Le jeudi 26 octobre 2006 à 07:46 -0400, Alex Karasulu a écrit : > Ersin Er wrote: > > 0 Java5 migration > > -1 Subprojects merging > > > > MINA's current project structure seems almost perfect to me. It gives > > you dependency granulalarity and provides Java5 support where needed. > > I do not see a strong reason to do the both moves. But if Java 1.4 is > > preventing MINA to support new features, Java5 move can be considered > > more seriously. > > I was questioning the Java 5 move myself. > > However I think there are some real reasons why MINA should move in that > direction: concurrent libs, better performance with improved locking > constructs. I think eventually we're going to have an even better MINA > with 1.5. > > We still have the 1.0 branch for those interested in sticking to 1.4. > > What I do recommend is that MINA jumps to version 1.5 instead of 1.1 to > give a cue to users that something big has changed. A switch in JDK > versions is not a simple jump like the one from 0.9 to 1.0. It's a bit > more serious than that. You do want to give the users some kind of > hint. When they see you have jumped from 1.0 to 1.5 they will ask why > at least internally and take a closer look to see the differences if not > directly ask on the list. > > If we bump MINA to 1.1 and unsuspecting users move to 1.1 they're going > to be surprised and perhaps a bit unhappy when they realize JDK 1.4 is > no longer supported. Again they'll have to dig around or ask the ML but > in the end they're not going to be happy. > > In the end I think it's important to have some consistency in the way we > do versioning. I know there are some loop holes but we'll be consistent > in showing that something big has happened by skipping a few minor > versions as did the Tomcat folks. > > Alex
i think you made a point here. Jumping to 1.5 or 2.0 is an idea. I'm sure as a user I'll think 1.1 is nearly compatible with 1.0; but the API greatly changed and JDK requirement too. Julien
