I don't think we're debating if, we're debating when. I get your point
and I would agree if this was any different than before. The reality
though is that *every* single minor release of Camel was *not* 100%
backwards compatible. We absolutely strive for backward compatibility
and 2.9.0 should not be worse. We should be 100% compatible in patch
releases though.
The reality is that when we'll start working on 3.x we won't have it for
at least 9 months, most likely. Nobody will migrate to a 3.0-Mx
milestone release. It's also likely that many won't migrate to a new 2.x
while we actively work on 3.x. In the patch releases we only have a
small subset of the new features on trunk.
I know some are kinda tired of 2.x and want to look at and talk about
the next great thing. We had this discussion a year ago and promises
were made that 3.0 will be out in Q1 2011. We don't yet know how 3.0
will look like, although many have their ideas. We know that we want a
cleaner api, we know we want it leaner and we know that we want shorter
build times and other few things. We can easily continue to work on that
and the other features that come with every release until we figure out
how 3.0 will look like. Doing it now is, imho, a distraction.
My $0.02,
Hadrian
On 09/24/2011 08:18 AM, Hiram Chirino wrote:
The fact that it's not 100% compatible IS why it's not a good idea to
call it 2.x. The point of version numbering schemes are to reflect
the compatibility between releases. If trunk is not a major release,
I don't know what is. And yes I would hope that the 3.x line tries to
keep as much backward compatible support with 2.x as possible.
Otherwise, you've got a brand new product in your hands and your 2.x
users will have a big wall to climb to to upgrade to the 3.x. As for
dropping deprecated APIs, I'd do that in a 3.1 or once we verify that
migration to 3.x is fully backward compatible from 2.x (I'm sure we
will miss some compatibility stuff in the first few releases of 3.x).
Regards,
Hiram
FuseSource
Web: http://fusesource.com/
On Fri, Sep 23, 2011 at 11:12 AM, Christian Schneider
<ch...@die-schneider.net> wrote:
I just talked to Dan and Hadrian about this idea. The problem is that
however we would call such a release people would take it as kind of a
milestone and it would probably not be used much.
So it is probably better to use the 2.9 and perhaps 2.10 release for that.
The 2.x naming will suggest that it is mostly compatible and we can mention
in the release not that it is not 100% compatbile.
For people who do not want the changes we can use the 2.8.x branch. So we
could keep porting back some improvements like Dan did for people who want
to stay 100% compatible.
Christian
Am 23.09.2011 16:43, schrieb Christian Schneider:
Basically that is a good idea. I also thought about a similar solution. I
would only not like to call it milestone as that typically implies that it
is instable. No sane admin will deploy a milestone release in production.
How about 3.0.0-compat or similar. So we could have one or more releases
of this that are fully production ready but will still contain the
@Deprecated classes. Then we can release Camel 3.0.0 that simply removes
these.
Christian
Am 23.09.2011 15:59, schrieb Willem Jiang:
I think the main concern of Christian is he afraid Camel 3.0 won't have
much user to use, and he explain us the recent back ports of Camel 2.8.2 is
because of the API change in Camel 2.9-SNAPSHOT.
How about we treat the Camel 2.9 as the Camel 3.0-m1. When we developed
Camel 2.0, we actual did three mile stone release to fill the gap of the API
change.
For the user who want to do some release just after Camel 3.0 RC is out,
he can keep using Camel 3.0 mile stone release.
For the user who just want use the stable Camel 2.x, he can chose Camel
2.8.x, as we already did a lot of merge of on the Camel 2.8.x branch.
Camel 3.0 Mx can tell the user what API change will be made, and we can
removed the @Deprecates API when the Camel 3.0 is shaped.
Any thoughts ?
--
--
Christian Schneider
http://www.liquid-reality.de
Open Source Architect
Talend Application Integration Division http://www.talend.com
