Hi guys,
even if I'm not ActiveMQ PMC, let me express my thoughts:
1/ in order to avoid to disturb the users, I would prefer to avoid to
have ActiveMQ 7 or whatever for HornetQ. IMHO, the branding provides
information to the user, and people may be "lost" if we "rename" HornetQ
as ActiveMQ x.
2/ even if it's not encouraged, it's possible to have subprojects in an
Apache project. For instance, it's the case in ServiceMix (with bundles,
and in the past with nmr, etc), and in Karaf (cellar, cave, decanter).
So I don't see problem to have Apache ActiveMQ Hornet and Apache
ActiveMQ (and Apache ActiveMQ Appollo which would make sense).
3/ the HornetQ donation requires all the license granting and donation
rule: just to be sure that all HornetQ contributors/committers are
really aware of the Apache donation.
Just my $0.02
Regards
JB
On 03/25/2015 03:58 PM, Chris Mattmann wrote:
Thanks Hadrian that helps to clarify things.
Large code donations to the ASF need to start with an
IP clearance short form especially if my read is right
below and this large code donation was not entirely authored
by the ActiveMQ PMC and those with ICLAs on file. Has this
been done? FYI:
http://incubator.apache.org/ip-clearance/ip-clearance-template.html
(note this links to the Incubator and yes I know that
this wasn’t done through the Incubator but the form is
still valid)
Furthermore, have their been discussions about those
members of the community that were not represented on the
ActiveMQ PMC? The PMC is the one that releases Apache code,
and so I’m wondering why the Apache ActiveMQ PMC brought in
a large code contribution consisting of authorship by people
that weren’t on the Apache ActiveMQ PMC - not fully out of
precedent but I would have expected to see an influx of those
members of HornetQ community on the ActiveMQ PMC if the
code base is changing direction and the community surrounding
it is as well. Also if HornetQ is a Redhat product or originally
was a Redhat product, we need paperwork on file such as a Software
Grant Agreement (SGA) that helps to cleanly bring large code
elements into the community:
https://www.apache.org/licenses/software-grant.txt
Note the above advice are typically things that would have
happened had this contribution come through the Incubator.
Finally - Apache doesn’t really do “sub projects” anymore.
It’s been a long time. The clarification on that is that a project
can have multiple “products” (aka Lucene which releases Solr,
Lucene, PyLucene, etc.) but these cannot be distinct projects
and communities. If they are, they need to be made as such
by an act of the Board (direct to TLP; spinning out), and/or
by going through the Incubator.
Cheers,
Chris
-----Original Message-----
From: Hadrian Zbarcea <hzbar...@gmail.com>
Reply-To: <dev@activemq.apache.org>
Date: Wednesday, March 25, 2015 at 4:45 AM
To: <dev@activemq.apache.org>
Subject: Re: [DISCUSS} HornetQ & ActiveMQ's next generation
Hi Chris,
There was a code donation that completed last year. It started on
07/08/2014 (in a thread named: Possible HornetQ donation to ActiveMQ)
and completed in Oct.
HornetQ was a long time project and community of RedHat. The idea, the
way I understood it at the time, was to take relevant parts better
implemented in HornetQ and rewrite parts of ActiveMQ that were showing
their age (Hiram pointed out a few in the other thread yesterday).
The HornetQ community opted to have the ActiveMQ pmc instead of the
incubator as the sponsoring entity. There are many RH people on the
ActiveMQ pmc, the technology space is the same (messaging), it probably
was considered a better fit and and easier way to build a community.
The HornetQ subproject opted to use the ActiveMQ6 name as the name of
the project. However, the subproject is kept independent and there are
efforts being made to align some of the features with the current
ActiveMQ (ver 5.x). I believe the expectation is that users will migrate
to hornetq eventually, based on superior technical merits. That is a
migration, not an upgrade, with minimal chances of going back. The
ActiveMQ6 name is probably intended to help with that and create the
perception that it is the same project.
Only a very small part of the current ActiveMQ community is actively
involved in HornetQ. There are concerns expressed by a few PMC and ASF
members that the activemq6 name creates an confusion. Hornetq is not yet
a stable community.
The proposal is to change the name for the HornetQ to something that
reflects the current status, and not activemq6. It it relevant to note,
that with hornetq being named activemq6, the current activemq project
has no possibility of having a major version upgrade. It was also noted
by community members (non-committers) as well (see Lionel Cons' email)
that there is a precedent that didn't succeed as anticipated to name
another ActiveMQ subproject (apollo) as activemq 6. The name is now
reused for HornetQ.
One analogy would be Microsoft for instance donating IIS to the ASF as a
httpd subproject and name it httpd3, because the current httpd is old
and has no future.
Chris, your thoughts on the issue are highly appreciated. This does not
provide the complete picture, but it's hopefully clear enough.
Hadrian
On 03/25/2015 10:07 AM, Chris Mattmann wrote:
Can someone please explain what is being discussed?
I’m sorry I don’t follow the subtleties here.
Is there a code donation being proposed to Apache
ActiveMQ?
Cheers,
Chris
-----Original Message-----
From: Rob Davies <rajdav...@gmail.com>
Reply-To: <dev@activemq.apache.org>
Date: Wednesday, March 25, 2015 at 1:47 AM
To: <dev@activemq.apache.org>
Subject: [DISCUSS} HornetQ & ActiveMQ's next generation
(was: HornetQ &
ActiveMQ's next generation)
Thanks Lionel - I agree.
The [VOTE] thread was getting a little verbose, and a little heated.
There were a lot of opinions, and a lot of assumptions and its likely
there was some miscommunication when HornetQ was donated to the
ActiveMQ
community.
On the plus side, its great that there are so many passionate members
of
the community.
It seems there is no consensus from the ActiveMQ community that HornetQ
should be the next generation of ActiveMQ - yet - and hence should be a
sub-project with its own name.
Personally, I believe there are a lot of advantages of starting
development of ActiveMQ 6 around a HornetQ core - but as Hadrian as
already pointed out - it does need to validate itself by growing its
own
diverse community first. I hope the ActiveMQ community as a whole gets
involved in the code donated from HornetQ and pushes it the right way.
Rob
Lionel Cons <mailto:lionel.c...@cern.ch>
25 March 2015
06:58
(for the sake of clarity, I
think that this important subject deserves more
than the [VOTE]
thread currently used, hence this new thread...)
Apollo (tagline =
"ActiveMQ's next generation of messaging") started in 2010
as an
ActiveMQ sub-project in the hope of becoming ActiveMQ 6. At that time,
the
latest ActiveMQ was 5.4.
Almost 5 years later, ActiveMQ is now
5.11 and some of the Apollo developments
(like LevelDB or MQTT) have
been merged into ActiveMQ 5.x. FWIW, Apollo is
still officially
advertised as "the core of the 6.0 broker" in
http://activemq.apache.org/new-features-in-60.html.
In
parallel, last year, the HornetQ codebase has been donated to ActiveMQ.
The
ActiveMQ 6 RC assembled so far is HornetQ with Apollo's tagline,
"ActiveMQ's
next generation of messaging", hence the confusion.
For
me, the fundamental question to answer is: has it been _decided_ that
HornetQ
will be the core of the next generation of ActiveMQ?
If the
answer is yes then HornetQ can be called ActiveMQ 6.0 and we should get
a
stable, feature complete ActiveMQ 5.x replacement a few minor versions
later
(who trusts a .0 version anyway?).
If the answer is no
(or not yet) then HornetQ should probably appear as an
ActiveMQ
sub-project, just like Apollo (still) is. HornetQ can evolve there
and
come closer to ActiveMQ "the next generation". Then, the ActiveMQ
project
should decide what will be ActiveMQ 6.
Cheers,
Lionel
Cons
--
Jean-Baptiste Onofré
jbono...@apache.org
http://blog.nanthrax.net
Talend - http://www.talend.com
