[Pkg-samba-maint] Samba3/Samba4: my thoughts after SambaXP conference

[Christian Perrier]

As I more or less promised, let me here summarize my views about
samba3 and samba4 packages, after various discussions I had around at
the SambaXP conference.

All discussions with the Samba Team developers lead mostly to the same
conclusions:

-Samba 4 has no chance to be out in a usable state before the planned
 release date for Etch

-When it will be released (as "alpha", "beta", or whatever), the file
 and print services in it will not be complete (or maybe completely
 missing such as printing services currently)

-When it will be released as alpha or beta, it will be functional as
 an AD domain controller

-in a very far future, and without any visibility, samba4 could fully
replace samba3 (but Andrew Bartlett was really evasive about this for
obvious reasons of non visibility in long term)

So, a first conclusion seems pretty obvious: there is absolutely no
point in trying to have some samba4 code in etch.

That said, the next discussion is more about the "release after
etch". Let's assume for the moment that samba4 will be ready at some
moment during our own release cycle post-etch. That seems pretty
likely at this moment.

From my understanding of the samba users need, one can quite safely
bet that there will be some need for us keeping samba3 in official
Debian releases at that moment....just leaving users time to consider
the code ready for File and Print Services.

In the same time, there will be a strong demand for samba4 in the
etch+1 release so that people can imagine building AD controllers with
it (this is what I feel from discussions with overall samba users,
whether they use Debian or not).

All this leads me (and this seems shared by Noèl, Jelmer Vernooj,
Andrew Bartlett and, though we spoke about this a little bit less,
Jerry Carter) to conclude that, at some moment in the future, we will
need to provide both versions and plan to eventually release with them
both.

We still of course have a lot of time to think about it, as my first
conclusion (also shared by the very same people) is that samba4 can
continue to live its life in experimental as is.

However, for post-etch, this more or less draws the coexistence of
both version in unstable, not only because it will give the software
more exposure (OK, I read threads about that topic and, even though I
don't completely share the strict "unstable is for things we want to
release only", I accept it) but also because we will want to release
with both, quite likely.

I'd like to have concurrent point of views here, especially those from
Steinar, Steve and Eloy, so that we can hopefully explain in a wider
audience (-devel?) what are these plans.


-- 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.alioth.debian.org/pipermail/pkg-samba-maint/attachments/20060427/d7113384/attachment.pgp