On 2009-12-29, at 12:48 PM, Kristian Rosenvold wrote: > Jason van Zyl wrote: > >> I honestly think it will be easier for people to get involved in the 3.0 >> codebase. > > Somewhere along the line in the flurry of emails that have been coming > along today I got the impression that a transition to guice has been > called off for M3. >
Taken us longer to make the adapter then we expected, and if we want 3.0 out in a reasonable timeframe it's just not something we can do. > I think that's a shame, and having only recently gotten a little "more > involved" with the maven code-base, I have felt the pain that plexus can > inflict. > In all honestly it affects almost no users. You're an advanced user, and we've had the same problems with Guice and we've had to create some shim code and make some changes in Guice as well. Nothing is perfect but leaving Plexus in 3.0 isn't going to affect average users. If the ITs work I don't see much value in rushing the switch. We testing all the edges with Nexus and when that's done we'll try Maven. > I will try not to start any plexus-bashing, but given the discussion > that has been going on today about how M3 is a significant cleanup of > inner workings, I find stopping before getting rid of plexus strange. > Then you don't really understand how much work it is. It's impossible for anyone to see this when they don't work on this full-time. > I would be more than willing to spend a fair amount of time to make such > a transition happen, the sooner the better. Especially if there already > is a working plexus/guice adapter that can be used as basis. I get this > feeling I missed some discussion somewhere ? I don't think you honestly understand the ramifications of changing the underlying system. The grid has slowly been built up over time to try and automate much of this and the ITs catch almost everything these days but it's not a small or trivial job. When we have found problems Stuart has been able to work with Jesse to patch Guice but we're still finding things and we fix them as we go. I think it's wiser to fully convert one system completely first, in this case Nexus, and then attempt the next system. > > But then again, we all must release sometime - I just needed to vent > my frustration with plexus. Switching to guice also has a motivational > effect on people wishing to work with the maven codebase; developers > like to work with code they feel is "hot". Well, I'm not really concerned as a first priority whether you think the code is "hot". I care about it being stable, making sure the grid is in place to catch anything, and that the ITs actually represent something we can rely on. The Plexus shim for Guice is here: http://svn.sonatype.org/spice/trunk/spice-inject/ Stuart has started playing around trying to get this to work with Maven but, again, we're focusing on Nexus first. But if you want to setup some builds and give feedback, patch guice, help setup the tests on the grid for validation and help write new ITs then that would be great. Myself, Stuart and Benjamin would be happy to get you started and point you in the right direction. Given what I know about how the resources available to work on this and the general level of activity in the core I don't see Guice getting in for 3.0. But if it can be proven that a version of Maven running on top of Guice is a viable replacement then users won't notice and it can go in when it's ready somewhere along with 3.0.x line. > > Regards > > Kristian Rosenvold > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org > For additional commands, e-mail: dev-h...@maven.apache.org > Thanks, Jason ---------------------------------------------------------- Jason van Zyl Founder, Apache Maven http://twitter.com/jvanzyl ---------------------------------------------------------- --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
