ant-ivy-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "riki.eng" <>
Subject Re: Migration to IVY
Date Fri, 05 Sep 2008 20:31:43 GMT

I did this kind of work for my company. We had a very similar situation.

The solution I adopted was pretty manual but was giving me a good
representation of what the current state was.
Basically I got a "snapshot" of the libraries that we were using. I
rearranged each one of them following a maven2 repo format and generated ivy
and pom files based on the way we are currently using them. The reason why
we didn't get the pom files from the main maven repo is because I didn't
want to introduce additional libraries (dependencies) in this process. Like
I said, the end result had to be exactly what it was before, but in a
structure that was getting us ready for the integration with ivy.
In the first stage (for backwards compatibility) the libraries were still
stored in SVN, later we will pull them out.

Hope this helps...


Mathew Joseph-4 wrote:
> Hello there,
>    The 3rd party dependencies that our project depends on are stored in
> subversion and are retrieved as part of an svn checkout along with the
> source code. Dependency management is non existent. We are about to
> migrate to an IVY based system and I wanted to know what is the
> experience of people who might have attempted something similar.
> Steps I will need to take.
> 1. Identify exact version number of the dependencies and transitive
> dependencies that we are using 
> 2. Move them into a maven or custom repository format hosted either in
> subversion or plain file system served by http.
> I believe the idea is to ensure that I only resolve the first level
> dependency in my scripts and all transitive dependencies are pulled in
> by IVY. However, this means my custom repository needs to have either
> via pom.xml or ivy.xml defined for all first level dependencies. 
> I was wondering that it will be very time consuming for me to recreate
> the artifact repository from scratch. Do you guys think it makes sense
> to create a ivy.xml with version numbers and have ivy resolve them from
> a public repository. Once I do a retrieve, I then use the local cache as
> the first version of my custom repository and then start adding in all
> the dependencies that I did not find in public repos. This will ensure
> the repository has ivy.xml for all dependencies.
> Then close out ivysettings.xml to use only the internal http resolver as
> opposed to looking out in public repos.
> I am more inclined to use a maven repository style as opposed to a
> custom one since I can then use maven repository managers to handle
> proxying and mirroring other public repos.
> My thoughts are a bit random on this, but I hope you get the general
> picture.. 
> regards,
> Mathew Joseph
> Build & Release Engineer
> Bravura Solutions

View this message in context:
Sent from the ivy-user mailing list archive at

View raw message