On Sat, 2007-09-01 at 10:54 -0400, Derrell Lipman wrote:
> On 8/31/07, Andrew Bartlett <abartlet@xxxxxxxxx> wrote:
> > I see three options, for the short and long term:
> > - ship as-is (with ability to display users fixed in LDB browser)
> > - ship without any SWAT
> > - ship with SWAT from TP1
> I had a lot of fun creating the qooxdoo framework for the new SWAT. I
> think it provided a fairly easy way for people to add modules towards
> a fully-functional administration tool. The learning curve for
> qooxdoo, although not really difficult, is not tiny either. I'd love
> to see this SWAT continue to be developed, but the last year has shown
> that there are not a lot of developers looking to learn about and
> develop it.
One of the biggest challenges of a project the size of Samba4 is keeping
everything maintained. It is frustrating to see good code rot, just
because it doesn't have someone looking after it regularly.
Particularly when you know how much effort went into creating it in the
> Furthermore, as much as I really like qooxdoo, I think I
> agree with Jerry that a more generic API for Samba, not requiring a
> browser interface, is a better long-term way to go. SWAT, however,
> exists (albeit quite incomplete) and no othere API does. I think I
> recommend leaving it in until there's a good replacement. If nothing
> else, during alpha, some new developers may see it and have a desire
> to improve upon it.
> If SWAT is left in, I still hope to find the time to migrate it to the
> current release of qooxdoo, which will eliminate much of the size
> issue causes in Samba4.
My gut feeling is to leave the code in the first alpha, but not to
advertise it. We will see how things go, and in the less rushed move to
the next alpha, decide if we want to continue with a web management
Authentication Developer, Samba Team http://samba.org
Samba Developer, Red Hat Inc.