Mailing-List: contact derby-dev-help@db.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Derby Development" <derby-dev@db.apache.org>
Mime-Version: 1.0 (Apple Message framework v619)
In-Reply-To: <41262C44.4020002@debrunners.com>
References: <41262C44.4020002@debrunners.com>
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <8E8B9BFC-F2F9-11D8-8EED-003065CBB010@nonintuitive.com>
Content-Transfer-Encoding: 7bit
From: Cloudscape Testing <fuzzylogic@nonintuitive.com>
Subject: Re: [VOTE] on repository layout
Date: Fri, 20 Aug 2004 15:37:45 -0700
To: "Derby Development" <derby-dev@db.apache.org>

Looks like my mail program decided to reset my username for me. 
Hopefully this will go through as Andrew McIntyre instead of Cloudscape 
Testing. :)

I vote for option 1, because I would think that a developer who is 
checking out the code from svn should be able to check out one 
directory (in this case, derby/code) and get everything necessary to 
build a binary distribution and nothing more.

That means though, that I would include the user documentation in the 
code part of the repository, since we would want to keep that in sync 
with the code that it documents. The developer would also then have the 
current doc set in their build tree for reference, and a person could 
build a full binary distribution from the source distribution.

Are there source files for the docs, which are then generated by 
forrest? If so, I would think the repository would look something like:

derby/code {trunk/branches/tags}
- derby code
- derby tests
- compiled user documentation

derby/site {trunk/branches/tags}
- website pages
- source for derby documentatio
- derby technical papers

my .02$,
andrew