Premature version bump has resulted in a cancellation - Jason
On Mon, Jun 4, 2018, at 11:50 AM, Jason E Bailey wrote: > Eh, I'm a bit confused over why a jdk requirement change is not > considered breaking, and I don't see what the problem is with iterating > the release rather than the version. But I'm good with changing it back. > > I'll cancel this when I get the chance and reset the releases and the pom. > > - Jason > > On Mon, Jun 4, 2018, at 11:39 AM, Robert Munteanu wrote: > > On Mon, 2018-06-04 at 08:48 -0400, Jason E Bailey wrote: > > > The major version change is from a release perspective. The only > > > change to the versioning which OSGi uses is > > > org.apache.sling.api.resource which went from 2.11 to 2.12 > > > > > > Ioan brought up the issue as part of the pull request that the > > > upgrade to jdk 8 is a significant change. If someone is running > > > sling on a jdk 7 environment then this release will be broken for > > > them. I looked around at other Apache projects and there seems to be > > > a trend that upgrades to JRE support results in a major release > > > upgrade. > > > > > > This would also allow support, if there was ever a need, to do a > > > release for the jdk7 version after this release. > > > > This is not our current practice - we bumped versions from 5 to 6 and 7 > > without bumping major versions so I'd suggest we keep doing that. I > > think the bigger suprise would be that we increase the major version > > component without an actual breaking change :-) > > > > Thanks, > > > > Robert
