SWAT in Samba4

[Gerald (Jerry) Carter]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Andrew,

>> For what it's worth, IMO SWAT is a waste of your effort.  Your
>> energy would be better spent on DC functionality.  Your decision
>> though.
> 
> That's why I'm trying to cut it back, so I don't 
> waste valuable pre-release cycles fixing bugs in it.
> It is the only component that consistently had bugs in it
> before most of the TP releases.  It is also the only
> component without any testsuite.  I don't see this as a
> coincidence. 
> 
> The other option is to, as you suggest, remove it completely.  
> 
> I do very much appreciate your views on this. 

In some ways, it would be better to leave it broken in the
alpha releases so that people who are testing realize that
is it *really* broken.  Things that cause the most pain
tend to get more attention.  I'd leave it broken for a while.
If it becomes clear that no one is stepping up to the plate
to fix it, it can be removed later on.

While we're on the topic, I'd really reconsider your views
on not doing a Samba4DC.  The recent posts have testified to
the confusion between the SAMBA_4_0 and Samba 3 feature
set.  Unless you get more developers working on Samba 4,
a complete replacement of Samba 3's functionality is really
far off which is what you would need to do a Samba 4.0.0
release.

If it were me in your position, I would commit to shipping
the features that I would support.  If no one maintains, SWAT
don't ship SWAT.  If no one maintains printing, don't worry
about printing.  What *you* care about is the DC functionality.
So work on that.

This types of cases really lend themselves to a DVCS solution
and local branches.  Your tree would be the DC tree.  Someone
else could maintain SWAT, etc...  If at some point all the trees
added up to a complete 4.0 release, them pool them together
into a master tree and release.

But right now, you have to under-promise and over-deliver.  And
what you can deliver on is an AD DC.  And that is what people
really are chomping at the bit for.  Let Samba 3.x give you cover
on the stuff you don't really care about.  We are all working
towards the technology proposed in SAMBA_4_0 in other ways.  But
it's the DC work that has the most freedom at this point since
there is nothing to compare it against.   You don't have to do
everything.  Right now, just do what nobody else does.

Feel free to file this advice in whatever file system it makes
most sense.  Just my thoughts.  I'd hate for you to waste your
time on busy work or redoing what's already been done when you
have the potential to accomplish so much more.

Remember that all code really throwaway code at some point.
It's the knowledge that we carry forward to improve the next
generation that really counts.  So don't get caught up in
"perfect" releases or "final" versions as they are really just
a fantasy for obsessed release managers :-)





cheers, jerry
=====================================================================
Samba                                    ------- http://www.samba.org
Centeris                         -----------  http://www.centeris.com
"What man is a man who does not make the world better?"      --Balian
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFG5qJnIR7qMdg1EfYRApm8AJ9oqb+jg1ZoA89AKq3oiLb/szu5SQCgjf6H
zWBxXCFPnPb9KqSvhyV4Yyo=
=4u6w
-----END PGP SIGNATURE-----