Hi Barry,
Could you merge this docs patch into the oss.sgi.com cvs tree please?
thanks mate.
---------------------------- Original Message ----------------------------
Subject: Bug#446877: attr(1) implies it only works with XFS
From: "Reuben Thomas" <rrt@xxxxxxxx>
Date: Wed, October 17, 2007 10:47 pm
To: "Nathan Scott" <nscott@xxxxxxxxxx>
Cc: 446877@xxxxxxxxxxxxxxx
--------------------------------------------------------------------------
On Wed, 17 Oct 2007, Nathan Scott wrote:
> All sounds good to me - please send a patch with those changes.
Attached.
To confirm what I've done:
0. I've replaced ".P" by ".PP", in the pages I've edited as according to
groff_man(7), they are aliases for the same command, and I wasted a few
minutes checking what was going on when I found them both. In attr.5, .PP
was already used exclusively, whereas in attr.1 both were.
1. In attr.1, I just moved the text that was already there mentioning
getfattr and setfattr further up the page (and then changed a couple of
words at the start of the next paragraph to avoid ambiguity).
2. In attr.5, I removed the links to filesystem patches, mentioned ext4, JFS
and reiserfs, and, lower down, added details of their attribute support. I
removed implementation details, so that I could for example refer to XFS and
reiserfs in the same paragraph, removed the hope that ext2/3/4's limits
would one day be raised (since the relevant code is still the same in ext4
as it was in ext2!), but also removed the erroneous implication that all
extended attributes must fit on a single disk block in those FSes (in fact,
*each* must fit on a single disk block, so in practice, for small tags of
the sort given in the examples, any of the file systems is fine).
--
http://rrt.sc3d.org/ | golf, n. a good walk spoiled (Twain)
attr-man.patch
Description: Text Data
|