Mailing-List: contact ant-dev-help@jakarta.apache.org; run by ezmlm
Importance: Normal
Subject: Re: cvs commit: jakarta-ant/lib - New directory
To: ant-dev@jakarta.apache.org
Message-ID: <OFAA80638D.D998FB4B-ON852569A5.0048F0EF@raleigh.ibm.com>
From: "Sam Ruby" <rubys@us.ibm.com>
Date: Tue, 28 Nov 2000 09:12:37 -0500
MIME-Version: 1.0
Content-type: text/plain; charset=us-ascii

Peter Donald wrote:
>
> > I too am tired.  For example, I know from past experience
> > that as a committer if I were to -1 this, my objection
> > would be ignored.
>
> Nope it wouldn't - at least not by me ;) Only one person
> has expressed a -1 so far (Connor IIRC) and

Ok, for the record, -1.  It is also my reading of
http://jakarta.apache.org/site/decisions.html that one veto should be
sufficient, even if it is "only" Conor. ;-)

> I have offered a solution that would solve your objections -
> what do you think about it?

I'm sorry to say that for me, your proposal actually solves the wrong
problem.  If there was some way to reliably determine what packages were
used to build the product, I have no problem with a build process producing
an installation image that exactly matches the build.

What I would like to be able to do is easily and reliably control what
versions of dependent products are used in any particular build.  For
background, most of my work to date has been involved with integrating
multiple products.  (FYI - this is still true at this moment, but the
products are not Apache ones)  This task is often difficult and the way I
like to proceed is to start with the source of every product and build them
as if they were a single product.  This way when I find I have to debug a
problem, I can reliably bring up the source which matches my executable,
make a fix, build it, and run the results.

- Sam Ruby