Mike Kupfer wrote:
> >>>>> "Roland" == Roland Mainz <roland.mainz at nrubsig.org> writes:
> 
> Roland> - Would Sun like/dislike the diversion between OpenSolaris
> Roland>   vs. Solaris in this point ? AFAIK SchillyX and other
> Roland>   distributions did already expore this part (e.g. ksh93
> Roland>   residing at /usr/bin/ksh) quite successfully
> 
> Currently, the ON part of OpenSolaris is mechanically generated from
> Solaris.  At some point this will be flipped around, so that OpenSolaris
> contains the "master" workspace. 

Offtopic: Will this include the full history of changes in Solaris, too
? It would be nice to see which patches/changes were applied over the
time and for which bugid/RFE...

> Either way, any divergence has to be
> coded into the makefiles, using a switch to determine whether ksh means
> ksh88 or ksh93.
> 
> I don't think anyone in Sun would be happy with such a divergence.  I'll
> ask if folks would accept it, temporarily,

That would be nice... it may help to figure out which stuff actually
breaks (I do not expect much breage if we do testing with the
genunix.org prototype first, see below).

> but I'd also like to explore
> alternatives.

Agreed.

> One alternative would be to manually change /usr/bin/ksh to mean ksh93.
> You'd have to redo this any time you installed new Solaris Express bits.
> But building and BFUing OpenSolaris would leave /usr/bin/ksh alone.

Yes, but this limits testing a lot... we will not receive much feedback
whether ksh93 causes any scripts and/or applications to break... ;-/

> Roland> Cyril Plisko <cyril.plisko at gmail.com> from genunix.org said we
> Roland> could have a branch in their SVN (=Subversion) respository -
> Roland> which may be ideal for getting a prototype running.
> 
> We should also have Subversion support on opensolaris.org "soon".  But
> if the team wants to use Subversion, I guess using genunix.org would
> help things get started more quickly.

The idea of using a branch at genunix.org was to create a prototype for
development, testing and evaluation.

> The alternative would be for April to deliver snapshots as tarballs;
> external contributions would come back as source patches.

>From my experience that isn't always a good idea since this may cause
nasty fun with getting patches applied which were created for older
versions... such a kind of manual patch integration can consume horrible
amounts of time... ;-(
I'd like to avoid that April has all the work with patch integration&co
(which can be very annoying) and that's why I was running around and
asking for SVN access...

> Of course, Subversion would be easier for collaborative development than
> tarballs and source patches.

That was the original idea...

> Roland, IIRC, you mentioned some interest in helping with development
> work, but you also have concerns about whether Sun would integrate ksh93
> and then say there's nothing else to do.  I know *I* can't make any
> guarantees about what happens after ksh93 is added to the source base,
> and I'm not sure what guarantees April can make, if any.

Yes, I know.... but for now I would be glad to get some actual work done
and get ksh93+libshell.so integrated into OpenSolaris - and then decide
the next steps...

> What I'm trying get at is that if there are external folks who are
> likely to contribute code, then it makes sense to consider using
> genunix.org for the development work.  But if April is going to do all
> the development work, it might be simpler to use the tarball approach.

I'd like to do development and others like Felix have expressed
interest, too. And it may be nice to have SVN including commit logs
(send as emails through a mailinglist (maybe this one) so we can track
the progress).

----

Bye,
Roland

-- 
  __ .  . __
 (o.\ \/ /.o) roland.mainz at nrubsig.org
  \__\/\/__/  MPEG specialist, C&&JAVA&&Sun&&Unix programmer
  /O /==\ O\  TEL +49 641 7950090
 (;O/ \/ \O;)

Reply via email to