Mailing-List: contact dev-help@excalibur.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Excalibur Developers List" <dev@excalibur.apache.org>
Received-SPF: neutral (hermes.apache.org: local policy)
From: Niclas Hedhman <niclas@hedhman.org>
Organization: Private
To: "Excalibur Developers List" <dev@excalibur.apache.org>,
   Berin Loritsch <bloritsch@d-haven.org>
Subject: Re: Excalibur project FOG=-100
Date: Fri, 10 Sep 2004 22:28:11 +0800
User-Agent: KMail/1.5
References: <200409101944.23628.niclas@hedhman.org>
 <4141B421.40508@d-haven.org>
In-Reply-To: <4141B421.40508@d-haven.org>
Cc: Gump code and data <general@gump.apache.org>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="windows-1252"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200409102228.11217.niclas@hedhman.org>

On Friday 10 September 2004 22:03, Berin Loritsch wrote:
> Niclas Hedhman wrote:
> > Hi,
> > I have seen that the Excalibur project bluntly consider Gump an
> > "annoyance", and pragmatically killed all the hard effort that I and
> > Stefan put into get Excalibur sources in order earlier this year.
>
> It was not intended to slight you or Stefan in any way.  The truth is
> that we really aren't sure how to get Maven builds and Gump to be happy
> together.  It's not something that is common knowlege and it is not easy
> to maintain.

Excalibur under Avalon was building with Maven, and had an Ant wrapper for 
Gump in place, mainly courtesy of Leo Simons I think. 
To keep it running can't be that big a deal.

Also, since then, I believe that the Maven runner has been enhanced, and 
should be able to build against your Maven POMs...

> <snip/>
>
> > I am considering reverting the excalibur projects in Gump back to the
> > Avalon version until the Excalibur community smartens up in respect to
> > Gump. At least the Avalon-excalibur codebase is build operational, except
> > for some alt-rmi stuff, which I perhaps can get working anyway.
>
> You are one of the few people who know how to get Gump running, but it
> isn't fair to you to be the only one maintaining it.  I'm not aware of
> any good step by step instructions to get Gump working again.

It is just about as simple as it could possibly get, but I agree that docs are 
spread out and hard to locate, many concepts are unclear who the target 
audience is (most the average maintainer doesn't need to know).
BUT, Stefan and Adam provides excellent support, and welcomes anyone making an 
effort.

> Alot of the Excalibur codebase is dependant on a number of other
> projects which does complicate things too.  If those dependencies
> change, then making it work again is hard mainly because of the
> lag time.  Make a change to the config file, wait 8 hours, wash/rinse/
> repeat until it comes clean.

I haven't checked, but the same goes down the line in even stronger waves. 
There were 200 projects in ASF depending on Avalon material, and I assume 
(have not checked) that there are something similar for Excalibur.
Ignoring such fact doesn't help.

> The real problem is making it easy.

No, the "real problem" is to be a bit more careful when one make the changes 
of the codebase, and ensure that Gump and everything downstream still are 
happy. Ignorance of this just creates problems for others at some point in 
time later. Small steps would basically never created this problem. There has 
been two major steps here; The move out of Avalon was one, in which case Gump 
was just ignored. Followed by a quite large 'clean up' operation, which also 
ignored any tie-back to Gump.

It is a damage that has been done, and I am not a Excalibur committer and 
won't be the one who tries to put it back in order.

> > Cheers
> > Niclas
> >
> > P.S. Really needed to get this off my chest, since there seems to be
> > certain levels of hippocracy floating around...
>
> To be honest, I am pro the concept of Gump, but all my attempts at
> maintaining the Gump build have been very frustrating.  I tried to
> get it set up on my local machine, but I don't have the space or the
> patience to get it working.  At least that is my reality.  I just wish
> it were easier to make Gump work and to test the Gump build.

I totally agree. Will it happen? I don't think any time soon, but it should be 
moving in the right direction by now. Whether Python is the best tool has 
also been debated, and I have no other opinion than it keeps many Java 
experts out of the loop. OTOH, the Java-based Gump had its own weird 
problems.


Cheers
Niclas
-- 
   +------//-------------------+
  / http://www.bali.ac        /
 / http://niclas.hedhman.org / 
+------//-------------------+


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@excalibur.apache.org
For additional commands, e-mail: dev-help@excalibur.apache.org
Apache Excalibur Project -- URL: http://excalibur.apache.org/