Mailing-List: contact gump-help@jakarta.apache.org; run by ezmlm
Reply-To: <ajack@TrySybase.com>
From: "Adam Jack" <ajack@TrySybase.com>
To: "'Gump code and data'" <gump@jakarta.apache.org>
Subject: RE: Gump in Limbo? (was RE: Specifying ANT files)
Date: Mon, 23 Jun 2003 13:37:56 -0600
Message-ID: <00ba01c339be$f3e6e230$b1c5ea43@wdn086>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: 7bit
In-Reply-To: <3EF4680B.8020906@apache.org>
Importance: Normal

	I can't tell you what you what to invest in.

No worries, I wasn't expecting you too. I am going into this w/ my eyes wide
open.

	However, whatever choice you make will influence others.

Nice to believe, but not so much as the choice you make... ;-)

	Each based on the assertion
	that Gump won't take off until it is rewritten.

I wouldn't be so presumptive. Gump has "taken off" (IMHO) it is achieving
what it set out to, it does so daily. If other folks don't install it for
private projects, so be it. I wonder if there are three reasonable
objectives for gump (1) central continuous integration builds for open
source, as an "automated build agent" (2) same but for private projects (3)
general multi-project build tool. Maybe gump isn't used for (3) yet, and
maybe it oughtn't be. Heck maybe that is a different beast, that isn't gump,
who knows. Whatever, I think gump needs to be judged against the three
separate objectives. A re-write oughtn't be to make it succeed (it has) but
to allow it to grow further -- perhaps into (2)/(3).

Python seems a good tool for developing all aspects of Gump (interaction w/
command line tools, creating output, forking builders) so a one language
implementation other than a multiple language implementation. I don't know
how to write Python, but I believe it is the best tool for this job that
I've seen proposed.

As a users I would rather learn/debug/patch that do the same w/ the scripts
output today. A such, I'd certainly +1 on Python.

	If development of the Python base were to rekindle, I would contribute
	to the effort.

	I simply don't want to produce yet another codebase for
	which I am perceived as the primary, if not sole, author.

My only "problem" with the Python base was it was heading in the direction
of (3) whilst (initially) leaving (1) and (2) behind. I'd happily work with
a Python gump so long as the (1) and (2) features are supported. I would
like to help w/ (2) -- so folks could customize the HTML output, and perhaps
connect to other source code repository types (than CVS). As you said, if
development of the Python base were to rekindle I'd assist as best I can.

	> Further, I couldn't find Gump in bugzilla (am I just missing it?) so I
can't
	> even contribute enhancement requests/bug reports. I have a few I need to
do.
	> Could somebody point me to the right place?

	That I can help with.  Project gump created.  Two components are
	defined: java/xslt and python.

Thanks, I'll plug some in later this week.

regards

Adam