[
https://issues.apache.org/jira/browse/MJAVADOC-400?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14660057#comment-14660057
]
Raymond DeCampo commented on MJAVADOC-400:
------------------------------------------
I have also run into this issue although my project structure was different.
Projects A1, A2, ... are modules of project A. Then project A is a module of
project P. Project P is also the parent of A and A1, A2, ...
The javadoc plugin aggregate-jar goal is bound to the package phase in project
A. Running mvn install on project A directly has the expected effect of
building the javadoc jar.
However if project A is built indirectly by building project P, then the
javadoc jar for project A is not built.
I am using maven 3.3.3 and version 2.10.3 of maven-javadoc-plugin.
> aggregate-jar does nothing if called as part of a reactor build
> ---------------------------------------------------------------
>
> Key: MJAVADOC-400
> URL: https://issues.apache.org/jira/browse/MJAVADOC-400
> Project: Maven Javadoc Plugin
> Issue Type: Bug
> Affects Versions: 2.9.1
> Environment: Reproduced with Oracle Java 1.7.0_45, Maven 3.0.5. Also
> reproduced with Maven 3.2.2.
> Reporter: Philip Aston
>
> The attached {{reproduction.tgz}} contains a multi-module project
> _com.repo:parent_, and two child modules, _com.repo:module_a_ and
> _com.repo:module_b_. _com.repo:module-a_ is a jar module that contains a Java
> class. _com.repo:module-b_ depends on _com.repo:module-a_ and has uses
> {{javadoc:aggregate-jar}} to create a javadoc attachment.
> If I change into the {{module_b}} directory and run {{mvn clean install}}, a
> {{com.repo:module-b:javadoc}} attachment is not created.
> However, if I run {{mvn clean install}} from the top level, a
> {{com.repo:module-b:javadoc}} attachment is *not* created. I believe this to
> be a bug.
> Running with -X shows that the aggregate-jar mojo is configured, but never
> executed.
> If I change to {{module-b}} and run with -X, I see an
> MultipleArtifactsNotFoundException, but it otherwise behaves as I would
> expect. This looks similar to the symptoms of MVN-4618.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)