[ 
https://issues.apache.org/jira/browse/KARAF-6119?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16789417#comment-16789417
 ] 

ASF subversion and git services commented on KARAF-6119:
--------------------------------------------------------

Commit ab787412baa16f64800aced5934565ccac2881e1 in karaf's branch 
refs/heads/karaf-4.2.x from Freeman Fang
[ https://gitbox.apache.org/repos/asf?p=karaf.git;h=ab78741 ]

[KARAF-6119]karaf-maven-plugin assembly goal doesn't handle timestamp snapshots

(cherry picked from commit 11bda060bfc7fa7dd7893618a407ab87431a24f5)


> karaf-maven-plugin assembly goal doesn't handle locked snapshots
> ----------------------------------------------------------------
>
>                 Key: KARAF-6119
>                 URL: https://issues.apache.org/jira/browse/KARAF-6119
>             Project: Karaf
>          Issue Type: Bug
>          Components: karaf
>    Affects Versions: 4.2.2
>            Reporter: Mike Mansell
>            Assignee: Freeman Fang
>            Priority: Major
>              Labels: maven
>
> When using the assembly goal with the karaf-maven-plugin, and a feature has a 
> bundle that has an explicit snapshot version 
> (<bundle>mvn:GROUPID/ARTIFACTID/0.2.0-20190111.021945-8</bundle>), and that 
> bundle is injected into the system maven directory layout, it incorrectly 
> names the VERSION directory component.
> Currently, it creates a path like
> {{GROUPID/ARTIFACTID/0.2.0-20190111.021945-8/ARTIFACTID-0.2.0-20190111.021945-8.jar}}
> but that will not resolve when using a Maven resolver (such as the Aether 
> that is used in the Karaf runtime). Instead the path needs to be 
> {{GROUPID/ARTIFACTID/0.2.0-SNAPSHOT/ARTIFACTID-0.2.0-20190111.021945-8.jar}}
> It seems that the fix is to add a "base version" capability to the 
> org.apache.karaf.util.maven.Parser class similar to how the 
> org.apache.maven.artifact.DefaultArtifact handles it.
> I've done that in my local copy and seems fine, and I'm happy to provide a 
> PR, however, I don't understand why the Parser class has a 
> getArtifactPath(String) variant. Shouldn't it just use the m_version instance?



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to