On the selfish side, personally I'd say that I have no compelling reason to upgrade (heck Hibernate is still using 1.9) because none of the bugs/requests I have reported have been addressed. These range from LONG standing, simple ones like standardized "provided" configuration support to more complex ones.
Given that I would actually want to help (either because an RC includes fixes for one or more of my reports, or because I am a good citizen) both of your suggestions are good. Out of curiosity, why limit latest to just rc? On Wed, May 7, 2014 at 10:50 AM, Daz DeBoer <darrell.deb...@gradleware.com>wrote: > G'day > It appears that a few regressions slipped into the 1.12 release. Here's > the list of possible regressions that I'm aware of: > > - GRADLE-3076 <http://issues.gradle.org/browse/GRADLE-3076>: Gradle > 1.12-rc-2 fails with UnsatisfiedLinkError on some Linux versions > - This was reported in 1.12-rc-2. Was it fixed for 1.12? > - GRADLE-3079 <http://issues.gradle.org/browse/GRADLE-3079>: Method in > build.gradle not available in imported scripts (regression in 1.12) > - This is the classloader regression in "apply from" that Luke is > working on > - GRADLE-3080 <http://issues.gradle.org/browse/GRADLE-3080>: gradle > 1.12 now downloads snapshots when I apply a version range > - Not sure if this has been verified, but I suspect it's due to the > changes to version listing that I did > - GRADLE-3081 <http://issues.gradle.org/browse/GRADLE-3081>: gradle > 1.12 tries to resolve custom packaging with .jar extension > - Not sure if this has been investigated/verified, but it sounds > credible > - GRADLE-3082 <http://issues.gradle.org/browse/GRADLE-3082>: "class > loader scope is locked" error when building GroovyFX with Gradle 1.12 > - This has been verified by Peter > > I guess we need to decide which of these should be fixed for 2.0, and make > fixing them a priority. We should also update the "Affects Version" for any > that are verified regressions. > > Regarding the underlying issue of regressions: the RC phase for 1.12 was > over 2 weeks long, so perhaps the reason these weren't caught is that > there's little pressure for people to try out the RC, particular when we > release so frequently. I'm not sure there's any "solution" to this, but > some things that might help are: > > - Release more frequently (4-6 weeks), so that regressions don't cause > too much pain and each release contains less stuff > - Make it easy for users to configure a build to always run with the > latest gradle. Something like "latest.rc" in the wrapper config, or a > command-line override. That way, more people might setup CI jobs that would > catch regressions in the RC phase. > > -- > Darrell (Daz) DeBoer > Principal Software Engineer, *Gradleware <http://gradleware.com>* > > Join us for Gradle Summit 2014, June 12th and 13th in Santa Clara, CA: > http://www.gradlesummit.com >
