On 3/15/07, Ross Gardler <[EMAIL PROTECTED]> wrote:
Xavier Hanin wrote: > On 3/15/07, Scott Goldstein <[EMAIL PROTECTED]> wrote: ... > Then if we see that users would be interested to share their Ivy files in a > central repository, we will talk with the repository team to see if they > are > ok to host Ivy files. But this require an important amount of work of > maintenance, and the Ivy community is not strong enough for the moment > IMHO. Over at Apache Forrest we are moving to Ivy (very slowly at present). We are discussing whether to use the Maven2 repos or whether to create a local repo to Forrest with the intention of passing it over to the Ivy project once we are satisfied all dependencies are correctly described in ivy.xml files.
The decision is not easy. On one hand you get a bunch of existing metadata, but not always of enough quality or flexibility. On the other hand you've got a lot of work to write your own metadata, with the big advantage to keep control over them. For a private project I would advise to avoid using maven repository anyway (or any other public repository), because keeping control over the repository is very important (see the best practices page for justification [1]). For an open source project it's different, maybe the solution is somewhere in between, leverage the maven2 repo and write Ivy files only for modules for which you are not satisfied with their pom. Feel free to come back to this list (or the [EMAIL PROTECTED]) if you want further opinions or suggestions. Since Forrest is based on Apache Cocoon there is a fairly large number
of modules we will be working with. However, it will be much easier for us to just use the Maven ones.
I don't know if you've talked with Jean-Baptiste Quenot, he made a very good work at migrating back Cocoon build to Ant+Ivy. Cocoon didn't make the switch, but maybe he worked on metadata too which could be interesting for you. So, I read this thread with interest and look forward to understanding
the best way forward for the Forrest project (right now I change my mind every time I think about it, the community is also undecided at this point). > Another point is that with Ivy you can quite easily have distributed > repositories, and I'm not sure that a central repository is the best. If > Ivy > gain momentum, I would prefer seeing each project using Ivy provide their > own Ivy repository with their modules, and an ivyconf.xml for this > repository with proper per module configuration of resolvers. Can you point to the relevant doc for me to do background reading on this?
For the moment there's only reference documentation: The configuration [2] page gives an example where you define a resolver for all modules from the same org (with further details on the module element page [3]). There is also the configuration file include page [4]. HTH, - Xavier [1] http://incubator.apache.org/ivy/doc/bestpractices.html [2] http://incubator.apache.org/ivy/doc/configuration.html [3] http://incubator.apache.org/ivy/doc/configuration/module.html [4] http://incubator.apache.org/ivy/doc/configuration/include.html Ross
