Mailing-List: contact ivy-user-help@ant.apache.org; run by ezmlm
Precedence: bulk
Reply-To: ivy-user@ant.apache.org
Received-SPF: pass (athena.apache.org: domain of adrian@last.fm designates
 170.20.116.180 as permitted sender)
Message-ID: <48F5C7BA.1040009@last.fm>
Date: Wed, 15 Oct 2008 11:36:42 +0100
From: Adrian Woodhead <adrian@last.fm>
User-Agent: Thunderbird 2.0.0.6 (X11/20070824)
MIME-Version: 1.0
To: ivy-user@ant.apache.org
Subject: Re: newbie: Instructions on hosting enterprise repository
References: <48F4C07D.7060704@reading.ac.uk> <48F4D806.8050307@last.fm>
 <48F5B0AC.5040000@reading.ac.uk>
In-Reply-To: <48F5B0AC.5040000@reading.ac.uk>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit


Garry Smith wrote:
> That is good to hear.  We currently have jars in our SVN that are 
> organised in the same way you mention (by 
> organisation/module/version). To move to the Ivy way of resolving 
> dependencies it sounds like we just have to include the appropriate 
> ivy metatdata for each existing jar.   I will have a play with the 
> 'creating a repository' tutorial to get an idea of the ivy 
> configuration files that will be needed in the SVN.
We found that for most cases we could just put the jar files in with no 
extra metadata and only in exceptional cases (e.g.jars which depend on 
other jars, a module consisting of more than one jar, the jar file name 
different to the module name) did we need to also add ivy.xml files.

> Great! This would fit my project's needs. Some of our developer's 
> don't want to build the components written by other developers in our 
> team.  They would rather each developer  uploads  a binary version of 
> their component(s) to a repository, which all other developer's then 
> retrieve as part of their build. Hence my arrival at the Ivy website 
> the other day.  I also want our automated builds to build all project 
> components from source daily (not that I don't trust the developers, 
> but... :-) )
Yup, this is exactly what we do and why Ivy was such a help in allowing 
sharing of functionality between different projects.

> On a side note, if we decided not to use SVN to host the Ivy 
> repository, but still wanted developers to check in modules from their 
> development machines, what would be the recommended setup? Are people 
> using Apache2 and Webdav to do this natively from Ivy? Is there 
> another recommended way that works natively from Ivy?
I'm pretty sure quite a few people go the Apache/Webdav route, not sure 
about other approaches, maybe someone else wants to comment?