geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jacek Laskowski" <ja...@laskowski.net.pl>
Subject Re: Where to develop plugins (was: Re: JPA Plugin patch)
Date Fri, 01 Sep 2006 15:27:10 GMT
On 9/1/06, Aaron Mulder <ammulder@alumni.princeton.edu> wrote:
> There are several issues here:

I knew you'd have helped me to sort out some issues. I did know!
Thanks Aaron for taking your precious time dealing with my comments. I
wish there were a way to buy you a beer or two for your careful
explanations. Ok, enough for the starter...let's move on to further
drill down into your reasoning of moving to SF. I feel we're drawing
to a close....

> 1) Do I think we were wrong to develop the JPA plugin outside
> Geronimo?  No.  We can argue about this as long as you like.  Do Maven
> committers work on Mojo plugins?  I'd be surprised if none of them
> did.  Are you saying that nothing but server/* is RTC?  ("I'd merely
> point out that any activity done outside the trunk or bug fixes don't
> have to be RTCed.")  That wasn't my understanding -- doesn't RTC cover
> devtools and xbean?  Still, it's irrelevant, because Cliff and Roberto
> are not committers and could not have worked easily in the sandbox
> either.  Cheer up -- Chariot could have tried to make this
> proprietary.

As far as I know, only Geronimo trunk is in the RTC hug. I kept
wondering why XBean (as a subproject) was in it too - possibly
they^H^H^H^Hwe liked to be hugged...ekhm...loved too ;-)

I don't think devtools is RTCed, either. Only Geroniom trunk is in
great love of RTC ;-)

Re: developers developing plugins - you're certainly right, we're not
developing one monolithic codebase with all the features one could
dream of. It's impossible and I understand why people could develop
their plugins outside. They simply don't have necessary karma to do it
here. That's a very important matter I've not thought of before.

What I was not happy with was you as a Geronimo committer who's been
doing plugin development outside. Now, when I think about it a bit
more I think I get the gist of it - you couldn't work with your
buddies, actually, they couldn't work as easily as they can do outside
- the rest of the jpa plugin team don't have the karma to do it here.
Got it and have to think about it a little to let my mind settle.

> 2) Should the code be merged into Geronimo?  We're happy to donate the
> code if Geronimo wants it.  I wouldn't want to do that until G 1.1.1
> and the first installable version of the plugin is released, but
> hopefully that will be quite soon.  If you want to start the wheels
> turning for a donation, be my guest.  Is it enough to post code to a
> Jira with the Apache flag on?

I think so, but having read your comments above I think it would be a
pain for your fellows, wouldn't it? They couldn't develop the plugin
in an easy way, but sending patches on and on before they get commit
and even then their work could be RTCed. I like your explanation of
the plugin being part of Geronimo only when a user (not a Geronimo
committer) really needs it  via declaring it as a dependency.

> 3) You must not understand the plugin system very well to make
> statements like "It's also disruptive to the community as they need to
> look it up in their notes where the plugin comes from rather than
> download it from a Geronimo space.".  A user doesn't need to know
> where a plugin was built -- all plugins can be included in the plugin
> repository and this one certainly will be.

The last sentence was the main point of my statement. It will be or
will not. It depends on what is a list of repos in Geronimo distro.
Provided the plugin finds its place in an official repo that's
available in Geronimo console, it's fine, but you can't be sure it
will. I don't see a reason it will not, but that's part of my
statement.


> I think your comments show that we have a fundamental disagreement
> over the purpose of the Geronimo project.

As you might've seen after my replied above, not necessarily. I'm just
trying to understand your reasoning and the only way I'd chosen was to
challenge it - very risky, but I knew you'd not minded ;-)

>  I believe Geronimo should
> hold the essential glue that ties a J2EE server together.  I do not
> believe that it should hold more than that.  XBean is used standalone,
> so while I'd like Geronimo to adopt it, I think it should be a TLP.  I
> think OpenEJB and ActiveMQ should be TLPs.  I don't think Jetty or
> Tomcat or HOWL or Axis should be rolled into Geronimo.  It would be
> great if the Transaction and Connector core code were split off and
> maintained as part of Jencks.  What's left for Geronimo?  The
> configuration and deployment system, the console, the module
> configuration, the deployment plan syntax, the DConfigBeans,
> everything that ties all those critical components of a J2EE server
> together, and gives the developer a consistent experience across
> module types.

Well, let's drill it down and think about it a little more. If all
these parts you mentioned were separated as optional dependencies (or
plugins) would Geronimo become another IoC container? How different
would it be than XBean or Spring or HiveMind? Do we need another one?
It would be great if we could strip off these additional services and
provide plugins so that XBean or any other IoC container pull them
down and build Geronimo THE container. It's not a new idea in our
team, is it?

> Certainly I would like to see more plugins developed, and I really
> don't feel any particular urge for that to be done within the Geronimo
> project.  I think we should let go and encourage all plugin
> development no matter who or where.  But as I said, if the community
> wants to take over the development of certain plugins, that's OK too
> -- I just don't think it's necessary.  And I certainly object to being
> told that I have a responsibility to do plugin development *only*
> within the Geronimo project, because I don't think that's what the
> project is about.

I hope you've already noticed it wasn't my intention you drawing such
conclusions. Just challanging yours to become better, leaner, happier,
etc. ;-)

Your explanations were as perfect as the movie you recommended - The
Princess Bride - which I liked very much. Thanks Aaron!

Jacek

-- 
Jacek Laskowski
http://www.laskowski.net.pl

Mime
View raw message