Hi there, attr(1) is intended only for XFS - it is an IRIX-compatibility command, and there are parts of the tool that only work on the XFS filesystem (in particular, those options that deal with the ROOT namespace are XFS-specific).
On Tue, 2007-10-16 at 12:07 +0100, Reuben Thomas wrote: > Package: attr > Version: 1:2.4.39-1 > Severity: minor > > attr(1) says: "Extended attributes implement the ability for a user to > attach name:value pairs to objects within the XFS filesystem." > > In fact, other file systems can use extended attributes. I suggest you Indeed they can (the above sentence does not say they cannot, however). Two paragraphs down in the man page is this: "This document describes the attr command, which is mostly compatible with the IRIX command of the same name. It is thus aimed specifically at users of the XFS filesystem - for filesystem independent extended attribute manipulation, consult the getfattr(1) and setfattr(1) docu‐ mentation." > change "within the XFS filesystem" to "on supported filesystems, > including ...", and put a list of the commonest file systems that > support extended attributes. Or if most of them do, simply write "on > supported filesystems. Most commonly-used filesystems support extended > attributes." or "Most filesystems used on fixed media support extended > attributes", which I suspect is closer to the truth. Or simply stop > the sentence at "supported file systems." I'll go for f) "none of the above". :^) > By the way, with reference to bug #304765, a few words in attr(1) to > the effect that the file system must be mounted "user_xattr" would be > helpful. For example: Helpful perhaps, but not correct for all cases - this is a filesystem- specific mount option (other filesystems - like JFS/XFS - dont require it, while reiserfs/ext2/3/4 do). I do agree it would be useful to document this somewhere - the attr(5) man page is a general discussion of extended attributes on Linux, so it would seem a suitable place - do you want to send me a patch? > You could put this as the last sentence of the first paragraph of > attr(1). I think then you could safely close bug #304765. I think the paragraph on FILESYSTEM DIFFERENCES in attr(5) would be a better spot. And perhaps also on the getfattr/setfattr man pages? cheers. -- Nathan
