[ https://issues.apache.org/jira/browse/MDEPLOY-234?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Robert Scholte closed MDEPLOY-234. ---------------------------------- Resolution: Duplicate Assignee: Robert Scholte I'm going to close this as duplicate. MDEPLOY-226 is the main issue related to the experimental deployAtEnd-option and which might have a good fix for it. > Nothing is deployed if deployAtEnd=true and extension is added in submodule > ---------------------------------------------------------------------------- > > Key: MDEPLOY-234 > URL: https://issues.apache.org/jira/browse/MDEPLOY-234 > Project: Maven Deploy Plugin > Issue Type: Bug > Components: deploy:deploy > Affects Versions: 2.8.2 > Environment: Maven 3.5.3 > Reporter: Piotr Paczynski > Assignee: Robert Scholte > Priority: Major > Attachments: project.zip, result.txt > > > It seems that when *deploytEnd=true* and one of the submodules in the reactor > adds an *extension*, then nothing gets deployed. The last module to build > says "Deploying ... at end". > > I'm attaching a minimal project that is affected by the issue. To reproduce, > unpack it and issue the following command in the main dir (I've used Maven > 3.5.3). > {noformat} > mvn deploy -DaltDeploymentRepository=test::default::file:target/repo > {noformat} > Outcome: Nothing is deployed. The *target/repo* is not created. See output > attached. > Expected result: Deployment of both modules is performed during submodule's > deploy phase. > > I've debugged it and think I understand the problem: > 1. When the parent module is built, the DeployMojo class is loaded and the > {noformat} > private static final AtomicInteger readyProjectsCounter > {noformat} > field is initialized to 0. > 2. Then it's bumped to 1 when deploy of "parent" is executed. > 3. Then, the second module is built (submodule) and, because it contributes > an extension, DepoyMojo class is re-loaded (by Guice?) and the > {noformat} > private static final AtomicInteger readyProjectsCounter{noformat} > field is reinitialized to 0. > 4. Then it's bumped to 1 when deploy of "submodule" is executed, but since *1 > < 2* the deploy is not performed at all. > > Not sure how to fix it, anyone? -- This message was sent by Atlassian JIRA (v7.6.3#76005)