I'll add my opinion as a Maven user and someone outside of the dev team: My first hand choice would be a 3.0.4 with Aether and Sisu fixes. My take on the discussions is that including Aether 1.12 is possible license vice, but there was some concerns having a dependency on something not being within the control of the devs. But if it was at Eclipse everyone (or at least a majority) was fine with it. So, if the ongoing vote passes, wouldn't it be ok to release 3.0.4 with the latest existing Aether and Sisu releases from Sonatype (here I assume that what's being moved to Eclipse is that exact code). Any fixes for that code would then be handled over at Eclipse and used in future Maven releases. After that, the work could start to adapt Maven in what ever way to use the upcoming releases from the new Eclipse projects. It should give you time to do this without rushing and at the same time keep us users happy. :-) I'm also thinking that the release process (vote) on 3.0.4 could start in parallel with the ongoing Aether/Sisu Eclipse/EPL vote and have a dependency to that outcome. Would that be possible? If so, 3.0.4 (if the vote passes) could be released right after Aether and Sisu is accepted at Eclipse (next week I take it if everything goes according to plan).
If the above isn't feasible, I would rather see 3.0.4 without Aether/Sisu fixes released soon than 3.0.4 with Aether/Sisu fixes released in a month. In my opinion it's more important to get a new version out than trying to get as many bug fixes as possible fixed. I know this might not make sense from a usage perspective, but I'm more worried about what people thinks just looking at how long it takes between releases. Release early, release often. /Anders On Fri, Aug 19, 2011 at 13:45, Jason van Zyl <[email protected]> wrote: > I think we should wait for the vote to finish as it appears the vote will > pass and then a release can be made that users actually want. That fixes are > are most requested fall into the category of requiring the newer versions of > Aether and Sisu. > > I'm certain users would be more interested in having a version with Aether > 1.12 which has been released. So if we don't want to cut a release with that > in it then waiting another month, I believe, is better than rushing out some > release now with 6 fixes which are minor and making all users drop in > libraries themselves to fix the issues that are hurting them most. If we want > to wait for official Eclipse releases then we should just wait. > > Users who are on the dev list feel free to chime in. > > ------------------------------------ > Summary of the issues > ------------------------------------ > > [MNG-5064][1] mvn -nsu (--no-snapshot-updates) should not download snapshots > (and break local builds) > [MNG-5131][2] Wrong encoding for encrypted passwords > [MNG-5113][3] NullPointerException on javadoc site generation > [MNG-5137][4] Reactor resolution does not work for forked multi module builds > [MNG-5096][5] <exclusion> on <dependency> with <type>test-jar</type> doesn't > work in maven 3 > [MNG-5135][6] Regression: in some cases aggregator mojo is unable to resolve > dependencies with custom packaging > > Fixes not pushed back to the ASF as these are dependent on fixes in Eclipse > Aether and Eclipse Sisu: > > [MNG-5042][7] Regression: CloningClassLoader causes StackOverflowError in > groovy > [MNG-5056][8] Test dependencies get packaged into WAR file. > [MNG-5084][9] Resolver for plugins failing > [MNG-5087][10] Maven 3 dependency resolution fails until > maven-metadata-local.xml files (created by maven-invoker-plugin) are deleted > [MNG-5125] [11]Regression: mvn 3.0.3 is extreemly slow with a large number of > dependencies > [MNG-5138][12] Dependency conflicts are extremely opaque > > On Aug 19, 2011, at 7:32 AM, Mark Struberg wrote: > >> I think the outcome of the other vote doesn't change much. >> The following scenarios are possible: >> >> a.) The vote fails and maven stays at aether 1.11. In this case we still >> should ship a 3.0.4 to fix a few bugs on our side. >> >> b.) The vote passes and aether @eclipse must change package names + GAV then >> it will certainly take a bit time on both sides. So users still could >> upgrade to aether-1.12 manually if they want. >> >> c.) The vote passes and aether GAV and package names are allowed to remain >> com.sonatype.* In this case users might also manually update if they want. >> >> In any case we should ship maven-3.0.5 once there is a new aether version >> shipped by the Eclipse Foundation AND we did run sufficient tests with this >> version on our side. >> >> LieGrue, >> strub >> >> >> --- On Fri, 8/19/11, Jason van Zyl <[email protected]> wrote: >> >>> From: Jason van Zyl <[email protected]> >>> Subject: Re: [DISCUSS] release maven-3.0.4 now with the old aether >>> To: "Maven Developers List" <[email protected]> >>> Date: Friday, August 19, 2011, 11:09 AM >>> Mark, >>> >>> I think you're jumping the gun a bit. The vote that Arnaud >>> proposed is not finished yet. Many PMC members have not >>> voted yet and at the very least to short-circuit the vote we >>> need to reach a majority amongst the PMC. Once that happens >>> then I think we can start discussing a 3.0.4 release. >>> >>> On Aug 19, 2011, at 5:34 AM, Mark Struberg wrote: >>> >>>> Hi folks! >>>> >>>> I know aether @eclipse should be starting in the next >>> weeks. But it will certainly take a few more weeks to get >>> all the packages renamed (most probably), etc. >>>> >>>> Should we create a 3.0.4 with the old aether package >>> names now and start with the upgrade later? >>>> >>>> +1 from me. >>>> >>>> LieGrue, >>>> strub >>>> >>>> >>> --------------------------------------------------------------------- >>>> To unsubscribe, e-mail: [email protected] >>>> For additional commands, e-mail: [email protected] >>>> >>> >>> Thanks, >>> >>> Jason >>> >>> ---------------------------------------------------------- >>> Jason van Zyl >>> Founder, Apache Maven >>> http://twitter.com/jvanzyl >>> --------------------------------------------------------- >>> >>> Simplex sigillum veri. (Simplicity is the seal of truth.) >>> >>> >>> >>> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [email protected] >> For additional commands, e-mail: [email protected] >> > > Thanks, > > Jason > > ---------------------------------------------------------- > Jason van Zyl > Founder, Apache Maven > http://twitter.com/jvanzyl > --------------------------------------------------------- > > There's no sense in being precise when you don't even know what you're > talking about. > > -- John von Neumann > > > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
