Re: Viewing Contacts

On Saturday 13 December 2008 11:19:32 Ingo Klöcker wrote:
> On Saturday 13 December 2008, Anne Wilson wrote:
> > On Friday 12 December 2008 18:48:14 Ingo Klöcker wrote:
> > > On Friday 12 December 2008, Anne Wilson wrote:
> > > > The second thing that seems a cause for concern is that when I
> > > > added a record, std.vcf was backed up as std.vcf_5, not std.vcf_1
> > > > as expected.  This is nonsensical, both in the context of its own
> > > > directory and the context of the original resource's directory.
> > >
> > > This behavior does make sense if the backups are created in
> > > ringbuffer fashion, i.e. the oldest backup is replaced by the new
> > > backup and all other backups are left untouched. (Think "Duck,
> > > duck, goose" [1].)
> > >
> > > If the backups were created in "finite queue" fashion then the
> > > creation of a new backup would require renaming all other backups.
> > > (Think belt conveyor.)
> > >
> > > Your observation indicates that KAddressBook uses a ringbuffer for
> > > the backups.
> >
> > Ingo - there were no backups in that directory when I made the
> > change.  Now there is std.vcf and std.vcf_5.  It still seemed
> > illogical to me, so I browsed the directory on an older box, with
> > more backups, and was surprised at what I saw -
> >
> > std.vcf_1		07/07/08
> > std.vcf_2		16/09/08
> > std.vcf_3		30/07/08
> > std.vcf_4		18/09/08
> > std.vcf_5		11/07/08
> > std.vcf_6		25/10/08
> > std.vcf_7		06/07/08
> >
> > I'm pretty sure that older versions ran the conveyor-belt system, so
> > it looks as though the system has changed in recent versions.  This,
> > of course, has implications for anyone needing to restore an older
> > version, which thankfully is a rare occasion these days :-)
>
> Hmm, I found the following very old backups:
> -rw-r--r--  1 ingo users  16557 2006-12-04 19:03 std.vcf_1
> -rw-r--r--  1 ingo users  16557 2006-11-28 23:35 std.vcf_2
> -rw-r--r--  1 ingo users  16557 2006-11-22 00:03 std.vcf_3
> -rw-r--r--  1 ingo users  16557 2006-11-30 00:52 std.vcf_4
> -rw-r--r--  1 ingo users  16557 2006-12-01 01:34 std.vcf_5
> -rw-r--r--  1 ingo users  16557 2006-11-18 09:03 std.vcf_6
> -rw-r--r--  1 ingo users  16557 2006-12-03 23:49 std.vcf_7
>
> So from oldest to newest: 6 - 3 - 2 - 4 - 5 - 7 - 1
> I cannot make a pattern of this.
>
> KAddressBook 3.5.10 appears to use two underline characters for the
> backups. For those backups I get:
> -rw-r--r-- 1 ingo users 18230 2008-12-12 21:21 std.vcf__0
> -rw-r--r-- 1 ingo users 18230 2008-12-12 21:21 std.vcf__1
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__10
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__11
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__12
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__13
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__14
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__15
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__16
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__17
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__18
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__19
> -rw-r--r-- 1 ingo users 18230 2008-12-12 21:21 std.vcf__2
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__20
> -rw-r--r-- 1 ingo users 18230 2008-12-12 21:21 std.vcf__3
> -rw-r--r-- 1 ingo users 18230 2008-12-12 21:21 std.vcf__4
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__5
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__6
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__7
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__8
> -rw-r--r-- 1 ingo users 18174 2008-12-12 21:21 std.vcf__9
>
> So the (modification) time does not give any hint as to which backup is
> the most recent one. OTOH, the identical modification time together
> with the file size (backups 0-4 are larger than the other backups)
> indicates that version 3.5.10 uses the conveyor-belt system.
>
The box I was looking at is a CentOS5 box, my IMAP server.  There, the 
smallest file is

-rwxrwx---  1 anne users 52409 Jan 20  2008 std.vcf__20

with

-rwxrwx---  1 anne users 52931 Oct 25 10:05 std.vcf
-rwxrwx---  1 anne users 52689 Jan 20  2008 std.vcf__0
-rwxrwx---  1 anne users 52931 Jul  7 16:38 std.vcf_1
-rwxrwx---  1 anne users 52689 Jan 20  2008 std.vcf__1

so it looks as though std.vcf_1 is the latest backup on mine, 
with the double-underline backups (up to __20) being older ones.

All very confusing.  I guess the file size is really the only reliable 
indicator.

Anne

_______________________________________________
KDE PIM users mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/kdepim-users

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)

iEYEABECAAYFAklDs50ACgkQbMErw/n0TZo/fQCeKTN1bLTEcam9UVRARv1A4Jzf
B70An3AgD5uwtZPLt44XzLfh90FuFRD/
=Jd2n
-----END PGP SIGNATURE-----