Bug Scrub for Maven 3.2.0

A bug scrub is a review of all the bugs/issues for a specific target version to decide what issues will be addressed for the release.

 

The mailing list thread starts here: http://mail-archives.apache.org/mod_mbox/maven-dev/201401.mbox/%3CCA%2BnPnMw6X9iWc93EB3Zvd_097qLHyTvnKDJO-MmyEZh83rx66w%40mail.gmail.com%3E

Summary

Issues moved out of scope

Undecided

  • Sounds like an issue building the internal model. Additionally this would not be a change that affects other consumers and their processing of dependencies, so this looks like a valid candidate to me.
  • http://jira.codehaus.org/browse/MNG-2807 ciManagement from parent is not merging with children
    Same as MNG-3124. Both issues are related it would seem
  • http://jira.codehaus.org/browse/MNG-4173 Remove automatic version resolution for POM plugins
    This is somewhat reasonable, but we have already kicked this can down the road and it may hinder adoption. I would be happy to kick this one to 4.x on the basis that most existing poms were written with the assumption that you could avoid specifying the plugin version... and we even omit the plugin version in the asf parent pom for some stuff...
  • http://jira.codehaus.org/browse/MNG-3092 Version ranges with non-snapshot bounds can contain snapshot versions
    Do we have a decision as to what we will do with this one? It is one of the longest discussions we have...
  • http://jira.codehaus.org/browse/MNG-5185 Improve "missing dependency" error message when _maven.repositories/_remote.repositories contains other repository ids than requested
    The attached patch does not address the real issue, namely being able to define specific repo id's as offline. I would be happy to take a stab at the real issue, but likely do not have the time. If nobody else has the time, we should move this to 3.2.x as it could be a patch level enhancement to the maven CLI options

Action required

  • http://jira.codehaus.org/browse/MNG-5353 Ignore pre-releases in exclusive upper bound [lw,up)
    ACTION: jvzyl will we be upping Aether to M4, in which case that will expose an alternative version range syntax that resolves this issue... OTOH that new syntax may cause issues for existing pom readers... in which case this becomes a push back to 4.x
     
  • http://jira.codehaus.org/browse/MNG-5205 Memory leak in StringSearchModelInterpolator
  • ACTION: krosenv to provide status update

Issues being worked on

Issues added to scope

 

  • No labels

1 Comment

  1. Robert Scholte

    In general, if I read dependencies related issues, I think it should be pushed to M4.x unless we can change the build-pom back to the expected format before Maven does the install & deploy