On 5/24/25 15:33, Elliotte Rusty Harold wrote: > On Sat, May 24, 2025 at 1:16 PM Tamás Cservenák <ta...@cservenak.net> wrote: >> ... aaand because some company/ies does/do that, >> we as an open source project/forget need/must stick to it? > The claim that we must move our JDK version forward because Java > 8/11/17 is not supported is fallacious. It is based on a false > premise. > > If you want to upgrade the minimum Java version, you need a better > reason than that. I haven't yet heard a reason strong enough to > convince me. Others have other opinions.
I would be curious what the strategy would have been if maven 4 would have decided to stay on JDK 8. Wait till 2030 and then throw maven away and build something new? Do one big jump from 8 to 2x and hope the community would accept it? Hope that vendors would keep supporting some slightly newer JDK versions beyond 2030 to get a few extra years out of it? If vendors would not stop supporting JDKs, maven would have 7 JDK versions to support in 2030 - this is quite a test matrix (not to mention the wasted resources). Trying to keep the delta between 'min' and 'current' constant sounds like the pragmatic approach for projects which plan to stick around. In fact many did and introduced some easy to remember, forward looking rule like "our latest major version supports 2 LTS + current". The crazy part about this discussion to me is that maven already has a solution for this. Maven 3 exists, runs on JDK 8 and is supported as long as the community wants to support it. Maven 4 could even do a RC more and be JDK 25+ without influencing anyone on JDK 8. >> So basically, all we did so far was "waste of resources" >> as Java 8 is there to stay, at least until 2030, right? > IMHO, yes. I would much rather people invested their time into fixing > Maven bugs and improving the health of the codebase rather than > continuously migrating code from one Java version to the next. Bugfixes and refactoring are luckily not mutually exclusive and are just regular project maintenance tasks. The cumulative time spent writing JDK version discussion mails is probably already more than the review of a fictional global JDK 17->21 migration PR would take. (that is another reason why many projects came up with an easy to remember min-version rule rather than having to repeat this kind of threads every few years) There is also no time pressure behind language level cleanups. They can (and often should) happen incrementally, as needed, without disrupting any other tasks. The point in time when the min requirements are bumped is only the starting point which unlocks more cleanup options - the rest follows organically. > That's > life with a volunteer project though. Developers are going to do what > they find fun rather than delivering user value. no fun allowed while producing value ;) I could flip the argument and say that trying to reduce aspects which are generally perceived as not-fun, probably creates value for everyone over the long term, including users. Esp in open source projects which rely on motivated developers spending some of their free time (or carefully negotiated work-time) helping to maintain it. what lucky situation to be in when developers perceive modern APIs / language idioms / cleaner code as more fun :) (btw: is it ok to use vote threads for discussions?) best regards, -mbien --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
