On Thu, 10 Jan 2019, 16:02 Daniel Farrell <[email protected] wrote: > On Thu, Jan 10, 2019, 7:31 PM Robert Varga <[email protected]> wrote: > >> Hello everyone, >> >> as I noted in previous updates, our Neon release stream looks rather >> good where JDK11 compatibility is concerned, with my local builds with >> JDK11 passing with minor fixes (submitted) up to and including >> openflowplugin. >> >> With the three version bumps tracked at TSC-186, TSC-187 and TSC-188, >> this should be readily reproducible without resorting to my local hacks. >> >> I would like to ask individual projects to do some build testing to >> ensure they work with JDK11 in the time we have left before Neon code >> freeze sets in (1/24/2019), but as this was never a requirement to >> participate nor a TSC-mandated goal, it really is up to them. >> >> The big question is how are we going to go about adopting Java 11 from >> the point of: >> - it being a fully supported runtime >> - it being the minimum required runtime >> >> We have one prior experience with this upgrade, which was the transition >> to Java 8, which was executed as follows: >> >> 1) We started preparation in Lithium, but have not completed it, >> actually regressing on some front >> >> 2) In Beryllium, we have added parallel verify jobs to run each patch >> with Java 7 and Java 8, with merge and autorelease jobs running with >> Java 7 only >> >> 3) In Boron, we have bumped required source to 1.8 and ran all >> verification and release jobs with Java 8 only >> >> If we follow this model and start executing now, Sodium would be the >> combined Java 8/Java 11 release delivered in September 2019 and >> Magnesium would be a Java 11-only release delivered in March 2020. >> >> The downside of this is the toll on our infrastructure, which means our >> infra costs for Sodium would essentially double -- hence it may not be >> an entirely feasible option. >> > > +1 > > I don't think this is a viable option given our need to substantially > reduce infra spending. > >> >> As an alternative we could skip the combined Java 8/Java 11 release, >> going directly to a Java 11-only release -- either with Sodium or with >> Magnesium. >> > > +1 > > It seems reasonable to target Sodium to me. >
Speaking in an individual and personal capacity as upstream contributor and TSC member, I would support this. Perhaps conditional to just making sure that we actually can successfully build and run and CSIT say at least all mananged projects in autorelease on Java 11 at the start of Sodium? I would, of course, want to avoid a "broken world" - no surprise there to anyone with me saying this? ;-) I would like to start the discussion around the options we have and >> preferences of both OpenDaylight projects and our downstreams. >> > Speaking officially for Red Hat, we may need a week or so to internally discuss and confirm how this aligns with plans and time frames. My expectation is that it will be OK, but give us some time, and I will get back speaking for us and not just myself. Ericsson friends, what would your position on this be? Everybody, please chime in. >> >> Regards, >> Robert >> >> P.S.: unlike when we first had our Java 8 discussion, Java 11 seems to >> be supported by all major Linux distributions right now. > > >> _______________________________________________ >> release mailing list >> [email protected] >> https://lists.opendaylight.org/mailman/listinfo/release >> > _______________________________________________ > Discuss mailing list > [email protected] > https://lists.opendaylight.org/mailman/listinfo/discuss >
_______________________________________________ Discuss mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/discuss
