Bug#174308: star should become standard tar
On Fri, Dec 27, 2002 at 02:47:14AM +0100, Alex de Landgraaf wrote:
> Quoting "Bryan W. Headley" <firstname.lastname@example.org>:
> > >>Also, Debian has always taken advantage of the rich set of features
> > >>offered by the GNU tools. I see no benefit to limiting ourselves to
> > >>the use of POSIX features. I certainly see no a priori reason to
> > >>adopt it as the standard to write to. POSIX itself is non-free, and
> > >>many of its standardization choices are motivated by compromises for
> > >>the sake of proprietary unix systems. I think such systems should be
> > >>supplanted, not catered to.
> > >
> > >
> > > So when is ACL-support for gtar going to be available? I'm a little
> > > miffed to learn that there are other free tar implementations out there
> > > that handle ACLs, while Debian and GNU tar are lagging behind.
> > >
> > Why not ask one of the gtar maintainers? Here's one,
> > Paul Eggert <email@example.com>
> > And they have an anon email address,
> > firstname.lastname@example.org
> Hmmm, is it me, or hasn't there been a GNU tar release this millenium?
> on ftp://ftp.gnu.org/gnu/tar, the last release was somewhere in '99.
> Now i agree on the "if it works, dont fix it" view carried out here by many,
> and Debian should be conservative with making changes that dramatically
> change the distribution, but do we want to depend on software that isn't being
> updated to new changes? otoh I don't see why ACL couldn't be added to GNU tar,
> it has a --posix parameter that seems like it could be used for it, it just
> needs to be reimplemented for the current POSIX 'standard' (isn't a standard
> that what a large group of people use? oh well, nm :o)
> My opinion: if people are still working on GNU tar, try to find out why ACL
> isn't being implemented. If GNU tar really is dead, then it's no use to keep
> defending it, changing to star (or another clone) should seriously be
> considered, once it is compatible.
I'm pretty sure that bug-gnu-utils is publically archived. Its
archives might be enlightening. The verdict was that Paul (I believe
it was him...) did not see any sufficiently widely-accepted set of
interfaces for ACLs to begin justifying the tar support. I also
believe he invited the person who asked to do certain things to move
the process along, and nothing ever came of it. It was within the past
month or so.
Amazing what actually asking can find out. GNU tar is by no means
stagnant, and Paul has a reputation as a bit of a POSIX/portability nut
himself. No offense intended.
MontaVista Software Debian GNU/Linux Developer