[Release Planning 3.6] Samba 3.6.0pre2

[Andrew Bartlett]

On Thu, 2010-11-18 at 23:38 +0100, Guenther Deschner wrote:
> Hi Karolin,
> 
> On 11/17/2010 10:34 PM, Stefan (metze) Metzmacher wrote:
> > Am 16.11.2010 17:15, schrieb Jeremy Allison:
> >> On Tue, Nov 16, 2010 at 04:02:38PM +0100, Karolin Seeger wrote:
> >>> Hey folks,
> >>>
> >>> are we ready for Samba 3.6.0pre2 or are there any important issues which
> >>> need to be addressed first?
> >>
> >> I think Guenther has something *VERY IMPORTANT* he needs to say
> >> on this matter :-). Simo too, probably :-).
> >
> > We need to fix the rpc client code, to avoid always using a read request
> > after the trans request.
> > I have a fix for that, but I need to clean it up a bit.
> > And it's also hard to backport it to 3.6 (from master) as some of Simo's
> > patches
> > are not backported yet.
> 
> Yeah, there is a whole bunch of changes still missing.
> 
> Of course - as usual - my fault :-)
> 
> I took several approaches to cherry pick all those while still at the 
> SDC, succeeded with some but failed miserably with most of them. As 
> metze pointed out: there are some subtile dependencies on other larger 
> changesets (auth merges, new rpc server) that are not merged yet. And 
> then I was notoriously bad on merging back the spoolss fixes after 
> andreas and simos rewrite...
> 
> To summarize: I have a not so good feeling for v3-6-test at least when 
> it comes to printing and spoolss... Alternatives now are:
> 
> a) come up with painfully hand-merged patches (where we never can be 
> sure of where we are exactly and if we got all important bits, also 
> further cherry-picking stays painful) or
> b) copy over master (which has cooled down a bit recently) for the next 
> pre-release (which would be a slap in the face to those who carefully 
> picked and merged to 3-6-test so far (I am certainly not amongst these)).
> 
> As I have the impression that at least master is way better tested and 
> at least used sporadically by developers (unlike 3.6), I strongly vote 
> for option b). But its not up to me to make that decision and then again 
> I can fully understand anyone voting against.

+1 for b)

I think re-branching from master is a good option, particularly now that
master is reinforced by autobuild.  

Also, I've seen how the auth changes I've made, and then the partial
merge of some of the NTLMSSP changes has caused quite considerable
confusion, and while I'm sure it was done with the best of care, I also
can't certify that the result is correct.  I made the decision not to
merge the auth changes into 3.6 on the misunderstanding that 3.6 was
shortly to be stabilised and released, but it has seen quite a deal of
development instead. 

Andrew Bartlett

-- 
Andrew Bartlett                                http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org
Samba Developer, Cisco Inc.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 190 bytes
Desc: This is a digitally signed message part
URL: <http://lists.samba.org/pipermail/samba-technical/attachments/20101119/a4eda492/attachment.pgp>