Hello. I'm currently trying to debug https://issues.apache.org/jira/browse/MASSEMBLY-848 and am reaching this line:
https://github.com/apache/maven-plugins/blob/trunk/maven-assembly-plugin/src/main/java/org/apache/maven/plugins/assembly/artifact/DefaultDependencyResolver.java#L232 I'm debugging based on the contents of an integration test, the original source files of which came from this example project: https://github.com/io7m/independent-versioning-20170207 When line 232 is reached in the debugger, the dependencyArtifacts parameter contains the two artifacts com.io7m.experimental:mod-a:jar:1.1.0:compile and com.io7m.experimental:mod-b:jar:1.2.0:compile and both of those claim to already have been resolved (isResolved() == true), which makes sense as they're reactor projects. Both are present in the given configSource.reactorProjects list. However, the code doesn't appear to make any use of this list or the fact that the artifacts are already resolved. It simply passes the list of artifacts to the resolver which then fails with an OverConstrainedVersionException because the reactor project artifacts aren't in any of the remote or local repositories. This seems to be the root cause of MASSEMBLY-848, but I can't work out whether this is expected behaviour, misuse of an API, an oversight, or possibly all or none of the above! Can anyone shed any light on this? M
pgpO43nka9r6m.pgp
Description: OpenPGP digital signature