On Wed, Apr 16, 2008 at 1:38 AM, Xavier Hanin <[EMAIL PROTECTED]> wrote:
> The link gives a 404, so I can only say what I think of the patch you > attached in your last e-mail. IMO, the patch quality is good overall. The patch has been updated a few times since I sent that email. New patch is here<http://ivyroundup.googlecode.com/svn/wiki/files/builder-resolver.patch>(or view directly<http://code.google.com/p/ivyroundup/source/browse/wiki/files/builder-resolver.patch>in SVN browser). > think it still requires some more test (I don't think there's a test of > m2resource for instance), and also some more comments in the code: I know > Ivy code itself is not a very good example, but including code means > maintaining it, and some more comments, especially in the xslt, may help. > I will add a m2resource access to the unit test and more code comments and update the patch. Besides the code quality, we need to evaluate the usefulness for the > community. ATM, even though I like the idea, it seems the number of people > interested is still low. Moreover, this resolver could be provided as a > plugin, and as such don't really *require* to be included in Ivy core. So > maybe you could start by providing it as an easy to use plugin, then we'd > integrate it in Ivy core once the community interest will be large enough. > We can still add a link or two in the official documentation (in the > resolvers and links page for instance) to favor the adoption. OTOH, I > encourage you to ping the ivy user mailing list about this project, you > can > even do it once a month to report progress and encourage people to join. > Blogging and announcing is also a good way to get traffic and raise a > community. I'd ask that you please reconsider this recommendation. I think the new resolver is generically useful. Initial feedback on the ivy users list has been encouraging. Including this in 2.0 would be a new feature addition that people would appreciate and help show Ivy development and innovation is alive. Moreover, nobody is going to bother to download a separate plugin. Making this a plugin is equivalent to killing the idea. We want to encourage people to use Ivy, in whatever ways they want... not make it hard to do so. The builder resolver is generic and not at all tied to the RoundUp project, which is entirely separate. Contrast with the ivyrep resolver (which is included in Ivy) which although now deprecated was originally tied to a specific repository. Also, I've included unit tests, complete documentation<http://ivyroundup.googlecode.com/svn/wiki/files/builder.html>, the code passes checkstyle, etc. I am already an Apache committer (from early Harmony project days) and am willing to maintain this little piece of the project. In short, if you think this is a good idea overall, I don't see what downside there is to going ahead and including it in Ivy now. The potential upside, however seems very positive. We need a well-organized community repository! > As I said in my previous e-mail I think better name consistency would > really > help. I don't have enough time to get actively involved ATM, but I'd be > happy to at least share my ideas, so ping me if you create a dedicated > mailing list (you can also use ivy user list at the beginning). > Thanks. Please consider adding to the RoundUp wiki site, especially the ModuleMaintainerGuidelines<http://code.google.com/p/ivyroundup/wiki/ModuleMaintainerGuidelines>page.. your experience would be very valuable here. I will work on setting up a mailing list too. Thanks, -Archie -- Archie L. Cobbs
