Re: Gump changes

[Sam Ruby]

Nicola Ken Barozzi wrote:

Of course this is just MHO, I am not going to vote on this thing, since
I don't feel entitled to, because of my minimal code involvement with Gump.

I recall Sam having problems about changes in Ant that broke his builds,
and thus the creation of Gump (http://jakarta.apache.org/gump/why.html)
... I smell that this will bring the same problems, but if Sam is ok...

To the best of my recollection, I have only issued one veto in my entire involvement with Apache, and it was ignored (note: that veto wasn't on Ant, but it matters not).

Vetoes are not the way to build a community.

If users would just show interest and some patches, they would easily
become committers, and I fail to see why moving the code round will help
this involvement.

A solution would be to make two repos: gump and gump-repos, and have
gump-repos open to all *Apache* committers.

Why?  You can already host descriptors outside of Gump's control.

The descriptors were initially all in the Gump repo for better control
and cross-project check.

Making a common repo would continue this tentative, while not
relinquishing control of the whole Gump codebase to people that could
start bashing (am I being too defensive here?)

Is 'relinquishing' control a problem ?

Could be.

Chuckle. I have been *trying* to relinquish control for about a year now.

My ideal would be to have a reposititory in which every Apache committer has commit access. Those that wish to contribute are welcome to. All I would ask is that people who express an opinion are willing to do the work and provide support for the results.

If the 'bashing' results in
finding problems and improving gump - or even increasing awarness -
then it's should be welcomed.

Of course. *If*

There is a "competing" Gump project called Maven (which started in Gump
BTW).
Maven developers have access to j-c.
Gump developers don't have access to Maven.

It's the same reason why Maven developer would not want Maven moved
under j-c.

JvZ is a committer here. He contributed here for a while, then created an alternate proposal, then moved it to Turbine.

Will this make Maven work nicer with Gump while not assimilating it?
Could be a deal.
If not, I wouldn't like it... but hey, this is where I'm being
defensive, maybe I'm just plain wrong.

I'd love to see Gump assimilated by Maven. Or by Alexandria. Anybody willing to do the work is welcome to try.

As a start though I'd probably keep things as is, and suggest all
committers that ask the descriptor to be patched to move it to their
codebase and have it administered. Probably.

I think the real issue is where to host gump's code - i.e. the xslt and
java code.

My experience is that it is easier to keep these two together. I have no issue with people wanting to make changes to the xslt and/or java - as long as they are willing to do the work and support the results.

- Sam Ruby


--
To unsubscribe, e-mail: <mailto:alexandria-dev-unsubscribe@;jakarta.apache.org>
For additional commands, e-mail: <mailto:alexandria-dev-help@;jakarta.apache.org>