Don't forget that successive versions of some plugins may break backward compatibility and other such bad practices. Locking everyone in a large organization down to one version of such a plugin could be very limiting, since these things have to be phased in.

Also, I don't think we can pretend that we have all of the requirements or use cases. I've heard some pretty convoluted approaches to managing this sort of data, and I don't have the understanding of the environment to make a judgement about whether they are doing things in an unnecessarily complex way, or even whether their approach can be changed (political environment can have a big impact here).

I do think that mixins like this would be beneficial, and I'm really not at all convinced that it's a good idea to expand the scope of such mixins outside of the build element...so, basically plugin packs. To me, the term 'plugin pack' still needs some definition, so I'm willing to say that it's a good idea to provide this type of flexibility, and then help shape the concrete details a little bit to get at something reasonable.

As far as many of the complexities involved in your questions at the beginning of your reply, we can use tooling to help with that sort of visibility too. Also, supporting plugin packs for the wider world doesn't preclude having internal policies against their use in organizations...maybe we'd do better to help people set rules about how Maven is allowed to be used?

-john

On Sep 2, 2007, at 11:30 AM, Jason van Zyl wrote:

What are the real requirements?

They are:

1) An easy way to get a set of stable plugins that work together
2) To easily see what versions are contained in this set
3) To easily change or augment what is in this set

The current mechanism + toolings works. I know what's going to happen with plugin packs. Someone is going to want to change a version of a plugin they are using and "How do I find out what versions of plugins am I using", "How do I override what version of a plugin I'm using if it's specified in a plugin pack?", "Does plugin management win if it's in a plugin pack?". "I found a bug in the way plugin packs are processed and I can't get a plugin version I need, is there a work around?". "How do I configure plugins that have been defined in the plugin pack". So people are going to have to end up redeclaring bits to get configuration and execution information locked down and then you end up with a terrible configuration management problem.

A fully, and clear, literal way to define this is what is most practical. The questions below are also framed to bias the answers. For A, plugin packs are not the only solution. In very practical terms the total number of plugins is not that high. What people want to know is the stable set. The core processing required for the notion of a plugin pack will not be straight forward and it's not necessary and adds almost no value and I'm certain it will lead to greater complexity.

Users want 1), 2), and 3). The current mechanism plus minimal tooling, or no tooling if people cut and paste (big deal) covers those requirements. Plugin packs cover 1) and then it becomes another nightmare to maintain. I am not in favor of plugin packs.

On 1 Sep 07, at 7:53 PM 1 Sep 07, Brett Porter wrote:

Like the other poll, I'd like to hear from as many people as possible their opinion this topic (even if you just want to say '0' so we know where you stand).

[ ] (A) Having a way to include a set of plugins in one small POM fragment would be a useful feature to have (if you have a use case other than the already stated "standard plugins", please specify)
[ ] (B) Pasting a snippet in from the web site is sufficient
[ ] (C) No opinion
[ ] (D) Undecided
[ ] (E) Other (please specify)

Thanks,
Brett

--
Brett Porter - [EMAIL PROTECTED]
Blog: http://www.devzuz.org/blogs/bporter/

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


Thanks,

Jason

----------------------------------------------------------
Jason van Zyl
Founder and PMC Chair, Apache Maven
jason at sonatype dot com
----------------------------------------------------------




---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


---
John Casey
Committer and PMC Member, Apache Maven
mail: jdcasey at commonjava dot org
blog: http://www.ejlife.net/blogs/john


Reply via email to