http://gwt-maven.googlecode.com/svn/branches/cc_20080814_automaticrefactor/
Instead of gwt-redux project I am now changing course a bit and
working in a branch of GWT-Maven. Many of the cleanups from gwt-redux
and in this branch, but I am not injecting all the GWT deps for all
phases based on the previous discussion and feedback (it was a bad
idea to go that way ;)).
At any rate, the branch works for my Linux tests thus far, and has a
lot of cleaned up stuff. Major points are:
1. No longer has setup and extract for automatic, instead uses
dependency plugin (see simplesample example in branch - this was done,
even though setup/extract was brilliant and is great, to reduce the
complexity and burden here, and achieve the same effect with standard
maven conventions)
2. Still supports concepts of "automatic" or "manual" mode. Automatic
uses dep plugin (as noted above) and central repo. Manual still uses
google.webtoolkit.home.
3. Uses the central repo for GWT deps - including native libs (with
classifiers, etc)
4. Injects gwt-dev and gwt-user for the run and compile scripts (only)
using a cleaned up utility class (along with a lot of other general
cleanup/refactor). Close to the same result as what we did before,
but much cleaner flow (IMO).
5. Many bug fixes are in the branch based on comments and patches in
the issue tracker. I commented on each issue in the tracker for those
that are in the branch, and made sure they are accepted - but didn't
close any yet.
6. Documentation has been updated and standardized too (uses plugin
plugin for goal info, for example). It's not published yet, but mvn
site demonstrates the new stuff.
7. I haven't tested the bean gen stuff. I don't use it, and I can't
get a hold of Cooper (who wrote it), haven't heard from him in
months. There are probably many bean gen issues, but I don't have
desire or time to get into it (there have been a lot of issues related
to it on the tracker for a long time too). If anyone here wants to
tackle that, let me know and I will add you to the project. If it
doesn't get much traction I might just remove that stuff (though it is
cool, it's not useful it its not maintained) and make it a separate
project (maybe should be a separate maven plugin for GWT anyway, that
would potentially be cleaner).
8. I don't know when it will make it to a release in the trunk, but
soon. I want to test more on mac (no difference with leopard in 1.5,
or not) and windows - and I want to test to make sure OLDER GWT
releases still work fine too.
As always, all feedback is welcome. Any contributions are welcome,
testers, documenters, coders, etc.