Subject: do we really have to "break" 'tar -o' *again*?!?!
To: NetBSD-current Discussion List <firstname.lastname@example.org>
From: Greg A. Woods <email@example.com>
Date: 02/03/1999 02:19:37
[ On Tuesday, February 2, 1999 at 15:31:52 (-0800), Todd Vierling wrote: ]
> Subject: CVS commit: src
> Also change meaning of pax-as-tar option "-o" to mean
> "use V7 output format" (same as GNU tar's -o).
>From The Single UNIX Specification, Version 2 (describing '-o'):
Assign to extracted files the user and group identifier of
the user running the program rather than those on the
GNU Tar has *always* been the lone one out, and *wrong*, on this.
I don't care how long NetBSD's /usr/bin/tar is/was GNU Tar, I think it's
a bad idea to avoid being backwards compatible with Real UNIX, and
gratuitously incompatible with the industry standards. It's even worse
to be gratuitously incompatible with other pax-as-tar systems.
(This was one of my complaints from the very first day I first used GNU
Tar, and I think one of the very first issues I had with NetBSD, tar
being one of those programs you often encounter first on a new system.)
If we really must have a way to produce V7-tar format archives with
pax-as-tar then I'd suggest '-O', which so far as I can tell has no
other meaning (with '-c', that is) in any other well known version of
tar, not even the one on SCO UNIX. Note that Standard UNIX doesn't
provide a way to create V7-tar archives any more, nor should they be
necessary, esp. given that even the original Pax could have been
compiled and used on V7 or any equivalently ancient systems.
Greg A. Woods
+1 416 218-0098 VE3TCP <firstname.lastname@example.org> <robohack!woods>
Planix, Inc. <email@example.com>; Secrets of the Weird <firstname.lastname@example.org>