Le 07/07/15 11:50, John Hartnup a écrit : > Interesting read. It looks like we consumers of Mina have some work ahead > of us if we choose to upgrade.
Yes. Hopefully, for the better... > > I would note that keeping the name while changing the API can be a real > nuisance for developers who work in a certain way. I copied some example > Jersey configuration, which didn't work. I spent days Googling for > solutions, finding lots of people apparently succeeding with the same code > as mine. Eventually someone on Stack Overflow said "Oh, those are Jersey 1 > properties. Jersey 2 is completely different." This is a very complex question. The name is the project, and it's hard to change it. This is also the reason we have major versions, vs minor versions. Also, consider that MINA 2 is likely to be maintained. At some point, MINA 2 is also showing its age : it's support NIO, not NIO2, and none of the new Java language goodies since Java 6 are simply ignored. So far, so good, because 99% of the time, it's enough. IMHO, I do think that MINA 2 should not evolve too much : people are used to it, it's doing the job, and there is not too much room for improvement without breaking the internal structure. Although, it's far from being perfect, and we would like to rethink many design decisions that are not so good, in restrospect. Bottom line, yes, for the users, it's not a lighweight decision to take. Also it's most certainly premature, as we are at the early stage of MINA 3 design. I would say taht those who want to experiment, and even want to be part of the MINA 3 effort, then it's a great opportunity. I insist : we welcome anyone wanting to pbring some value to the project !
