Joerg Schilling writes:
> James Carlson <james.d.carlson at Sun.COM> wrote:
> > I think you mean to say that you believe that the ARC members are
> > either deliberately or ignorantly _precluding_ the option of replacing
> > pax and other things by star in the future.
> >
> > This is not true.  We do understand the nature of this
> > incompatibility.  We're not fools.  And there is no attempt or intent
> > to preclude _any_ future project.
> 
> Why then do you try to include new incompatible options that prevent to 
> implement the option of replacing the programs?

I didn't try to include anything.  I don't know why you're confusing
me with the project team or with the ARC itself.

Nor do I believe that the project team proposing this change is
attempting to prevent any such thing.

This project (like many that have preceded it) forces a requirement
onto such a future replacement project.  It does _not_ prevent such a
project.

> > The nature of the incompatibility is that, when invoked as
> > /usr/bin/tar or /usr/bin/pax, the 'star' implementation will be forced
> > to adopt the command line behavior of those older tools, the file
> > formats, and the documented behaviors.  That requirement for backward
> > compatibility extends even to the options and features that the author
> > of star dislikes, such as the older '-@' and the new '-/' option.
> 
> This is not true:

Yes, it is true, because this is the nature of backward
compatibility.  It's one of the most important things we (in the ARC)
enforce.

You won't get to pass "go" without it.

> Even though the way of using -@ is questionable, the main problem with the 
> implementation of the functionality is that it is based on deprecated methods
> from POSIX.1-1988 instead of being based on portable and vendor tagged 
> POSIX.1-2001 methods.

In evaluating compatible change in OpenSolaris, that doesn't actually
matter.

You may be right that it's poorly designed.  You may be right that a
suitable "expert" would have done it differently.  However, none of
that matters to a user who is trying to extract an archive created by
an older system using this existing format, or create an archive to
transfer to such a system.  What matters in those cases is
compatibility.

> philosophy. The fact that Sun did not update this preliminary implementation 
> by a fully POSIX.1-2001 based extended tar format did throw Sun's tar back to 
> the stoneage. All additions that have been implemented in sun tar lateron,

That may be true.  It's also irrelevant.

> > When invoked as 'star', the utility may do as it likes.  It needn't
> > support anything the other existing utilities do.  But when invoked as
> > /usr/bin/tar, that's when backward compatibility becomes important.
> 
> This was implemented, but this kind of compatibility will be destroyed by 
> accepting the current ARC case.

Nonsense.  It forces _more_ work onto such a future project, but it
does _not_ prevent it.  You've failed to explain how this case
actually _prevents_ anything from occurring in the future.

> Do you know the difference between the term "star" and the term "tar"?

No, of course not.  Like most of the people on this list, I live to be
insulted.

-- 
James Carlson, Solaris Networking              <james.d.carlson at sun.com>
Sun Microsystems / 1 Network Drive         71.232W   Vox +1 781 442 2084
MS UBUR02-212 / Burlington MA 01803-2757   42.496N   Fax +1 781 442 1677

Reply via email to