do > they just dump the directory as it appears on disk? From what I've seen they just dump it (although Heaven knows what those cunning guys at Sun are up into :) ), but I'd say it's just as irrelevant as the fact that Linux gives you an error. Csaba |