Invited to join efforts with the Codehaus Mojo plugin - thoughts?

[Charlie Collins]

I was invited by the Apache/Codehaus GWT plugin creator/maintainer,
Nicolas De Loof, to join efforts with him and combine our projects.

That project is here: http://mojo.codehaus.org/gwt-maven-plugin/.

The new project, the joining of the two, would be under the Mojo/
Codehaus/Apache rules, and not here at Google Code hosting (not nearly
as much autonomy).

I am inclined to say sure and join the projects, after all they do
have the same goals (not Maven "goals" but overall "goals," you known
what I mean ;)). Additionally the Maven standard way to include a
plugin seems to be at the Codehaus Mojo (I don't know all the rules,
but that does seem to be the path into real mavenhood for a plugin).
In fact I suggested that we join efforts several months back when I
first noticed the Codehaus Mojo plugin (but got no response at that
time). I am not in business here to try to be "the" plugin or whatnot,
but just to provide a good plugin that we all can use, no matter where
it lives (when I created this project, years ago now, it was the only
such offering).

Even though I am *inclined* to join, I am not sure if it will actually
happen. There are several potential issues with the process. First, I
am not the only person here at this project, so I can't really just
decide for all of us. What do other project contributors here think?
Second, I have to be voted in to join the Mojo project through the
Apache foundation - not sure how that works (and I can't just add
others, one of the reasons I put the project here to begin with, I
don't prefer the oligarchy, I prefer just letting anyone that wants to
help in). Third, I am not sure how well the projects will mesh - I
think GWT-Maven is further along, but I am not really familiar enough
with the Mojo plugin to know how well it works, how robust it is, etc
(it seems simple at the outset, but as anyone who has worked on things
here knows there are many subtleties to actually getting this plugin
to work in the wide variety of use cases people throw at it).

Even if Nicolos and I do ultimately decide to join, there is no time
frame yet, no hurry (no changes to this project are planned, it will
still be maintained, etc). I would like to get feedback from the
community here before anything happens. So, let me know what you
think, please.

[jieryn]

I think a unification would be in the best interests of the user base;
there was already a lot of confusion as to which plugin should be used
and configured, and the recent deployment of codehaus plugin to
central did not help.

I reviewed the codehaus plugin about 6 months ago, and the 1.0.0
release doesn't seem to have improved things. The codehaus plugin
seems to be significantly far behind the google plugin in: features,
maturity, actual use/deployment.

I don't see any aspects to the codehaus plugin that would be
beneficial to the google plugin. So, if a unification of the plugins
would simply be the replacement of codehaus plugin in whole with the
google plugin, then I wholeheartedly vote yes. I think it would be
advantageous for the google plugin to fall under codehaus because, as
you say, it is the path to true maven pluginhood.

Perhaps the codehaus plugin team could serve as a cleansing of the
google plugin codebase, putting things into the formality expected of
Maven plugins, etc. That would certainly be an advantage; especially,
and I mean no offense Charlie (and team) as I am quite grateful to
your hard work, there seems to be some peculiarities and uncertainties
about how to do things the Maven way, at times.

[Darren Hartford]

I'm not a developer on this project, but a long time user of maven and the maven gwt plugin.  From what I've seen of this group, and Charlie in particular, they have been very proactive (see Charlie's comment about trying to avoid the 'gwt' prefix collision). I, from a trust/emotional standpoint, would prefer to keep this gwt-maven project hosted as-is with google because many people are indeed already using the gwt-maven project, Charlie and team have built up trust in the project (very important), and large amount of feedback has gone back into this project.  To up and move it would cause a lot of potential issues that most people would like to avoid. It is very frustrating that the 'gwt' prefix collision has happened, attempts to avoid it were made, and there is no way for Charlie and team to go but to sidestep the issue by renaming the gwt-maven project's prefix because codehaus decided they had the right to the prefix (which they may based on 'tier 1'-style maven plugin support, not saying one way or the other). my two coppers, -D --- On Tue, 11/25/08, jieryn <jie...@gmail.com> wrote:

[Arthur Kalmenson]

I have to agree with jieryn, I think that unifying the plugin would be
in the best interest of the user base. It'll make sure that there is a
good official Maven plugin and one that has a larger team behind it.
As it stands, both plugins have their own useful features, so I'd like
to see them combined.

Charlie, you and the team have done an excellent job and I love the
plugin, but you can definitely use more help. The more people there
are behind a project the better.

--
Arthur Kalmenson

[Charlie Collins]

I was kind of thinking along the exact same lines as Arthur there. I
don't want to be the only main guy on this thing, having more people
involved would be beneficial for many reasons. Don't get me wrong,
others have helped here too, it's not just me - but I have definitely
been the *main* creator/developer/maintainer - others seem to come and
go (which is understandable, this project is something like 2-3 years
old, who knows why they heck I even do it ;)).

I think it would be in the best interest of everyone, in the long run,
if the projects join and as many dedicated people as we can get are
involved. Also, if this were a different type of project I might
resist this a lot more. I really like the autonomy (and the tools) of
a Google Code project, but the bottom line is this IS a *Maven*
plugin, doing things more in the Maven standard path only makes sense
(even if the Maven standard path is not necessarily something I am
very familiar with, or entirely comfortable with).

That said I also agree with Darren about the trust and the
"frustration" that moving this project may cause. I hope, if and when
it comes to it, that I can mitigate that by being very involved on the
Apache/Codehaus side, and by not removing or just shutting down this
project, but rather keeping it going until everyone is comfortable
with the new project (until we all agree the new one runs fine, meets
all the requirements, handles the use cases, etc).

[Mirko Nasato]

My point of view is that since the Codehaus plugin has been deployed
to the central repository first, that plugin now becomes *the* gwt
maven plugin. Despite the fact that this project is actually more
advanced.

So it would be beneficial to join the Codehaus project that has a
stronger "brand" (so to speak) with this one that has more features.

But with this project being LGPL and the Codehaus one using the Apache
license it's easy to foresee licensing issues in copying code
directly.

Cheers

Mirko


2008/11/25 Charlie Collins <charlie...@gmail.com>:

[jie...@gmail.com]

IANAL BUUUUTTTT... Isn't this mitigated by Charlie, and the other
developers, owning the copyright to the code? They could simply change
the license at their whim. This is exactly how projects like MySQL can
offer dual licensed stuff, etc.

[eldzi]

I vote "yes" if the status, developer rights and community mood will
not be disturbed by the merge and be fluently transformed to one
project. Personally I like the way Charley leads this project and how
everybody is willing to help to shift it to be better. Codehaus plugin
has nice features (async interface generation, eclipse launchers) so
by merging this two projects everybody will benefit. This gwt-maven
plugin seems to be more mature and used but that is for further code
examination.

From license point of view if we state it is under ASF and then merge
with Codehaus there shouldn't be a problem.

Finally it will be good to see what will change by joining Codehaus
from developer's view (the voice that matters ;)

On 26. Nov, 04:13 h., jie...@gmail.com wrote:
> IANAL BUUUUTTTT...   Isn't this mitigated by Charlie, and the other
> developers, owning the copyright to the code? They could simply change
> the license at their whim. This is exactly how projects like MySQL can
> offer dual licensed stuff, etc.
>

[Mirko Nasato]

2008/11/26 eldzi <eld...@gmail.com>:

>
> From license point of view if we state it is under ASF and then merge
> with Codehaus there shouldn't be a problem.
>

Yes, but you may need some sort of agreement signed by all the people
who contributed some code, like this one

http://www.apache.org/licenses/icla.txt

Although for a Codehaus project maybe that's not required.

Cheers

Mirko

[Mirko Nasato]

Is it already official then?

http://jira.codehaus.org/browse/MGWT-3

"googlecode gwt-maven and mojo gwt-maven-plugin are merging to Mojo

You will get ASAP your favorite goals available from Mojo plugin"

Cheers

Mirko

[Charlie Collins]

It's underway now yes. The Apache/Codehaus gang voted me in, and
Nicolas is working on merging what we have with his as a first step,
we are not changing anything here.

I will keep maintaining this plugin as we go, until I feel like the
Codehaus one is up to snuff, and then I will put it to you guys to try
it, test it, etc. AFTER everyone thinks the Codehaus one is good to
go, then I will likely suspend this project and direct everyone to the
codehaus plugin. In my mind it may take a month or more to get to that
point (the point where the other is set, and this one can be
deprecated), so this project is still the priority for now. (Nicolas
says "ASAP" to get our goals over into his plugin, because he is
handling that, but I won't throw the switch here until I can sign off
on the new plugin having all we want, and until I understand all the
Codehaus politics and such - how to do a release - etc.)

For now everything *here* stays the same. I am going to try to get a
new release in the next few days that changes the prefix.

[Mirko Nasato]

That's excellent news!

(And in the meantime since we're not affected by the prefix clash here
we'll just stick with the current version.)

Cheers

Mirko


2008/11/27 Charlie Collins <charlie...@gmail.com>:

[Arthur Kalmenson]

Congratulations Charlie! I look forward to your collaboration (and the
prefix update :P).

--
Arthur Kalmenson

[Charlie Collins]

Thanks guys, I will try to make it as smoothe as possible.

Also, as to the current version, it seems to be working fine for me
too, without the prefix change. I don't understand exactly why it's
affecting some people and not others, but for the record it works here
as is too.

On Nov 27, 7:53 am, "Mirko Nasato" <mirko.nas...@gmail.com> wrote:
> That's excellent news!
>
> (And in the meantime since we're not affected by the prefix clash here
> we'll just stick with the current version.)
>
> Cheers
>
> Mirko
>

> 2008/11/27 Charlie Collins <charlie.coll...@gmail.com>:

[Mirko Nasato]

2008/11/28 Charlie Collins <charlie...@gmail.com>:

>
> Also, as to the current version, it seems to be working fine for me
> too, without the prefix change. I don't understand exactly why it's
> affecting some people and not others, but for the record it works here
> as is too.
>

That's what I was trying to understand in the other thread, before we
jumped to the "what should the new prefix be" debate. ;-)

Maybe there's an explanation, e.g. it doesn't work with maven 2.0.8
but it does work with 2.0.9 or something like that.

Cheers

Mirko