Don Cragun wrote:
> After a lengthy discussion, this case was approved during today's PSARC
> meeting. J?rg's issue (the star utility uses -/ to mean something
> different from the way cpio, pax, and tar will use -/) was discussed in
> detail, but PSARC did not find his arguments compelling.
>
> - Don
>
I see that I missed this meeting and discussion. Its approved.
Without commenting about the decision, I'd like throw out a
few observations even though its already quite obvious that too
many observations were made. :-)
1) Any references about a star option "introduced in 1994" is
irrelevant. In 1994, star was just one of hundreds (thousands)
of FOSS programs. Sun/Solaris/PSARC can't track or
mediate random FOSS programs.
2) On June 16th 2004 when PSARC 2004/480 was approved, star
became relevant. This is only an approval for "sfw" integration.
This doesn't make it a hard rule that star/tar should share option
flags, but the project team should investigate this, expose the
issue (if there is one) to PSARC and decide their proposed path.
PSARC can "yea" or "nea" this.
3) I believe several people have started to work with Joerg on an
additional fast-track (for star) with a dependency on 2004/480.
(Hence, only a review of "goodies" Joerg has added since that
time.) I know Basabi spent some time with this.
Since Basabi was busy, I tried to help. I was unable to get Jeorg
to produce a proposal I was comfortable with. This was mostly
about "getting the words right", rather than any serious contentions
about the content. I think Joerg might have assumed I would
write it, rather than edit it - not sure.
Unfortunately, I was struck by an unexpected health problem
(on June 2, to be exact). I was completely away from work for
about 3 weeks and then 3 or 4 weeks at part time. Let's just say
that I haven't worked on this. (Then again, Joerg was also quiet
about this also.)
Let's assume that somebody gets this fast-track done and approved.
I think it shouldn't be hard.
I *think* Joerg's position was that star should be /usr/bin/star (as
opposed to /usr/sfw/bin/star), but *not* a replacement for tar.
There was supposed to be a "statement of intent" that star would
eventually replace tar.
Had this happened, then it is clear that conflicts between tar and
tar would have been addressed.
4) The fact that 2004/480 wasn't integrated is also (mostly)
irrelevant. The architecture was approved. If some other
project team finds that "it was in the way", then some PSARC
approved remedy would result. (We need not discuss how
this would be done.)
5) The fact that 2004/480 specified /usr/sfw/bin/star is relevant.
That is a reflection of importance of "flag compatibility" (or
non-importance).
- jek3
